{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 3.0, "eval_steps": 500, "global_step": 23583, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.00012721027859051011, "ewc_loss": 0.0, "ewc_loss_diag": 0.0, "ewc_loss_parallel": 0.0, "grad_norm": 4.265416145324707, "learning_rate": 0.0, "loss": 0.9782, "mean_token_accuracy": 0.7547097206115723, "num_tokens": 38293.0, "step": 1 }, { "epoch": 0.00025442055718102023, "ewc_loss": 0.0, "ewc_loss_diag": 0.0, "ewc_loss_parallel": 0.0, "grad_norm": 4.474235534667969, "learning_rate": 4.23908435777872e-09, "loss": 1.0326, "mean_token_accuracy": 0.7414199113845825, "num_tokens": 79961.0, "step": 2 }, { "epoch": 0.0003816308357715303, "ewc_loss": 2.611244553918368e-13, "ewc_loss_diag": 1.8648277366750676e-17, "ewc_loss_parallel": 2.6020852139652106e-16, "grad_norm": 4.450266361236572, "learning_rate": 8.47816871555744e-09, "loss": 0.9023, "mean_token_accuracy": 0.7777742147445679, "num_tokens": 118063.0, "step": 3 }, { "epoch": 0.0005088411143620405, "ewc_loss": 3.311129148642067e-12, "ewc_loss_diag": 2.0990154059319366e-16, "ewc_loss_parallel": 3.3029134982598407e-15, "grad_norm": 4.457966327667236, "learning_rate": 1.271725307333616e-08, "loss": 0.9761, "mean_token_accuracy": 0.7541318535804749, "num_tokens": 149306.0, "step": 4 }, { "epoch": 0.0006360513929525506, "ewc_loss": 2.546585164964199e-11, "ewc_loss_diag": 1.6514567491299204e-15, "ewc_loss_parallel": 2.531308496145357e-14, "grad_norm": 4.246554374694824, "learning_rate": 1.695633743111488e-08, "loss": 0.9984, "mean_token_accuracy": 0.7504554390907288, "num_tokens": 192477.0, "step": 5 }, { "epoch": 0.0007632616715430606, "ewc_loss": 1.4370016288012266e-10, "ewc_loss_diag": 9.71445146547012e-15, "ewc_loss_parallel": 1.4299672557172016e-13, "grad_norm": 4.64027214050293, "learning_rate": 2.11954217888936e-08, "loss": 0.9609, "mean_token_accuracy": 0.7626928687095642, "num_tokens": 226246.0, "step": 6 }, { "epoch": 0.0008904719501335708, "ewc_loss": 5.165929906070232e-10, "ewc_loss_diag": 4.529709940470639e-14, "ewc_loss_parallel": 5.115907697472721e-13, "grad_norm": 5.0900421142578125, "learning_rate": 2.543450614667232e-08, "loss": 1.0128, "mean_token_accuracy": 0.7551836967468262, "num_tokens": 263593.0, "step": 7 }, { "epoch": 0.001017682228724081, "ewc_loss": 1.1423253454267979e-09, "ewc_loss_diag": 7.860379014346108e-14, "ewc_loss_parallel": 1.1368683772161603e-12, "grad_norm": 4.313149929046631, "learning_rate": 2.967359050445104e-08, "loss": 0.9101, "mean_token_accuracy": 0.7725673913955688, "num_tokens": 298202.0, "step": 8 }, { "epoch": 0.001144892507314591, "ewc_loss": 1.7971615307033062e-09, "ewc_loss_diag": 1.1546319456101628e-13, "ewc_loss_parallel": 1.7834622667578515e-12, "grad_norm": 4.465803146362305, "learning_rate": 3.391267486222976e-08, "loss": 1.0363, "mean_token_accuracy": 0.7476455569267273, "num_tokens": 340163.0, "step": 9 }, { "epoch": 0.0012721027859051012, "ewc_loss": 7.159542292356491e-09, "ewc_loss_diag": 4.973799150320701e-13, "ewc_loss_parallel": 7.105427357601002e-12, "grad_norm": 4.657181739807129, "learning_rate": 3.815175922000848e-08, "loss": 1.0424, "mean_token_accuracy": 0.7432540059089661, "num_tokens": 372703.0, "step": 10 }, { "epoch": 0.0013993130644956112, "ewc_loss": 1.0011717677116394e-08, "ewc_loss_diag": 7.034373084024992e-13, "ewc_loss_parallel": 9.947598300641403e-12, "grad_norm": 4.1998515129089355, "learning_rate": 4.23908435777872e-08, "loss": 0.977, "mean_token_accuracy": 0.7514074444770813, "num_tokens": 414187.0, "step": 11 }, { "epoch": 0.0015265233430861213, "ewc_loss": 1.3445969671010971e-08, "ewc_loss_diag": 8.775202786637237e-13, "ewc_loss_parallel": 1.3358203432289883e-11, "grad_norm": 4.611934661865234, "learning_rate": 4.662992793556592e-08, "loss": 1.0082, "mean_token_accuracy": 0.7497714161872864, "num_tokens": 446133.0, "step": 12 }, { "epoch": 0.0016537336216766315, "ewc_loss": 4.21423465013504e-08, "ewc_loss_diag": 3.439026841078885e-12, "ewc_loss_parallel": 4.18367562815547e-11, "grad_norm": 4.733109474182129, "learning_rate": 5.086901229334464e-08, "loss": 1.0514, "mean_token_accuracy": 0.7319505214691162, "num_tokens": 477130.0, "step": 13 }, { "epoch": 0.0017809439002671415, "ewc_loss": 6.426125764846802e-08, "ewc_loss_diag": 4.320099833421409e-12, "ewc_loss_parallel": 6.366462912410498e-11, "grad_norm": 4.273692607879639, "learning_rate": 5.510809665112336e-08, "loss": 1.0086, "mean_token_accuracy": 0.7585204243659973, "num_tokens": 521754.0, "step": 14 }, { "epoch": 0.0019081541788576518, "ewc_loss": 7.916241884231567e-08, "ewc_loss_diag": 5.6274984672199935e-12, "ewc_loss_parallel": 7.867129170335829e-11, "grad_norm": 4.645796775817871, "learning_rate": 5.934718100890208e-08, "loss": 0.9558, "mean_token_accuracy": 0.7665998935699463, "num_tokens": 560311.0, "step": 15 }, { "epoch": 0.002035364457448162, "ewc_loss": 9.918585419654846e-08, "ewc_loss_diag": 6.622258297284134e-12, "ewc_loss_parallel": 9.868017514236271e-11, "grad_norm": 4.39243745803833, "learning_rate": 6.358626536668081e-08, "loss": 1.0334, "mean_token_accuracy": 0.7411091327667236, "num_tokens": 595214.0, "step": 16 }, { "epoch": 0.002162574736038672, "ewc_loss": 1.955777406692505e-07, "ewc_loss_diag": 1.887201506178826e-11, "ewc_loss_parallel": 1.9372237147763371e-10, "grad_norm": 4.847776412963867, "learning_rate": 6.782534972445952e-08, "loss": 1.0435, "mean_token_accuracy": 0.7417346239089966, "num_tokens": 631427.0, "step": 17 }, { "epoch": 0.002289785014629182, "ewc_loss": 3.5390257835388184e-07, "ewc_loss_diag": 2.830802259268239e-11, "ewc_loss_parallel": 3.510649548843503e-10, "grad_norm": 4.85832405090332, "learning_rate": 7.206443408223825e-08, "loss": 1.0942, "mean_token_accuracy": 0.7347165942192078, "num_tokens": 671155.0, "step": 18 }, { "epoch": 0.0024169952932196924, "ewc_loss": 4.76837158203125e-07, "ewc_loss_diag": 3.2741809263825417e-11, "ewc_loss_parallel": 4.729372449219227e-10, "grad_norm": 5.5069169998168945, "learning_rate": 7.630351844001696e-08, "loss": 0.9984, "mean_token_accuracy": 0.753947377204895, "num_tokens": 704515.0, "step": 19 }, { "epoch": 0.0025442055718102024, "ewc_loss": 5.550682544708252e-07, "ewc_loss_diag": 3.8198777474462986e-11, "ewc_loss_parallel": 5.529727786779404e-10, "grad_norm": 4.477396488189697, "learning_rate": 8.054260279779568e-08, "loss": 0.9739, "mean_token_accuracy": 0.7597118616104126, "num_tokens": 745414.0, "step": 20 }, { "epoch": 0.0026714158504007124, "ewc_loss": 6.593763828277588e-07, "ewc_loss_diag": 4.661160346586257e-11, "ewc_loss_parallel": 6.548361852765083e-10, "grad_norm": 4.602478504180908, "learning_rate": 8.47816871555744e-08, "loss": 1.0254, "mean_token_accuracy": 0.741905689239502, "num_tokens": 779384.0, "step": 21 }, { "epoch": 0.0027986261289912225, "ewc_loss": 7.7858567237854e-07, "ewc_loss_diag": 5.206857167650014e-11, "ewc_loss_parallel": 7.748894859105349e-10, "grad_norm": 4.4505391120910645, "learning_rate": 8.902077151335312e-08, "loss": 1.0133, "mean_token_accuracy": 0.7495947480201721, "num_tokens": 812994.0, "step": 22 }, { "epoch": 0.0029258364075817325, "ewc_loss": 1.0505318641662598e-06, "ewc_loss_diag": 1.0504663805477321e-10, "ewc_loss_parallel": 1.04046193882823e-09, "grad_norm": 4.408501148223877, "learning_rate": 9.325985587113183e-08, "loss": 0.9538, "mean_token_accuracy": 0.7668715715408325, "num_tokens": 849351.0, "step": 23 }, { "epoch": 0.0030530466861722425, "ewc_loss": 2.115964889526367e-06, "ewc_loss_diag": 1.864464138634503e-10, "ewc_loss_parallel": 2.0954757928848267e-09, "grad_norm": 4.650966644287109, "learning_rate": 9.749894022891056e-08, "loss": 0.9975, "mean_token_accuracy": 0.7481536865234375, "num_tokens": 880319.0, "step": 24 }, { "epoch": 0.003180256964762753, "ewc_loss": 2.8312206268310547e-06, "ewc_loss_diag": 2.2646418074145913e-10, "ewc_loss_parallel": 2.7939677238464355e-09, "grad_norm": 4.62087345123291, "learning_rate": 1.0173802458668929e-07, "loss": 1.0919, "mean_token_accuracy": 0.7259002923965454, "num_tokens": 921902.0, "step": 25 }, { "epoch": 0.003307467243353263, "ewc_loss": 3.4868717193603516e-06, "ewc_loss_diag": 2.546585164964199e-10, "ewc_loss_parallel": 3.4633558243513107e-09, "grad_norm": 4.635985851287842, "learning_rate": 1.05977108944468e-07, "loss": 1.0277, "mean_token_accuracy": 0.7429295182228088, "num_tokens": 959531.0, "step": 26 }, { "epoch": 0.003434677521943773, "ewc_loss": 4.112720489501953e-06, "ewc_loss_diag": 2.764863893389702e-10, "ewc_loss_parallel": 4.0745362639427185e-09, "grad_norm": 4.141330718994141, "learning_rate": 1.1021619330224672e-07, "loss": 0.9438, "mean_token_accuracy": 0.7642838954925537, "num_tokens": 997280.0, "step": 27 }, { "epoch": 0.003561887800534283, "ewc_loss": 4.4405460357666016e-06, "ewc_loss_diag": 3.255991032347083e-10, "ewc_loss_parallel": 4.423782229423523e-09, "grad_norm": 4.814310073852539, "learning_rate": 1.1445527766002544e-07, "loss": 1.0819, "mean_token_accuracy": 0.7312692403793335, "num_tokens": 1034608.0, "step": 28 }, { "epoch": 0.003689098079124793, "ewc_loss": 5.0961971282958984e-06, "ewc_loss_diag": 3.583409124985337e-10, "ewc_loss_parallel": 5.0640664994716644e-09, "grad_norm": 4.225417613983154, "learning_rate": 1.1869436201780416e-07, "loss": 0.9639, "mean_token_accuracy": 0.7620096802711487, "num_tokens": 1075907.0, "step": 29 }, { "epoch": 0.0038163083577153036, "ewc_loss": 5.543231964111328e-06, "ewc_loss_diag": 3.9108272176235914e-10, "ewc_loss_parallel": 5.50062395632267e-09, "grad_norm": 4.427147388458252, "learning_rate": 1.229334463755829e-07, "loss": 1.0138, "mean_token_accuracy": 0.7546048760414124, "num_tokens": 1114492.0, "step": 30 }, { "epoch": 0.003943518636305814, "ewc_loss": 6.0498714447021484e-06, "ewc_loss_diag": 4.256435204297304e-10, "ewc_loss_parallel": 6.024492904543877e-09, "grad_norm": 4.291891098022461, "learning_rate": 1.2717253073336161e-07, "loss": 0.9358, "mean_token_accuracy": 0.7731562852859497, "num_tokens": 1151596.0, "step": 31 }, { "epoch": 0.004070728914896324, "ewc_loss": 8.404254913330078e-06, "ewc_loss_diag": 8.87666828930378e-10, "ewc_loss_parallel": 8.32369551062584e-09, "grad_norm": 4.651190280914307, "learning_rate": 1.3141161509114031e-07, "loss": 1.0234, "mean_token_accuracy": 0.751585841178894, "num_tokens": 1191475.0, "step": 32 }, { "epoch": 0.004197939193486834, "ewc_loss": 1.3768672943115234e-05, "ewc_loss_diag": 1.3460521586239338e-09, "ewc_loss_parallel": 1.367880031466484e-08, "grad_norm": 4.301553726196289, "learning_rate": 1.3565069944891904e-07, "loss": 1.0032, "mean_token_accuracy": 0.7539398670196533, "num_tokens": 1230726.0, "step": 33 }, { "epoch": 0.004325149472077344, "ewc_loss": 1.8477439880371094e-05, "ewc_loss_diag": 1.6516423784196377e-09, "ewc_loss_parallel": 1.83936208486557e-08, "grad_norm": 4.196382999420166, "learning_rate": 1.3988978380669777e-07, "loss": 0.9669, "mean_token_accuracy": 0.7632488012313843, "num_tokens": 1273523.0, "step": 34 }, { "epoch": 0.004452359750667854, "ewc_loss": 2.2172927856445312e-05, "ewc_loss_diag": 1.82626536116004e-09, "ewc_loss_parallel": 2.1886080503463745e-08, "grad_norm": 4.254756927490234, "learning_rate": 1.441288681644765e-07, "loss": 0.9693, "mean_token_accuracy": 0.7627424001693726, "num_tokens": 1312013.0, "step": 35 }, { "epoch": 0.004579570029258364, "ewc_loss": 2.5510787963867188e-05, "ewc_loss_diag": 1.9936123862862587e-09, "ewc_loss_parallel": 2.5262124836444855e-08, "grad_norm": 4.597464561462402, "learning_rate": 1.4836795252225522e-07, "loss": 0.9968, "mean_token_accuracy": 0.7516172528266907, "num_tokens": 1350139.0, "step": 36 }, { "epoch": 0.004706780307848874, "ewc_loss": 2.8133392333984375e-05, "ewc_loss_diag": 2.1100277081131935e-09, "ewc_loss_parallel": 2.7939677238464355e-08, "grad_norm": 4.29860258102417, "learning_rate": 1.5260703688003392e-07, "loss": 0.967, "mean_token_accuracy": 0.7589424252510071, "num_tokens": 1387670.0, "step": 37 }, { "epoch": 0.004833990586439385, "ewc_loss": 3.0517578125e-05, "ewc_loss_diag": 2.240994945168495e-09, "ewc_loss_parallel": 3.026798367500305e-08, "grad_norm": 4.062749862670898, "learning_rate": 1.5684612123781264e-07, "loss": 0.9722, "mean_token_accuracy": 0.7589689493179321, "num_tokens": 1431658.0, "step": 38 }, { "epoch": 0.004961200865029895, "ewc_loss": 3.314018249511719e-05, "ewc_loss_diag": 2.546585164964199e-09, "ewc_loss_parallel": 3.282912075519562e-08, "grad_norm": 4.183523178100586, "learning_rate": 1.6108520559559137e-07, "loss": 0.9045, "mean_token_accuracy": 0.777551531791687, "num_tokens": 1467751.0, "step": 39 }, { "epoch": 0.005088411143620405, "ewc_loss": 3.504753112792969e-05, "ewc_loss_diag": 2.8085196390748024e-09, "ewc_loss_parallel": 3.4691765904426575e-08, "grad_norm": 4.358862400054932, "learning_rate": 1.6532428995337007e-07, "loss": 1.0037, "mean_token_accuracy": 0.7465317249298096, "num_tokens": 1507078.0, "step": 40 }, { "epoch": 0.005215621422210915, "ewc_loss": 3.6716461181640625e-05, "ewc_loss_diag": 2.954038791358471e-09, "ewc_loss_parallel": 3.655441105365753e-08, "grad_norm": 4.264369964599609, "learning_rate": 1.695633743111488e-07, "loss": 0.99, "mean_token_accuracy": 0.755617618560791, "num_tokens": 1544856.0, "step": 41 }, { "epoch": 0.005342831700801425, "ewc_loss": 3.886222839355469e-05, "ewc_loss_diag": 3.157765604555607e-09, "ewc_loss_parallel": 3.864988684654236e-08, "grad_norm": 4.448814868927002, "learning_rate": 1.7380245866892752e-07, "loss": 0.9776, "mean_token_accuracy": 0.7604350447654724, "num_tokens": 1576776.0, "step": 42 }, { "epoch": 0.005470041979391935, "ewc_loss": 4.029273986816406e-05, "ewc_loss_diag": 3.346940502524376e-09, "ewc_loss_parallel": 4.0046870708465576e-08, "grad_norm": 4.156569957733154, "learning_rate": 1.7804154302670624e-07, "loss": 0.9685, "mean_token_accuracy": 0.759522020816803, "num_tokens": 1613819.0, "step": 43 }, { "epoch": 0.005597252257982445, "ewc_loss": 4.220008850097656e-05, "ewc_loss_diag": 4.0745362639427185e-09, "ewc_loss_parallel": 4.1676685214042664e-08, "grad_norm": 4.502416610717773, "learning_rate": 1.8228062738448497e-07, "loss": 1.0271, "mean_token_accuracy": 0.7471545338630676, "num_tokens": 1654834.0, "step": 44 }, { "epoch": 0.005724462536572955, "ewc_loss": 6.628036499023438e-05, "ewc_loss_diag": 7.799826562404633e-09, "ewc_loss_parallel": 6.51925802230835e-08, "grad_norm": 4.120793342590332, "learning_rate": 1.8651971174226367e-07, "loss": 0.9961, "mean_token_accuracy": 0.7508350014686584, "num_tokens": 1691238.0, "step": 45 }, { "epoch": 0.005851672815163465, "ewc_loss": 9.107589721679688e-05, "ewc_loss_diag": 1.0011717677116394e-08, "ewc_loss_parallel": 9.033828973770142e-08, "grad_norm": 3.7467639446258545, "learning_rate": 1.907587961000424e-07, "loss": 0.884, "mean_token_accuracy": 0.7720924019813538, "num_tokens": 1734446.0, "step": 46 }, { "epoch": 0.005978883093753975, "ewc_loss": 0.000118255615234375, "ewc_loss_diag": 1.1757947504520416e-08, "ewc_loss_parallel": 1.1734664440155029e-07, "grad_norm": 3.9790782928466797, "learning_rate": 1.9499788045782112e-07, "loss": 0.9498, "mean_token_accuracy": 0.7588983774185181, "num_tokens": 1770266.0, "step": 47 }, { "epoch": 0.006106093372344485, "ewc_loss": 0.00013828277587890625, "ewc_loss_diag": 1.30385160446167e-08, "ewc_loss_parallel": 1.3690441846847534e-07, "grad_norm": 4.005307674407959, "learning_rate": 1.9923696481559985e-07, "loss": 1.0075, "mean_token_accuracy": 0.7407536506652832, "num_tokens": 1807979.0, "step": 48 }, { "epoch": 0.006233303650934996, "ewc_loss": 0.00014972686767578125, "ewc_loss_diag": 1.391163095831871e-08, "ewc_loss_parallel": 1.4901161193847656e-07, "grad_norm": 3.714128017425537, "learning_rate": 2.0347604917337857e-07, "loss": 0.8976, "mean_token_accuracy": 0.7693262696266174, "num_tokens": 1851061.0, "step": 49 }, { "epoch": 0.006360513929525506, "ewc_loss": 0.00015926361083984375, "ewc_loss_diag": 1.4842953532934189e-08, "ewc_loss_parallel": 1.5739351511001587e-07, "grad_norm": 3.5893588066101074, "learning_rate": 2.0771513353115727e-07, "loss": 0.8536, "mean_token_accuracy": 0.7845635414123535, "num_tokens": 1887647.0, "step": 50 }, { "epoch": 0.006487724208116016, "ewc_loss": 0.0001697540283203125, "ewc_loss_diag": 1.5599653124809265e-08, "ewc_loss_parallel": 1.6763806343078613e-07, "grad_norm": 4.193203449249268, "learning_rate": 2.11954217888936e-07, "loss": 1.02, "mean_token_accuracy": 0.7432173490524292, "num_tokens": 1924843.0, "step": 51 }, { "epoch": 0.006614934486706526, "ewc_loss": 0.00017642974853515625, "ewc_loss_diag": 1.594889909029007e-08, "ewc_loss_parallel": 1.7415732145309448e-07, "grad_norm": 4.1089630126953125, "learning_rate": 2.1619330224671472e-07, "loss": 0.9464, "mean_token_accuracy": 0.7624565362930298, "num_tokens": 1967116.0, "step": 52 }, { "epoch": 0.006742144765297036, "ewc_loss": 0.00018215179443359375, "ewc_loss_diag": 1.6763806343078613e-08, "ewc_loss_parallel": 1.8067657947540283e-07, "grad_norm": 3.730487108230591, "learning_rate": 2.2043238660449345e-07, "loss": 0.9282, "mean_token_accuracy": 0.7581252455711365, "num_tokens": 1999575.0, "step": 53 }, { "epoch": 0.006869355043887546, "ewc_loss": 0.00018596649169921875, "ewc_loss_diag": 1.8510036170482635e-08, "ewc_loss_parallel": 1.8440186977386475e-07, "grad_norm": 4.161142826080322, "learning_rate": 2.2467147096227215e-07, "loss": 1.0157, "mean_token_accuracy": 0.7414747476577759, "num_tokens": 2039396.0, "step": 54 }, { "epoch": 0.006996565322478056, "ewc_loss": 0.0001926422119140625, "ewc_loss_diag": 2.0023435354232788e-08, "ewc_loss_parallel": 1.909211277961731e-07, "grad_norm": 4.04760217666626, "learning_rate": 2.2891055532005087e-07, "loss": 1.0474, "mean_token_accuracy": 0.7320062518119812, "num_tokens": 2080258.0, "step": 55 }, { "epoch": 0.007123775601068566, "ewc_loss": 0.00019741058349609375, "ewc_loss_diag": 2.1420419216156006e-08, "ewc_loss_parallel": 1.955777406692505e-07, "grad_norm": 3.969210624694824, "learning_rate": 2.331496396778296e-07, "loss": 1.024, "mean_token_accuracy": 0.7412161827087402, "num_tokens": 2116552.0, "step": 56 }, { "epoch": 0.007250985879659076, "ewc_loss": 0.00020313262939453125, "ewc_loss_diag": 2.200249582529068e-08, "ewc_loss_parallel": 2.0116567611694336e-07, "grad_norm": 3.6672236919403076, "learning_rate": 2.3738872403560833e-07, "loss": 0.9224, "mean_token_accuracy": 0.765477180480957, "num_tokens": 2161097.0, "step": 57 }, { "epoch": 0.007378196158249586, "ewc_loss": 0.00020503997802734375, "ewc_loss_diag": 2.293381839990616e-08, "ewc_loss_parallel": 2.0302832126617432e-07, "grad_norm": 3.532313823699951, "learning_rate": 2.4162780839338705e-07, "loss": 0.8786, "mean_token_accuracy": 0.7758214473724365, "num_tokens": 2201190.0, "step": 58 }, { "epoch": 0.007505406436840096, "ewc_loss": 0.0002079010009765625, "ewc_loss_diag": 2.3865140974521637e-08, "ewc_loss_parallel": 2.0489096641540527e-07, "grad_norm": 4.035674571990967, "learning_rate": 2.458668927511658e-07, "loss": 1.0167, "mean_token_accuracy": 0.7401031255722046, "num_tokens": 2233725.0, "step": 59 }, { "epoch": 0.007632616715430607, "ewc_loss": 0.000209808349609375, "ewc_loss_diag": 2.5262124836444855e-08, "ewc_loss_parallel": 2.0675361156463623e-07, "grad_norm": 3.8553760051727295, "learning_rate": 2.501059771089445e-07, "loss": 0.9636, "mean_token_accuracy": 0.7509307265281677, "num_tokens": 2274459.0, "step": 60 }, { "epoch": 0.007759826994021117, "ewc_loss": 0.0002288818359375, "ewc_loss_diag": 2.735760062932968e-08, "ewc_loss_parallel": 2.2631138563156128e-07, "grad_norm": 3.65663480758667, "learning_rate": 2.5434506146672323e-07, "loss": 0.9112, "mean_token_accuracy": 0.7661662101745605, "num_tokens": 2317530.0, "step": 61 }, { "epoch": 0.007887037272611627, "ewc_loss": 0.0003490447998046875, "ewc_loss_diag": 3.771856427192688e-08, "ewc_loss_parallel": 3.4458935260772705e-07, "grad_norm": 3.6496036052703857, "learning_rate": 2.5858414582450195e-07, "loss": 1.0005, "mean_token_accuracy": 0.747557520866394, "num_tokens": 2355260.0, "step": 62 }, { "epoch": 0.008014247551202136, "ewc_loss": 0.000499725341796875, "ewc_loss_diag": 5.145557224750519e-08, "ewc_loss_parallel": 4.954636096954346e-07, "grad_norm": 3.145308494567871, "learning_rate": 2.6282323018228063e-07, "loss": 0.8845, "mean_token_accuracy": 0.7710457444190979, "num_tokens": 2397135.0, "step": 63 }, { "epoch": 0.008141457829792647, "ewc_loss": 0.00060272216796875, "ewc_loss_diag": 6.239861249923706e-08, "ewc_loss_parallel": 5.960464477539062e-07, "grad_norm": 3.173894166946411, "learning_rate": 2.6706231454005935e-07, "loss": 0.9236, "mean_token_accuracy": 0.7588900923728943, "num_tokens": 2440206.0, "step": 64 }, { "epoch": 0.008268668108383158, "ewc_loss": 0.000652313232421875, "ewc_loss_diag": 6.891787052154541e-08, "ewc_loss_parallel": 6.444752216339111e-07, "grad_norm": 3.427717685699463, "learning_rate": 2.713013988978381e-07, "loss": 0.9615, "mean_token_accuracy": 0.7407726645469666, "num_tokens": 2469565.0, "step": 65 }, { "epoch": 0.008395878386973667, "ewc_loss": 0.00070953369140625, "ewc_loss_diag": 7.636845111846924e-08, "ewc_loss_parallel": 7.003545761108398e-07, "grad_norm": 2.9039881229400635, "learning_rate": 2.755404832556168e-07, "loss": 0.8909, "mean_token_accuracy": 0.7620964050292969, "num_tokens": 2514314.0, "step": 66 }, { "epoch": 0.008523088665564178, "ewc_loss": 0.00074005126953125, "ewc_loss_diag": 8.288770914077759e-08, "ewc_loss_parallel": 7.338821887969971e-07, "grad_norm": 3.205686569213867, "learning_rate": 2.7977956761339553e-07, "loss": 0.9341, "mean_token_accuracy": 0.751189112663269, "num_tokens": 2558466.0, "step": 67 }, { "epoch": 0.008650298944154687, "ewc_loss": 0.000762939453125, "ewc_loss_diag": 8.800998330116272e-08, "ewc_loss_parallel": 7.562339305877686e-07, "grad_norm": 2.743039608001709, "learning_rate": 2.8401865197117426e-07, "loss": 0.8717, "mean_token_accuracy": 0.763230562210083, "num_tokens": 2596843.0, "step": 68 }, { "epoch": 0.008777509222745198, "ewc_loss": 0.0007781982421875, "ewc_loss_diag": 9.313225746154785e-08, "ewc_loss_parallel": 7.711350917816162e-07, "grad_norm": 2.8852081298828125, "learning_rate": 2.88257736328953e-07, "loss": 0.8601, "mean_token_accuracy": 0.7659891843795776, "num_tokens": 2632933.0, "step": 69 }, { "epoch": 0.008904719501335707, "ewc_loss": 0.00078582763671875, "ewc_loss_diag": 9.872019290924072e-08, "ewc_loss_parallel": 7.748603820800781e-07, "grad_norm": 2.9093499183654785, "learning_rate": 2.924968206867317e-07, "loss": 0.8414, "mean_token_accuracy": 0.7805991768836975, "num_tokens": 2673526.0, "step": 70 }, { "epoch": 0.009031929779926218, "ewc_loss": 0.00079345703125, "ewc_loss_diag": 1.0337680578231812e-07, "ewc_loss_parallel": 7.82310962677002e-07, "grad_norm": 3.013561964035034, "learning_rate": 2.9673590504451043e-07, "loss": 0.8824, "mean_token_accuracy": 0.764911949634552, "num_tokens": 2707770.0, "step": 71 }, { "epoch": 0.009159140058516728, "ewc_loss": 0.00078582763671875, "ewc_loss_diag": 1.0477378964424133e-07, "ewc_loss_parallel": 7.748603820800781e-07, "grad_norm": 3.1285810470581055, "learning_rate": 3.009749894022891e-07, "loss": 0.8441, "mean_token_accuracy": 0.7769902348518372, "num_tokens": 2745236.0, "step": 72 }, { "epoch": 0.009286350337107238, "ewc_loss": 0.00078582763671875, "ewc_loss_diag": 1.0989606380462646e-07, "ewc_loss_parallel": 7.748603820800781e-07, "grad_norm": 2.979541063308716, "learning_rate": 3.0521407376006783e-07, "loss": 0.9193, "mean_token_accuracy": 0.7538456320762634, "num_tokens": 2780291.0, "step": 73 }, { "epoch": 0.009413560615697748, "ewc_loss": 0.00075531005859375, "ewc_loss_diag": 1.1455267667770386e-07, "ewc_loss_parallel": 7.450580596923828e-07, "grad_norm": 3.0118134021759033, "learning_rate": 3.0945315811784656e-07, "loss": 0.9706, "mean_token_accuracy": 0.7467426657676697, "num_tokens": 2814771.0, "step": 74 }, { "epoch": 0.009540770894288259, "ewc_loss": 0.000743865966796875, "ewc_loss_diag": 1.2014061212539673e-07, "ewc_loss_parallel": 7.338821887969971e-07, "grad_norm": 2.743212938308716, "learning_rate": 3.136922424756253e-07, "loss": 0.8449, "mean_token_accuracy": 0.7742516994476318, "num_tokens": 2861375.0, "step": 75 }, { "epoch": 0.00966798117287877, "ewc_loss": 0.000728607177734375, "ewc_loss_diag": 1.2479722499847412e-07, "ewc_loss_parallel": 7.152557373046875e-07, "grad_norm": 2.592716693878174, "learning_rate": 3.17931326833404e-07, "loss": 0.8388, "mean_token_accuracy": 0.7705625295639038, "num_tokens": 2896139.0, "step": 76 }, { "epoch": 0.009795191451469279, "ewc_loss": 0.000713348388671875, "ewc_loss_diag": 1.2945383787155151e-07, "ewc_loss_parallel": 7.003545761108398e-07, "grad_norm": 2.5802979469299316, "learning_rate": 3.2217041119118274e-07, "loss": 0.7981, "mean_token_accuracy": 0.7858733534812927, "num_tokens": 2940269.0, "step": 77 }, { "epoch": 0.00992240173005979, "ewc_loss": 0.000690460205078125, "ewc_loss_diag": 1.3317912817001343e-07, "ewc_loss_parallel": 6.780028343200684e-07, "grad_norm": 2.6887741088867188, "learning_rate": 3.2640949554896146e-07, "loss": 0.847, "mean_token_accuracy": 0.770038902759552, "num_tokens": 2971115.0, "step": 78 }, { "epoch": 0.010049612008650299, "ewc_loss": 0.000682830810546875, "ewc_loss_diag": 1.387670636177063e-07, "ewc_loss_parallel": 6.668269634246826e-07, "grad_norm": 2.40413761138916, "learning_rate": 3.3064857990674013e-07, "loss": 0.7407, "mean_token_accuracy": 0.7953504323959351, "num_tokens": 3003727.0, "step": 79 }, { "epoch": 0.01017682228724081, "ewc_loss": 0.00066375732421875, "ewc_loss_diag": 1.4435499906539917e-07, "ewc_loss_parallel": 6.48200511932373e-07, "grad_norm": 2.451906442642212, "learning_rate": 3.3488766426451886e-07, "loss": 0.779, "mean_token_accuracy": 0.7883368730545044, "num_tokens": 3045557.0, "step": 80 }, { "epoch": 0.010304032565831319, "ewc_loss": 0.00064849853515625, "ewc_loss_diag": 1.4901161193847656e-07, "ewc_loss_parallel": 6.332993507385254e-07, "grad_norm": 2.9928200244903564, "learning_rate": 3.391267486222976e-07, "loss": 0.8484, "mean_token_accuracy": 0.7654428482055664, "num_tokens": 3080478.0, "step": 81 }, { "epoch": 0.01043124284442183, "ewc_loss": 0.000637054443359375, "ewc_loss_diag": 1.564621925354004e-07, "ewc_loss_parallel": 6.221234798431396e-07, "grad_norm": 2.494173288345337, "learning_rate": 3.433658329800763e-07, "loss": 0.873, "mean_token_accuracy": 0.7625874876976013, "num_tokens": 3119205.0, "step": 82 }, { "epoch": 0.010558453123012339, "ewc_loss": 0.000652313232421875, "ewc_loss_diag": 1.6391277313232422e-07, "ewc_loss_parallel": 6.370246410369873e-07, "grad_norm": 2.72756028175354, "learning_rate": 3.4760491733785504e-07, "loss": 0.8945, "mean_token_accuracy": 0.7553989887237549, "num_tokens": 3151816.0, "step": 83 }, { "epoch": 0.01068566340160285, "ewc_loss": 0.000762939453125, "ewc_loss_diag": 1.7508864402770996e-07, "ewc_loss_parallel": 7.450580596923828e-07, "grad_norm": 2.6443681716918945, "learning_rate": 3.5184400169563376e-07, "loss": 0.8634, "mean_token_accuracy": 0.7617502212524414, "num_tokens": 3189765.0, "step": 84 }, { "epoch": 0.010812873680193359, "ewc_loss": 0.0009918212890625, "ewc_loss_diag": 1.9185245037078857e-07, "ewc_loss_parallel": 9.685754776000977e-07, "grad_norm": 2.4153692722320557, "learning_rate": 3.560830860534125e-07, "loss": 0.8776, "mean_token_accuracy": 0.7598438858985901, "num_tokens": 3228988.0, "step": 85 }, { "epoch": 0.01094008395878387, "ewc_loss": 0.001251220703125, "ewc_loss_diag": 2.076849341392517e-07, "ewc_loss_parallel": 1.2293457984924316e-06, "grad_norm": 2.490994691848755, "learning_rate": 3.603221704111912e-07, "loss": 0.8852, "mean_token_accuracy": 0.7576684951782227, "num_tokens": 3270648.0, "step": 86 }, { "epoch": 0.01106729423737438, "ewc_loss": 0.0015716552734375, "ewc_loss_diag": 2.3469328880310059e-07, "ewc_loss_parallel": 1.5497207641601562e-06, "grad_norm": 2.1929240226745605, "learning_rate": 3.6456125476896994e-07, "loss": 0.8314, "mean_token_accuracy": 0.7721054553985596, "num_tokens": 3312134.0, "step": 87 }, { "epoch": 0.01119450451596489, "ewc_loss": 0.0018157958984375, "ewc_loss_diag": 2.5704503059387207e-07, "ewc_loss_parallel": 1.7955899238586426e-06, "grad_norm": 2.2320728302001953, "learning_rate": 3.688003391267486e-07, "loss": 0.8381, "mean_token_accuracy": 0.7703381776809692, "num_tokens": 3351904.0, "step": 88 }, { "epoch": 0.0113217147945554, "ewc_loss": 0.002044677734375, "ewc_loss_diag": 2.7008354663848877e-07, "ewc_loss_parallel": 2.0116567611694336e-06, "grad_norm": 1.9995008707046509, "learning_rate": 3.7303942348452734e-07, "loss": 0.7888, "mean_token_accuracy": 0.7815930843353271, "num_tokens": 3389610.0, "step": 89 }, { "epoch": 0.01144892507314591, "ewc_loss": 0.002197265625, "ewc_loss_diag": 2.9616057872772217e-07, "ewc_loss_parallel": 2.16066837310791e-06, "grad_norm": 3.0706374645233154, "learning_rate": 3.7727850784230606e-07, "loss": 0.8621, "mean_token_accuracy": 0.7557845711708069, "num_tokens": 3425779.0, "step": 90 }, { "epoch": 0.01157613535173642, "ewc_loss": 0.0023345947265625, "ewc_loss_diag": 3.1851232051849365e-07, "ewc_loss_parallel": 2.3096799850463867e-06, "grad_norm": 2.029421329498291, "learning_rate": 3.815175922000848e-07, "loss": 0.8373, "mean_token_accuracy": 0.7706642150878906, "num_tokens": 3463938.0, "step": 91 }, { "epoch": 0.01170334563032693, "ewc_loss": 0.0024566650390625, "ewc_loss_diag": 3.46451997756958e-07, "ewc_loss_parallel": 2.428889274597168e-06, "grad_norm": 1.8373366594314575, "learning_rate": 3.857566765578635e-07, "loss": 0.7976, "mean_token_accuracy": 0.7800958156585693, "num_tokens": 3500494.0, "step": 92 }, { "epoch": 0.01183055590891744, "ewc_loss": 0.002410888671875, "ewc_loss_diag": 3.632158041000366e-07, "ewc_loss_parallel": 2.384185791015625e-06, "grad_norm": 1.6896772384643555, "learning_rate": 3.8999576091564224e-07, "loss": 0.7592, "mean_token_accuracy": 0.7859750986099243, "num_tokens": 3540901.0, "step": 93 }, { "epoch": 0.01195776618750795, "ewc_loss": 0.0022735595703125, "ewc_loss_diag": 3.8370490074157715e-07, "ewc_loss_parallel": 2.2351741790771484e-06, "grad_norm": 1.659101128578186, "learning_rate": 3.9423484527342097e-07, "loss": 0.7866, "mean_token_accuracy": 0.7774301767349243, "num_tokens": 3579275.0, "step": 94 }, { "epoch": 0.012084976466098461, "ewc_loss": 0.0022430419921875, "ewc_loss_diag": 4.023313522338867e-07, "ewc_loss_parallel": 2.1904706954956055e-06, "grad_norm": 1.6406285762786865, "learning_rate": 3.984739296311997e-07, "loss": 0.7561, "mean_token_accuracy": 0.7837125062942505, "num_tokens": 3614755.0, "step": 95 }, { "epoch": 0.01221218674468897, "ewc_loss": 0.002197265625, "ewc_loss_diag": 4.302710294723511e-07, "ewc_loss_parallel": 2.1457672119140625e-06, "grad_norm": 1.714302659034729, "learning_rate": 4.027130139889784e-07, "loss": 0.8103, "mean_token_accuracy": 0.7681736350059509, "num_tokens": 3655744.0, "step": 96 }, { "epoch": 0.012339397023279481, "ewc_loss": 0.0021820068359375, "ewc_loss_diag": 4.470348358154297e-07, "ewc_loss_parallel": 2.130866050720215e-06, "grad_norm": 1.644943118095398, "learning_rate": 4.0695209834675715e-07, "loss": 0.7927, "mean_token_accuracy": 0.7800840139389038, "num_tokens": 3697257.0, "step": 97 }, { "epoch": 0.012466607301869992, "ewc_loss": 0.002166748046875, "ewc_loss_diag": 4.6938657760620117e-07, "ewc_loss_parallel": 2.115964889526367e-06, "grad_norm": 1.5761724710464478, "learning_rate": 4.111911827045358e-07, "loss": 0.7447, "mean_token_accuracy": 0.7916526794433594, "num_tokens": 3735737.0, "step": 98 }, { "epoch": 0.012593817580460501, "ewc_loss": 0.0020751953125, "ewc_loss_diag": 4.880130290985107e-07, "ewc_loss_parallel": 2.0265579223632812e-06, "grad_norm": 1.5698784589767456, "learning_rate": 4.1543026706231454e-07, "loss": 0.7582, "mean_token_accuracy": 0.7869675755500793, "num_tokens": 3774212.0, "step": 99 }, { "epoch": 0.012721027859051012, "ewc_loss": 0.0020294189453125, "ewc_loss_diag": 4.954636096954346e-07, "ewc_loss_parallel": 1.9818544387817383e-06, "grad_norm": 1.5943713188171387, "learning_rate": 4.1966935142009327e-07, "loss": 0.7467, "mean_token_accuracy": 0.7877858877182007, "num_tokens": 3807640.0, "step": 100 }, { "epoch": 0.012848238137641521, "ewc_loss": 0.0019989013671875, "ewc_loss_diag": 5.103647708892822e-07, "ewc_loss_parallel": 1.952052116394043e-06, "grad_norm": 1.539740800857544, "learning_rate": 4.23908435777872e-07, "loss": 0.7765, "mean_token_accuracy": 0.7827137112617493, "num_tokens": 3848908.0, "step": 101 }, { "epoch": 0.012975448416232032, "ewc_loss": 0.001922607421875, "ewc_loss_diag": 5.178153514862061e-07, "ewc_loss_parallel": 1.8700957298278809e-06, "grad_norm": 1.6352440118789673, "learning_rate": 4.281475201356507e-07, "loss": 0.7851, "mean_token_accuracy": 0.7757942080497742, "num_tokens": 3886341.0, "step": 102 }, { "epoch": 0.013102658694822541, "ewc_loss": 0.0018768310546875, "ewc_loss_diag": 5.21540641784668e-07, "ewc_loss_parallel": 1.8253922462463379e-06, "grad_norm": 1.4329984188079834, "learning_rate": 4.3238660449342945e-07, "loss": 0.6627, "mean_token_accuracy": 0.8075039386749268, "num_tokens": 3917111.0, "step": 103 }, { "epoch": 0.013229868973413052, "ewc_loss": 0.00177001953125, "ewc_loss_diag": 5.289912223815918e-07, "ewc_loss_parallel": 1.7136335372924805e-06, "grad_norm": 1.4135088920593262, "learning_rate": 4.3662568885120817e-07, "loss": 0.7452, "mean_token_accuracy": 0.7874698042869568, "num_tokens": 3959642.0, "step": 104 }, { "epoch": 0.013357079252003561, "ewc_loss": 0.00171661376953125, "ewc_loss_diag": 5.401670932769775e-07, "ewc_loss_parallel": 1.6614794731140137e-06, "grad_norm": 1.4378182888031006, "learning_rate": 4.408647732089869e-07, "loss": 0.7692, "mean_token_accuracy": 0.7822026610374451, "num_tokens": 4001028.0, "step": 105 }, { "epoch": 0.013484289530594072, "ewc_loss": 0.00167083740234375, "ewc_loss_diag": 5.587935447692871e-07, "ewc_loss_parallel": 1.6167759895324707e-06, "grad_norm": 1.501415491104126, "learning_rate": 4.451038575667656e-07, "loss": 0.7795, "mean_token_accuracy": 0.7750545740127563, "num_tokens": 4045249.0, "step": 106 }, { "epoch": 0.013611499809184581, "ewc_loss": 0.00167083740234375, "ewc_loss_diag": 5.662441253662109e-07, "ewc_loss_parallel": 1.6167759895324707e-06, "grad_norm": 1.4097678661346436, "learning_rate": 4.493429419245443e-07, "loss": 0.7817, "mean_token_accuracy": 0.7787725925445557, "num_tokens": 4082369.0, "step": 107 }, { "epoch": 0.013738710087775092, "ewc_loss": 0.00162506103515625, "ewc_loss_diag": 5.811452865600586e-07, "ewc_loss_parallel": 1.564621925354004e-06, "grad_norm": 1.4253381490707397, "learning_rate": 4.53582026282323e-07, "loss": 0.7638, "mean_token_accuracy": 0.7789257764816284, "num_tokens": 4117448.0, "step": 108 }, { "epoch": 0.013865920366365603, "ewc_loss": 0.0015869140625, "ewc_loss_diag": 5.923211574554443e-07, "ewc_loss_parallel": 1.5273690223693848e-06, "grad_norm": 1.3098384141921997, "learning_rate": 4.5782111064010175e-07, "loss": 0.7023, "mean_token_accuracy": 0.8007952570915222, "num_tokens": 4159132.0, "step": 109 }, { "epoch": 0.013993130644956112, "ewc_loss": 0.00152587890625, "ewc_loss_diag": 6.146728992462158e-07, "ewc_loss_parallel": 1.4677643775939941e-06, "grad_norm": 1.325980544090271, "learning_rate": 4.620601949978805e-07, "loss": 0.7355, "mean_token_accuracy": 0.7852635979652405, "num_tokens": 4199800.0, "step": 110 }, { "epoch": 0.014120340923546623, "ewc_loss": 0.00150299072265625, "ewc_loss_diag": 6.258487701416016e-07, "ewc_loss_parallel": 1.4454126358032227e-06, "grad_norm": 1.418068528175354, "learning_rate": 4.662992793556592e-07, "loss": 0.7405, "mean_token_accuracy": 0.7920355796813965, "num_tokens": 4244513.0, "step": 111 }, { "epoch": 0.014247551202137132, "ewc_loss": 0.001556396484375, "ewc_loss_diag": 6.51925802230835e-07, "ewc_loss_parallel": 1.4901161193847656e-06, "grad_norm": 1.3144358396530151, "learning_rate": 4.705383637134379e-07, "loss": 0.6847, "mean_token_accuracy": 0.8018754720687866, "num_tokens": 4282858.0, "step": 112 }, { "epoch": 0.014374761480727643, "ewc_loss": 0.0015716552734375, "ewc_loss_diag": 6.705522537231445e-07, "ewc_loss_parallel": 1.5050172805786133e-06, "grad_norm": 1.2730411291122437, "learning_rate": 4.7477744807121665e-07, "loss": 0.7349, "mean_token_accuracy": 0.7880138158798218, "num_tokens": 4325092.0, "step": 113 }, { "epoch": 0.014501971759318152, "ewc_loss": 0.0015869140625, "ewc_loss_diag": 6.891787052154541e-07, "ewc_loss_parallel": 1.519918441772461e-06, "grad_norm": 1.2337820529937744, "learning_rate": 4.790165324289954e-07, "loss": 0.6677, "mean_token_accuracy": 0.8063330054283142, "num_tokens": 4363459.0, "step": 114 }, { "epoch": 0.014629182037908663, "ewc_loss": 0.0016021728515625, "ewc_loss_diag": 7.115304470062256e-07, "ewc_loss_parallel": 1.5348196029663086e-06, "grad_norm": 1.2245457172393799, "learning_rate": 4.832556167867741e-07, "loss": 0.7415, "mean_token_accuracy": 0.788232147693634, "num_tokens": 4400065.0, "step": 115 }, { "epoch": 0.014756392316499172, "ewc_loss": 0.00164794921875, "ewc_loss_diag": 7.301568984985352e-07, "ewc_loss_parallel": 1.5720725059509277e-06, "grad_norm": 1.2297263145446777, "learning_rate": 4.874947011445528e-07, "loss": 0.7254, "mean_token_accuracy": 0.7899794578552246, "num_tokens": 4437828.0, "step": 116 }, { "epoch": 0.014883602595089683, "ewc_loss": 0.00168609619140625, "ewc_loss_diag": 7.525086402893066e-07, "ewc_loss_parallel": 1.6093254089355469e-06, "grad_norm": 1.2808712720870972, "learning_rate": 4.917337855023316e-07, "loss": 0.6941, "mean_token_accuracy": 0.7970995903015137, "num_tokens": 4475556.0, "step": 117 }, { "epoch": 0.015010812873680193, "ewc_loss": 0.001800537109375, "ewc_loss_diag": 7.82310962677002e-07, "ewc_loss_parallel": 1.7210841178894043e-06, "grad_norm": 1.2527943849563599, "learning_rate": 4.959728698601102e-07, "loss": 0.6963, "mean_token_accuracy": 0.8012926578521729, "num_tokens": 4512087.0, "step": 118 }, { "epoch": 0.015138023152270703, "ewc_loss": 0.0018768310546875, "ewc_loss_diag": 7.972121238708496e-07, "ewc_loss_parallel": 1.8030405044555664e-06, "grad_norm": 1.1700917482376099, "learning_rate": 5.00211954217889e-07, "loss": 0.6663, "mean_token_accuracy": 0.803689181804657, "num_tokens": 4552116.0, "step": 119 }, { "epoch": 0.015265233430861214, "ewc_loss": 0.00194549560546875, "ewc_loss_diag": 8.307397365570068e-07, "ewc_loss_parallel": 1.862645149230957e-06, "grad_norm": 1.2437807321548462, "learning_rate": 5.044510385756677e-07, "loss": 0.7115, "mean_token_accuracy": 0.7949564456939697, "num_tokens": 4591723.0, "step": 120 }, { "epoch": 0.015392443709451724, "ewc_loss": 0.0020294189453125, "ewc_loss_diag": 8.568167686462402e-07, "ewc_loss_parallel": 1.9371509552001953e-06, "grad_norm": 1.1067399978637695, "learning_rate": 5.086901229334465e-07, "loss": 0.6379, "mean_token_accuracy": 0.8153736591339111, "num_tokens": 4635800.0, "step": 121 }, { "epoch": 0.015519653988042234, "ewc_loss": 0.0020599365234375, "ewc_loss_diag": 8.828938007354736e-07, "ewc_loss_parallel": 1.9669532775878906e-06, "grad_norm": 1.122195839881897, "learning_rate": 5.129292072912251e-07, "loss": 0.7049, "mean_token_accuracy": 0.7961283922195435, "num_tokens": 4679908.0, "step": 122 }, { "epoch": 0.015646864266632744, "ewc_loss": 0.002105712890625, "ewc_loss_diag": 9.201467037200928e-07, "ewc_loss_parallel": 2.0116567611694336e-06, "grad_norm": 1.137087345123291, "learning_rate": 5.171682916490039e-07, "loss": 0.6863, "mean_token_accuracy": 0.7964316606521606, "num_tokens": 4720120.0, "step": 123 }, { "epoch": 0.015774074545223254, "ewc_loss": 0.0022125244140625, "ewc_loss_diag": 9.5367431640625e-07, "ewc_loss_parallel": 2.115964889526367e-06, "grad_norm": 1.1441127061843872, "learning_rate": 5.214073760067826e-07, "loss": 0.7089, "mean_token_accuracy": 0.7858878970146179, "num_tokens": 4758777.0, "step": 124 }, { "epoch": 0.015901284823813765, "ewc_loss": 0.0023040771484375, "ewc_loss_diag": 9.760260581970215e-07, "ewc_loss_parallel": 2.205371856689453e-06, "grad_norm": 1.1536668539047241, "learning_rate": 5.256464603645613e-07, "loss": 0.6699, "mean_token_accuracy": 0.8047294616699219, "num_tokens": 4794781.0, "step": 125 }, { "epoch": 0.016028495102404273, "ewc_loss": 0.0023651123046875, "ewc_loss_diag": 1.0207295417785645e-06, "ewc_loss_parallel": 2.2649765014648438e-06, "grad_norm": 1.1075328588485718, "learning_rate": 5.2988554472234e-07, "loss": 0.7173, "mean_token_accuracy": 0.7881619930267334, "num_tokens": 4836391.0, "step": 126 }, { "epoch": 0.016155705380994784, "ewc_loss": 0.0024261474609375, "ewc_loss_diag": 1.043081283569336e-06, "ewc_loss_parallel": 2.3245811462402344e-06, "grad_norm": 1.1445858478546143, "learning_rate": 5.341246290801187e-07, "loss": 0.6892, "mean_token_accuracy": 0.7983329892158508, "num_tokens": 4874193.0, "step": 127 }, { "epoch": 0.016282915659585295, "ewc_loss": 0.0025177001953125, "ewc_loss_diag": 1.0505318641662598e-06, "ewc_loss_parallel": 2.4139881134033203e-06, "grad_norm": 1.1483873128890991, "learning_rate": 5.383637134378975e-07, "loss": 0.6966, "mean_token_accuracy": 0.7920618057250977, "num_tokens": 4909311.0, "step": 128 }, { "epoch": 0.016410125938175806, "ewc_loss": 0.0025787353515625, "ewc_loss_diag": 1.0579824447631836e-06, "ewc_loss_parallel": 2.473592758178711e-06, "grad_norm": 1.123866319656372, "learning_rate": 5.426027977956762e-07, "loss": 0.6866, "mean_token_accuracy": 0.7998124957084656, "num_tokens": 4947409.0, "step": 129 }, { "epoch": 0.016537336216766316, "ewc_loss": 0.0026702880859375, "ewc_loss_diag": 1.0654330253601074e-06, "ewc_loss_parallel": 2.562999725341797e-06, "grad_norm": 1.13132905960083, "learning_rate": 5.468418821534549e-07, "loss": 0.6898, "mean_token_accuracy": 0.7945905923843384, "num_tokens": 4982792.0, "step": 130 }, { "epoch": 0.016664546495356824, "ewc_loss": 0.0027313232421875, "ewc_loss_diag": 1.087784767150879e-06, "ewc_loss_parallel": 2.6226043701171875e-06, "grad_norm": 1.0725339651107788, "learning_rate": 5.510809665112336e-07, "loss": 0.6611, "mean_token_accuracy": 0.802152156829834, "num_tokens": 5020795.0, "step": 131 }, { "epoch": 0.016791756773947335, "ewc_loss": 0.00274658203125, "ewc_loss_diag": 1.0952353477478027e-06, "ewc_loss_parallel": 2.637505531311035e-06, "grad_norm": 1.0956594944000244, "learning_rate": 5.553200508690124e-07, "loss": 0.6992, "mean_token_accuracy": 0.7929364442825317, "num_tokens": 5061519.0, "step": 132 }, { "epoch": 0.016918967052537846, "ewc_loss": 0.0027923583984375, "ewc_loss_diag": 1.125037670135498e-06, "ewc_loss_parallel": 2.682209014892578e-06, "grad_norm": 1.129529356956482, "learning_rate": 5.595591352267911e-07, "loss": 0.6948, "mean_token_accuracy": 0.7923251986503601, "num_tokens": 5094608.0, "step": 133 }, { "epoch": 0.017046177331128357, "ewc_loss": 0.002838134765625, "ewc_loss_diag": 1.1399388313293457e-06, "ewc_loss_parallel": 2.726912498474121e-06, "grad_norm": 1.1073336601257324, "learning_rate": 5.637982195845697e-07, "loss": 0.6956, "mean_token_accuracy": 0.7964053153991699, "num_tokens": 5131130.0, "step": 134 }, { "epoch": 0.017173387609718864, "ewc_loss": 0.00286865234375, "ewc_loss_diag": 1.1771917343139648e-06, "ewc_loss_parallel": 2.7418136596679688e-06, "grad_norm": 1.0832231044769287, "learning_rate": 5.680373039423485e-07, "loss": 0.6784, "mean_token_accuracy": 0.8011640906333923, "num_tokens": 5168921.0, "step": 135 }, { "epoch": 0.017300597888309375, "ewc_loss": 0.0028533935546875, "ewc_loss_diag": 1.2069940567016602e-06, "ewc_loss_parallel": 2.726912498474121e-06, "grad_norm": 1.0940850973129272, "learning_rate": 5.722763883001272e-07, "loss": 0.6757, "mean_token_accuracy": 0.7980347871780396, "num_tokens": 5206417.0, "step": 136 }, { "epoch": 0.017427808166899886, "ewc_loss": 0.002838134765625, "ewc_loss_diag": 1.259148120880127e-06, "ewc_loss_parallel": 2.7120113372802734e-06, "grad_norm": 1.123774528503418, "learning_rate": 5.76515472657906e-07, "loss": 0.7197, "mean_token_accuracy": 0.7873349785804749, "num_tokens": 5239984.0, "step": 137 }, { "epoch": 0.017555018445490397, "ewc_loss": 0.0029296875, "ewc_loss_diag": 1.30385160446167e-06, "ewc_loss_parallel": 2.7865171432495117e-06, "grad_norm": 1.085340976715088, "learning_rate": 5.807545570156846e-07, "loss": 0.6926, "mean_token_accuracy": 0.7971215844154358, "num_tokens": 5276444.0, "step": 138 }, { "epoch": 0.017682228724080904, "ewc_loss": 0.0029144287109375, "ewc_loss_diag": 1.3336539268493652e-06, "ewc_loss_parallel": 2.771615982055664e-06, "grad_norm": 1.0474369525909424, "learning_rate": 5.849936413734634e-07, "loss": 0.622, "mean_token_accuracy": 0.8159076571464539, "num_tokens": 5312064.0, "step": 139 }, { "epoch": 0.017809439002671415, "ewc_loss": 0.00286865234375, "ewc_loss_diag": 1.3634562492370605e-06, "ewc_loss_parallel": 2.726912498474121e-06, "grad_norm": 1.125514030456543, "learning_rate": 5.892327257312421e-07, "loss": 0.7098, "mean_token_accuracy": 0.7953469753265381, "num_tokens": 5350405.0, "step": 140 }, { "epoch": 0.017936649281261926, "ewc_loss": 0.0029449462890625, "ewc_loss_diag": 1.3783574104309082e-06, "ewc_loss_parallel": 2.8014183044433594e-06, "grad_norm": 1.0675404071807861, "learning_rate": 5.934718100890209e-07, "loss": 0.6595, "mean_token_accuracy": 0.8062784075737, "num_tokens": 5388338.0, "step": 141 }, { "epoch": 0.018063859559852437, "ewc_loss": 0.002899169921875, "ewc_loss_diag": 1.3932585716247559e-06, "ewc_loss_parallel": 2.7567148208618164e-06, "grad_norm": 1.0697739124298096, "learning_rate": 5.977108944467995e-07, "loss": 0.6622, "mean_token_accuracy": 0.8045392632484436, "num_tokens": 5432228.0, "step": 142 }, { "epoch": 0.018191069838442948, "ewc_loss": 0.0028839111328125, "ewc_loss_diag": 1.4081597328186035e-06, "ewc_loss_parallel": 2.7418136596679688e-06, "grad_norm": 1.0329945087432861, "learning_rate": 6.019499788045782e-07, "loss": 0.6409, "mean_token_accuracy": 0.8103231191635132, "num_tokens": 5469969.0, "step": 143 }, { "epoch": 0.018318280117033455, "ewc_loss": 0.002838134765625, "ewc_loss_diag": 1.4379620552062988e-06, "ewc_loss_parallel": 2.6971101760864258e-06, "grad_norm": 1.0111238956451416, "learning_rate": 6.06189063162357e-07, "loss": 0.6341, "mean_token_accuracy": 0.8084633350372314, "num_tokens": 5506262.0, "step": 144 }, { "epoch": 0.018445490395623966, "ewc_loss": 0.002777099609375, "ewc_loss_diag": 1.4528632164001465e-06, "ewc_loss_parallel": 2.6226043701171875e-06, "grad_norm": 1.059492826461792, "learning_rate": 6.104281475201357e-07, "loss": 0.6539, "mean_token_accuracy": 0.8024689555168152, "num_tokens": 5542870.0, "step": 145 }, { "epoch": 0.018572700674214477, "ewc_loss": 0.002838134765625, "ewc_loss_diag": 1.475214958190918e-06, "ewc_loss_parallel": 2.682209014892578e-06, "grad_norm": 1.0195198059082031, "learning_rate": 6.146672318779144e-07, "loss": 0.6658, "mean_token_accuracy": 0.7996755242347717, "num_tokens": 5579958.0, "step": 146 }, { "epoch": 0.018699910952804988, "ewc_loss": 0.002777099609375, "ewc_loss_diag": 1.5050172805786133e-06, "ewc_loss_parallel": 2.6226043701171875e-06, "grad_norm": 1.031572937965393, "learning_rate": 6.189063162356931e-07, "loss": 0.6791, "mean_token_accuracy": 0.7955799102783203, "num_tokens": 5623306.0, "step": 147 }, { "epoch": 0.018827121231395495, "ewc_loss": 0.002777099609375, "ewc_loss_diag": 1.519918441772461e-06, "ewc_loss_parallel": 2.6226043701171875e-06, "grad_norm": 1.0056041479110718, "learning_rate": 6.231454005934719e-07, "loss": 0.6072, "mean_token_accuracy": 0.8193692564964294, "num_tokens": 5660801.0, "step": 148 }, { "epoch": 0.018954331509986006, "ewc_loss": 0.0027313232421875, "ewc_loss_diag": 1.5497207641601562e-06, "ewc_loss_parallel": 2.5779008865356445e-06, "grad_norm": 1.0367779731750488, "learning_rate": 6.273844849512506e-07, "loss": 0.6399, "mean_token_accuracy": 0.8057074546813965, "num_tokens": 5696625.0, "step": 149 }, { "epoch": 0.019081541788576517, "ewc_loss": 0.0027618408203125, "ewc_loss_diag": 1.564621925354004e-06, "ewc_loss_parallel": 2.60770320892334e-06, "grad_norm": 1.0004078149795532, "learning_rate": 6.316235693090293e-07, "loss": 0.658, "mean_token_accuracy": 0.8023650050163269, "num_tokens": 5733491.0, "step": 150 }, { "epoch": 0.019208752067167028, "ewc_loss": 0.00274658203125, "ewc_loss_diag": 1.5944242477416992e-06, "ewc_loss_parallel": 2.592802047729492e-06, "grad_norm": 1.0125768184661865, "learning_rate": 6.35862653666808e-07, "loss": 0.6767, "mean_token_accuracy": 0.7954878807067871, "num_tokens": 5773550.0, "step": 151 }, { "epoch": 0.01933596234575754, "ewc_loss": 0.00274658203125, "ewc_loss_diag": 1.6167759895324707e-06, "ewc_loss_parallel": 2.5779008865356445e-06, "grad_norm": 1.0436125993728638, "learning_rate": 6.401017380245868e-07, "loss": 0.6704, "mean_token_accuracy": 0.7986307740211487, "num_tokens": 5813114.0, "step": 152 }, { "epoch": 0.019463172624348046, "ewc_loss": 0.002777099609375, "ewc_loss_diag": 1.6391277313232422e-06, "ewc_loss_parallel": 2.60770320892334e-06, "grad_norm": 1.0306360721588135, "learning_rate": 6.443408223823655e-07, "loss": 0.6996, "mean_token_accuracy": 0.7918697595596313, "num_tokens": 5852702.0, "step": 153 }, { "epoch": 0.019590382902938557, "ewc_loss": 0.00274658203125, "ewc_loss_diag": 1.6689300537109375e-06, "ewc_loss_parallel": 2.5779008865356445e-06, "grad_norm": 1.0352603197097778, "learning_rate": 6.485799067401442e-07, "loss": 0.767, "mean_token_accuracy": 0.7719191312789917, "num_tokens": 5894481.0, "step": 154 }, { "epoch": 0.019717593181529068, "ewc_loss": 0.002716064453125, "ewc_loss_diag": 1.6838312149047852e-06, "ewc_loss_parallel": 2.5480985641479492e-06, "grad_norm": 1.0037333965301514, "learning_rate": 6.528189910979229e-07, "loss": 0.6235, "mean_token_accuracy": 0.8071836233139038, "num_tokens": 5929501.0, "step": 155 }, { "epoch": 0.01984480346011958, "ewc_loss": 0.0027008056640625, "ewc_loss_diag": 1.7210841178894043e-06, "ewc_loss_parallel": 2.5331974029541016e-06, "grad_norm": 1.1164644956588745, "learning_rate": 6.570580754557017e-07, "loss": 0.6226, "mean_token_accuracy": 0.807403564453125, "num_tokens": 5966407.0, "step": 156 }, { "epoch": 0.019972013738710086, "ewc_loss": 0.0028076171875, "ewc_loss_diag": 1.7508864402770996e-06, "ewc_loss_parallel": 2.6226043701171875e-06, "grad_norm": 0.9982089996337891, "learning_rate": 6.612971598134803e-07, "loss": 0.5793, "mean_token_accuracy": 0.8232318162918091, "num_tokens": 6004059.0, "step": 157 }, { "epoch": 0.020099224017300597, "ewc_loss": 0.0027008056640625, "ewc_loss_diag": 1.773238182067871e-06, "ewc_loss_parallel": 2.518296241760254e-06, "grad_norm": 0.9858206510543823, "learning_rate": 6.65536244171259e-07, "loss": 0.6344, "mean_token_accuracy": 0.8062832355499268, "num_tokens": 6040516.0, "step": 158 }, { "epoch": 0.020226434295891108, "ewc_loss": 0.002685546875, "ewc_loss_diag": 1.780688762664795e-06, "ewc_loss_parallel": 2.5033950805664062e-06, "grad_norm": 1.012166142463684, "learning_rate": 6.697753285290377e-07, "loss": 0.7055, "mean_token_accuracy": 0.7867447137832642, "num_tokens": 6077801.0, "step": 159 }, { "epoch": 0.02035364457448162, "ewc_loss": 0.0027008056640625, "ewc_loss_diag": 1.8104910850524902e-06, "ewc_loss_parallel": 2.518296241760254e-06, "grad_norm": 0.9561834335327148, "learning_rate": 6.740144128868165e-07, "loss": 0.6229, "mean_token_accuracy": 0.8102702498435974, "num_tokens": 6114102.0, "step": 160 }, { "epoch": 0.020480854853072127, "ewc_loss": 0.002655029296875, "ewc_loss_diag": 1.8253922462463379e-06, "ewc_loss_parallel": 2.473592758178711e-06, "grad_norm": 0.9544920921325684, "learning_rate": 6.782534972445952e-07, "loss": 0.644, "mean_token_accuracy": 0.8048382997512817, "num_tokens": 6157342.0, "step": 161 }, { "epoch": 0.020608065131662637, "ewc_loss": 0.0026397705078125, "ewc_loss_diag": 1.8775463104248047e-06, "ewc_loss_parallel": 2.4586915969848633e-06, "grad_norm": 1.0172075033187866, "learning_rate": 6.82492581602374e-07, "loss": 0.6454, "mean_token_accuracy": 0.8050585985183716, "num_tokens": 6193058.0, "step": 162 }, { "epoch": 0.02073527541025315, "ewc_loss": 0.002716064453125, "ewc_loss_diag": 1.9073486328125e-06, "ewc_loss_parallel": 2.518296241760254e-06, "grad_norm": 1.044508934020996, "learning_rate": 6.867316659601526e-07, "loss": 0.6386, "mean_token_accuracy": 0.8099280595779419, "num_tokens": 6223616.0, "step": 163 }, { "epoch": 0.02086248568884366, "ewc_loss": 0.00274658203125, "ewc_loss_diag": 1.9222497940063477e-06, "ewc_loss_parallel": 2.5480985641479492e-06, "grad_norm": 0.9599901437759399, "learning_rate": 6.909707503179314e-07, "loss": 0.6385, "mean_token_accuracy": 0.8094019293785095, "num_tokens": 6264920.0, "step": 164 }, { "epoch": 0.02098969596743417, "ewc_loss": 0.0027008056640625, "ewc_loss_diag": 1.9371509552001953e-06, "ewc_loss_parallel": 2.5033950805664062e-06, "grad_norm": 0.970770001411438, "learning_rate": 6.952098346757101e-07, "loss": 0.6797, "mean_token_accuracy": 0.7941124439239502, "num_tokens": 6305270.0, "step": 165 }, { "epoch": 0.021116906246024678, "ewc_loss": 0.002685546875, "ewc_loss_diag": 1.9669532775878906e-06, "ewc_loss_parallel": 2.4884939193725586e-06, "grad_norm": 0.9285233616828918, "learning_rate": 6.994489190334889e-07, "loss": 0.5842, "mean_token_accuracy": 0.8209428787231445, "num_tokens": 6341945.0, "step": 166 }, { "epoch": 0.02124411652461519, "ewc_loss": 0.0026397705078125, "ewc_loss_diag": 1.996755599975586e-06, "ewc_loss_parallel": 2.4437904357910156e-06, "grad_norm": 0.933483362197876, "learning_rate": 7.036880033912675e-07, "loss": 0.5847, "mean_token_accuracy": 0.8230510950088501, "num_tokens": 6381565.0, "step": 167 }, { "epoch": 0.0213713268032057, "ewc_loss": 0.002655029296875, "ewc_loss_diag": 2.0265579223632812e-06, "ewc_loss_parallel": 2.4586915969848633e-06, "grad_norm": 0.9542022943496704, "learning_rate": 7.079270877490463e-07, "loss": 0.6331, "mean_token_accuracy": 0.8079312443733215, "num_tokens": 6419771.0, "step": 168 }, { "epoch": 0.02149853708179621, "ewc_loss": 0.002685546875, "ewc_loss_diag": 2.0563602447509766e-06, "ewc_loss_parallel": 2.4884939193725586e-06, "grad_norm": 0.9787636995315552, "learning_rate": 7.12166172106825e-07, "loss": 0.5912, "mean_token_accuracy": 0.8184528946876526, "num_tokens": 6451413.0, "step": 169 }, { "epoch": 0.021625747360386718, "ewc_loss": 0.002716064453125, "ewc_loss_diag": 2.0563602447509766e-06, "ewc_loss_parallel": 2.5033950805664062e-06, "grad_norm": 0.9378679990768433, "learning_rate": 7.164052564646038e-07, "loss": 0.5965, "mean_token_accuracy": 0.8185935020446777, "num_tokens": 6488883.0, "step": 170 }, { "epoch": 0.02175295763897723, "ewc_loss": 0.0026702880859375, "ewc_loss_diag": 2.0712614059448242e-06, "ewc_loss_parallel": 2.4586915969848633e-06, "grad_norm": 1.0034680366516113, "learning_rate": 7.206443408223824e-07, "loss": 0.6471, "mean_token_accuracy": 0.8036888837814331, "num_tokens": 6523669.0, "step": 171 }, { "epoch": 0.02188016791756774, "ewc_loss": 0.002716064453125, "ewc_loss_diag": 2.086162567138672e-06, "ewc_loss_parallel": 2.5033950805664062e-06, "grad_norm": 0.9443991184234619, "learning_rate": 7.248834251801612e-07, "loss": 0.6094, "mean_token_accuracy": 0.8122594356536865, "num_tokens": 6561795.0, "step": 172 }, { "epoch": 0.02200737819615825, "ewc_loss": 0.0027008056640625, "ewc_loss_diag": 2.1010637283325195e-06, "ewc_loss_parallel": 2.4884939193725586e-06, "grad_norm": 0.951756477355957, "learning_rate": 7.291225095379399e-07, "loss": 0.5948, "mean_token_accuracy": 0.8183512687683105, "num_tokens": 6604102.0, "step": 173 }, { "epoch": 0.02213458847474876, "ewc_loss": 0.002685546875, "ewc_loss_diag": 2.1010637283325195e-06, "ewc_loss_parallel": 2.473592758178711e-06, "grad_norm": 0.9541760087013245, "learning_rate": 7.333615938957187e-07, "loss": 0.6792, "mean_token_accuracy": 0.7965640425682068, "num_tokens": 6645412.0, "step": 174 }, { "epoch": 0.02226179875333927, "ewc_loss": 0.002685546875, "ewc_loss_diag": 2.115964889526367e-06, "ewc_loss_parallel": 2.473592758178711e-06, "grad_norm": 0.9800793528556824, "learning_rate": 7.376006782534972e-07, "loss": 0.5815, "mean_token_accuracy": 0.822920024394989, "num_tokens": 6678407.0, "step": 175 }, { "epoch": 0.02238900903192978, "ewc_loss": 0.0027313232421875, "ewc_loss_diag": 2.115964889526367e-06, "ewc_loss_parallel": 2.518296241760254e-06, "grad_norm": 0.9286218881607056, "learning_rate": 7.41839762611276e-07, "loss": 0.6147, "mean_token_accuracy": 0.8108910322189331, "num_tokens": 6720968.0, "step": 176 }, { "epoch": 0.02251621931052029, "ewc_loss": 0.0026702880859375, "ewc_loss_diag": 2.115964889526367e-06, "ewc_loss_parallel": 2.4586915969848633e-06, "grad_norm": 0.9648327231407166, "learning_rate": 7.460788469690547e-07, "loss": 0.6181, "mean_token_accuracy": 0.8144361972808838, "num_tokens": 6757286.0, "step": 177 }, { "epoch": 0.0226434295891108, "ewc_loss": 0.0027008056640625, "ewc_loss_diag": 2.1457672119140625e-06, "ewc_loss_parallel": 2.4884939193725586e-06, "grad_norm": 1.016223430633545, "learning_rate": 7.503179313268335e-07, "loss": 0.6507, "mean_token_accuracy": 0.8009672164916992, "num_tokens": 6790260.0, "step": 178 }, { "epoch": 0.02277063986770131, "ewc_loss": 0.002777099609375, "ewc_loss_diag": 2.175569534301758e-06, "ewc_loss_parallel": 2.562999725341797e-06, "grad_norm": 0.940656304359436, "learning_rate": 7.545570156846121e-07, "loss": 0.5938, "mean_token_accuracy": 0.8174359202384949, "num_tokens": 6828882.0, "step": 179 }, { "epoch": 0.02289785014629182, "ewc_loss": 0.002716064453125, "ewc_loss_diag": 2.175569534301758e-06, "ewc_loss_parallel": 2.5033950805664062e-06, "grad_norm": 0.9370916485786438, "learning_rate": 7.587961000423909e-07, "loss": 0.6091, "mean_token_accuracy": 0.8168485760688782, "num_tokens": 6867332.0, "step": 180 }, { "epoch": 0.02302506042488233, "ewc_loss": 0.0027008056640625, "ewc_loss_diag": 2.2202730178833008e-06, "ewc_loss_parallel": 2.473592758178711e-06, "grad_norm": 0.9380279183387756, "learning_rate": 7.630351844001696e-07, "loss": 0.7218, "mean_token_accuracy": 0.7806220054626465, "num_tokens": 6914101.0, "step": 181 }, { "epoch": 0.02315227070347284, "ewc_loss": 0.002716064453125, "ewc_loss_diag": 2.2351741790771484e-06, "ewc_loss_parallel": 2.4884939193725586e-06, "grad_norm": 0.9349440932273865, "learning_rate": 7.672742687579484e-07, "loss": 0.6163, "mean_token_accuracy": 0.8114298582077026, "num_tokens": 6952981.0, "step": 182 }, { "epoch": 0.02327948098206335, "ewc_loss": 0.00274658203125, "ewc_loss_diag": 2.2351741790771484e-06, "ewc_loss_parallel": 2.518296241760254e-06, "grad_norm": 0.960891604423523, "learning_rate": 7.71513353115727e-07, "loss": 0.5939, "mean_token_accuracy": 0.8163824677467346, "num_tokens": 6987534.0, "step": 183 }, { "epoch": 0.02340669126065386, "ewc_loss": 0.002777099609375, "ewc_loss_diag": 2.250075340270996e-06, "ewc_loss_parallel": 2.5480985641479492e-06, "grad_norm": 0.9388577342033386, "learning_rate": 7.757524374735058e-07, "loss": 0.6678, "mean_token_accuracy": 0.7966272830963135, "num_tokens": 7029343.0, "step": 184 }, { "epoch": 0.02353390153924437, "ewc_loss": 0.002777099609375, "ewc_loss_diag": 2.2798776626586914e-06, "ewc_loss_parallel": 2.5480985641479492e-06, "grad_norm": 0.9291725754737854, "learning_rate": 7.799915218312845e-07, "loss": 0.6056, "mean_token_accuracy": 0.8142961263656616, "num_tokens": 7067558.0, "step": 185 }, { "epoch": 0.02366111181783488, "ewc_loss": 0.00274658203125, "ewc_loss_diag": 2.3096799850463867e-06, "ewc_loss_parallel": 2.518296241760254e-06, "grad_norm": 0.9198174476623535, "learning_rate": 7.842306061890633e-07, "loss": 0.584, "mean_token_accuracy": 0.8154785633087158, "num_tokens": 7106164.0, "step": 186 }, { "epoch": 0.023788322096425393, "ewc_loss": 0.00274658203125, "ewc_loss_diag": 2.3096799850463867e-06, "ewc_loss_parallel": 2.518296241760254e-06, "grad_norm": 0.9682469964027405, "learning_rate": 7.884696905468419e-07, "loss": 0.639, "mean_token_accuracy": 0.8057491779327393, "num_tokens": 7145549.0, "step": 187 }, { "epoch": 0.0239155323750159, "ewc_loss": 0.0028076171875, "ewc_loss_diag": 2.3245811462402344e-06, "ewc_loss_parallel": 2.5779008865356445e-06, "grad_norm": 0.9863500595092773, "learning_rate": 7.927087749046207e-07, "loss": 0.6046, "mean_token_accuracy": 0.8154362440109253, "num_tokens": 7179697.0, "step": 188 }, { "epoch": 0.02404274265360641, "ewc_loss": 0.0028839111328125, "ewc_loss_diag": 2.3543834686279297e-06, "ewc_loss_parallel": 2.652406692504883e-06, "grad_norm": 0.9452886581420898, "learning_rate": 7.969478592623994e-07, "loss": 0.6799, "mean_token_accuracy": 0.7915682792663574, "num_tokens": 7220103.0, "step": 189 }, { "epoch": 0.024169952932196922, "ewc_loss": 0.0028533935546875, "ewc_loss_diag": 2.3990869522094727e-06, "ewc_loss_parallel": 2.60770320892334e-06, "grad_norm": 0.9080803394317627, "learning_rate": 8.011869436201782e-07, "loss": 0.5374, "mean_token_accuracy": 0.8360269069671631, "num_tokens": 7258059.0, "step": 190 }, { "epoch": 0.024297163210787433, "ewc_loss": 0.0028076171875, "ewc_loss_diag": 2.4139881134033203e-06, "ewc_loss_parallel": 2.562999725341797e-06, "grad_norm": 0.952570378780365, "learning_rate": 8.054260279779568e-07, "loss": 0.6492, "mean_token_accuracy": 0.8054138422012329, "num_tokens": 7295111.0, "step": 191 }, { "epoch": 0.02442437348937794, "ewc_loss": 0.0028839111328125, "ewc_loss_diag": 2.4586915969848633e-06, "ewc_loss_parallel": 2.637505531311035e-06, "grad_norm": 0.9686765074729919, "learning_rate": 8.096651123357356e-07, "loss": 0.6512, "mean_token_accuracy": 0.7998627424240112, "num_tokens": 7332703.0, "step": 192 }, { "epoch": 0.02455158376796845, "ewc_loss": 0.002899169921875, "ewc_loss_diag": 2.473592758178711e-06, "ewc_loss_parallel": 2.652406692504883e-06, "grad_norm": 0.9321699142456055, "learning_rate": 8.139041966935143e-07, "loss": 0.6543, "mean_token_accuracy": 0.7994179725646973, "num_tokens": 7376446.0, "step": 193 }, { "epoch": 0.024678794046558962, "ewc_loss": 0.002838134765625, "ewc_loss_diag": 2.4884939193725586e-06, "ewc_loss_parallel": 2.592802047729492e-06, "grad_norm": 0.9248453378677368, "learning_rate": 8.18143281051293e-07, "loss": 0.5965, "mean_token_accuracy": 0.8163964152336121, "num_tokens": 7416378.0, "step": 194 }, { "epoch": 0.024806004325149473, "ewc_loss": 0.002899169921875, "ewc_loss_diag": 2.518296241760254e-06, "ewc_loss_parallel": 2.652406692504883e-06, "grad_norm": 0.9107208251953125, "learning_rate": 8.223823654090716e-07, "loss": 0.6063, "mean_token_accuracy": 0.8161478042602539, "num_tokens": 7458178.0, "step": 195 }, { "epoch": 0.024933214603739984, "ewc_loss": 0.002899169921875, "ewc_loss_diag": 2.518296241760254e-06, "ewc_loss_parallel": 2.637505531311035e-06, "grad_norm": 0.8732800483703613, "learning_rate": 8.266214497668504e-07, "loss": 0.551, "mean_token_accuracy": 0.8315244913101196, "num_tokens": 7502436.0, "step": 196 }, { "epoch": 0.02506042488233049, "ewc_loss": 0.002838134765625, "ewc_loss_diag": 2.518296241760254e-06, "ewc_loss_parallel": 2.5779008865356445e-06, "grad_norm": 0.9160267114639282, "learning_rate": 8.308605341246291e-07, "loss": 0.609, "mean_token_accuracy": 0.8124930262565613, "num_tokens": 7539625.0, "step": 197 }, { "epoch": 0.025187635160921002, "ewc_loss": 0.0029296875, "ewc_loss_diag": 2.5480985641479492e-06, "ewc_loss_parallel": 2.6673078536987305e-06, "grad_norm": 0.9532270431518555, "learning_rate": 8.350996184824079e-07, "loss": 0.5853, "mean_token_accuracy": 0.8181537389755249, "num_tokens": 7570904.0, "step": 198 }, { "epoch": 0.025314845439511513, "ewc_loss": 0.0029754638671875, "ewc_loss_diag": 2.592802047729492e-06, "ewc_loss_parallel": 2.7120113372802734e-06, "grad_norm": 0.9145262241363525, "learning_rate": 8.393387028401865e-07, "loss": 0.5802, "mean_token_accuracy": 0.8217190504074097, "num_tokens": 7608652.0, "step": 199 }, { "epoch": 0.025442055718102024, "ewc_loss": 0.0029296875, "ewc_loss_diag": 2.60770320892334e-06, "ewc_loss_parallel": 2.6673078536987305e-06, "grad_norm": 0.9804658889770508, "learning_rate": 8.435777871979653e-07, "loss": 0.6536, "mean_token_accuracy": 0.8028731346130371, "num_tokens": 7647759.0, "step": 200 }, { "epoch": 0.02556926599669253, "ewc_loss": 0.0030059814453125, "ewc_loss_diag": 2.637505531311035e-06, "ewc_loss_parallel": 2.7418136596679688e-06, "grad_norm": 0.9997771382331848, "learning_rate": 8.47816871555744e-07, "loss": 0.5743, "mean_token_accuracy": 0.8195672035217285, "num_tokens": 7678918.0, "step": 201 }, { "epoch": 0.025696476275283042, "ewc_loss": 0.0030364990234375, "ewc_loss_diag": 2.637505531311035e-06, "ewc_loss_parallel": 2.771615982055664e-06, "grad_norm": 0.9234350323677063, "learning_rate": 8.520559559135228e-07, "loss": 0.5511, "mean_token_accuracy": 0.8283542394638062, "num_tokens": 7715797.0, "step": 202 }, { "epoch": 0.025823686553873553, "ewc_loss": 0.002960205078125, "ewc_loss_diag": 2.637505531311035e-06, "ewc_loss_parallel": 2.6971101760864258e-06, "grad_norm": 0.906855583190918, "learning_rate": 8.562950402713014e-07, "loss": 0.6554, "mean_token_accuracy": 0.8026520013809204, "num_tokens": 7763053.0, "step": 203 }, { "epoch": 0.025950896832464064, "ewc_loss": 0.00299072265625, "ewc_loss_diag": 2.682209014892578e-06, "ewc_loss_parallel": 2.7120113372802734e-06, "grad_norm": 1.014839768409729, "learning_rate": 8.605341246290802e-07, "loss": 0.6214, "mean_token_accuracy": 0.8130956888198853, "num_tokens": 7793956.0, "step": 204 }, { "epoch": 0.026078107111054575, "ewc_loss": 0.0030975341796875, "ewc_loss_diag": 2.7120113372802734e-06, "ewc_loss_parallel": 2.816319465637207e-06, "grad_norm": 0.9447134137153625, "learning_rate": 8.647732089868589e-07, "loss": 0.5998, "mean_token_accuracy": 0.814918577671051, "num_tokens": 7833997.0, "step": 205 }, { "epoch": 0.026205317389645082, "ewc_loss": 0.00299072265625, "ewc_loss_diag": 2.771615982055664e-06, "ewc_loss_parallel": 2.7120113372802734e-06, "grad_norm": 0.9368460178375244, "learning_rate": 8.690122933446377e-07, "loss": 0.6452, "mean_token_accuracy": 0.8055898547172546, "num_tokens": 7877711.0, "step": 206 }, { "epoch": 0.026332527668235593, "ewc_loss": 0.003021240234375, "ewc_loss_diag": 2.816319465637207e-06, "ewc_loss_parallel": 2.726912498474121e-06, "grad_norm": 0.9586076736450195, "learning_rate": 8.732513777024163e-07, "loss": 0.5679, "mean_token_accuracy": 0.8281447291374207, "num_tokens": 7918136.0, "step": 207 }, { "epoch": 0.026459737946826104, "ewc_loss": 0.003082275390625, "ewc_loss_diag": 2.8312206268310547e-06, "ewc_loss_parallel": 2.7865171432495117e-06, "grad_norm": 0.9201530814170837, "learning_rate": 8.774904620601951e-07, "loss": 0.6094, "mean_token_accuracy": 0.8153109550476074, "num_tokens": 7960847.0, "step": 208 }, { "epoch": 0.026586948225416615, "ewc_loss": 0.003021240234375, "ewc_loss_diag": 2.8312206268310547e-06, "ewc_loss_parallel": 2.726912498474121e-06, "grad_norm": 0.9368677735328674, "learning_rate": 8.817295464179738e-07, "loss": 0.5713, "mean_token_accuracy": 0.8199360370635986, "num_tokens": 7996567.0, "step": 209 }, { "epoch": 0.026714158504007122, "ewc_loss": 0.0030517578125, "ewc_loss_diag": 2.86102294921875e-06, "ewc_loss_parallel": 2.7567148208618164e-06, "grad_norm": 0.9540616869926453, "learning_rate": 8.859686307757526e-07, "loss": 0.5953, "mean_token_accuracy": 0.8167740106582642, "num_tokens": 8032441.0, "step": 210 }, { "epoch": 0.026841368782597633, "ewc_loss": 0.003082275390625, "ewc_loss_diag": 2.8759241104125977e-06, "ewc_loss_parallel": 2.7865171432495117e-06, "grad_norm": 0.9536646604537964, "learning_rate": 8.902077151335312e-07, "loss": 0.6278, "mean_token_accuracy": 0.8076046705245972, "num_tokens": 8072020.0, "step": 211 }, { "epoch": 0.026968579061188144, "ewc_loss": 0.0030670166015625, "ewc_loss_diag": 2.8759241104125977e-06, "ewc_loss_parallel": 2.771615982055664e-06, "grad_norm": 1.01482093334198, "learning_rate": 8.944467994913099e-07, "loss": 0.6732, "mean_token_accuracy": 0.7972130179405212, "num_tokens": 8105736.0, "step": 212 }, { "epoch": 0.027095789339778655, "ewc_loss": 0.003143310546875, "ewc_loss_diag": 2.905726432800293e-06, "ewc_loss_parallel": 2.8461217880249023e-06, "grad_norm": 0.9767781496047974, "learning_rate": 8.986858838490886e-07, "loss": 0.5248, "mean_token_accuracy": 0.833493709564209, "num_tokens": 8139021.0, "step": 213 }, { "epoch": 0.027222999618369163, "ewc_loss": 0.0031280517578125, "ewc_loss_diag": 2.9355287551879883e-06, "ewc_loss_parallel": 2.8312206268310547e-06, "grad_norm": 0.9121969938278198, "learning_rate": 9.029249682068674e-07, "loss": 0.5965, "mean_token_accuracy": 0.8145740032196045, "num_tokens": 8182361.0, "step": 214 }, { "epoch": 0.027350209896959674, "ewc_loss": 0.0030517578125, "ewc_loss_diag": 2.9802322387695312e-06, "ewc_loss_parallel": 2.7418136596679688e-06, "grad_norm": 1.0633560419082642, "learning_rate": 9.07164052564646e-07, "loss": 0.6108, "mean_token_accuracy": 0.8117790818214417, "num_tokens": 8218131.0, "step": 215 }, { "epoch": 0.027477420175550184, "ewc_loss": 0.0031890869140625, "ewc_loss_diag": 2.995133399963379e-06, "ewc_loss_parallel": 2.8908252716064453e-06, "grad_norm": 0.9659335613250732, "learning_rate": 9.114031369224248e-07, "loss": 0.5876, "mean_token_accuracy": 0.8228256702423096, "num_tokens": 8255886.0, "step": 216 }, { "epoch": 0.027604630454140695, "ewc_loss": 0.0031585693359375, "ewc_loss_diag": 3.0249357223510742e-06, "ewc_loss_parallel": 2.8461217880249023e-06, "grad_norm": 0.8998238444328308, "learning_rate": 9.156422212802035e-07, "loss": 0.5441, "mean_token_accuracy": 0.8293901085853577, "num_tokens": 8297196.0, "step": 217 }, { "epoch": 0.027731840732731206, "ewc_loss": 0.003082275390625, "ewc_loss_diag": 3.0547380447387695e-06, "ewc_loss_parallel": 2.771615982055664e-06, "grad_norm": 0.952968180179596, "learning_rate": 9.198813056379823e-07, "loss": 0.5855, "mean_token_accuracy": 0.8189936876296997, "num_tokens": 8331857.0, "step": 218 }, { "epoch": 0.027859051011321714, "ewc_loss": 0.0031585693359375, "ewc_loss_diag": 3.0547380447387695e-06, "ewc_loss_parallel": 2.8461217880249023e-06, "grad_norm": 0.919888973236084, "learning_rate": 9.24120389995761e-07, "loss": 0.5802, "mean_token_accuracy": 0.8192832469940186, "num_tokens": 8371120.0, "step": 219 }, { "epoch": 0.027986261289912225, "ewc_loss": 0.0030975341796875, "ewc_loss_diag": 3.0547380447387695e-06, "ewc_loss_parallel": 2.7865171432495117e-06, "grad_norm": 1.0838367938995361, "learning_rate": 9.283594743535397e-07, "loss": 0.7018, "mean_token_accuracy": 0.7866016626358032, "num_tokens": 8407502.0, "step": 220 }, { "epoch": 0.028113471568502735, "ewc_loss": 0.003265380859375, "ewc_loss_diag": 3.0547380447387695e-06, "ewc_loss_parallel": 2.9653310775756836e-06, "grad_norm": 0.946540355682373, "learning_rate": 9.325985587113184e-07, "loss": 0.554, "mean_token_accuracy": 0.828901469707489, "num_tokens": 8444360.0, "step": 221 }, { "epoch": 0.028240681847093246, "ewc_loss": 0.003173828125, "ewc_loss_diag": 3.069639205932617e-06, "ewc_loss_parallel": 2.86102294921875e-06, "grad_norm": 0.998786985874176, "learning_rate": 9.368376430690972e-07, "loss": 0.5658, "mean_token_accuracy": 0.8248047828674316, "num_tokens": 8476749.0, "step": 222 }, { "epoch": 0.028367892125683754, "ewc_loss": 0.003173828125, "ewc_loss_diag": 3.0994415283203125e-06, "ewc_loss_parallel": 2.86102294921875e-06, "grad_norm": 0.9506920576095581, "learning_rate": 9.410767274268759e-07, "loss": 0.5961, "mean_token_accuracy": 0.8177688121795654, "num_tokens": 8517157.0, "step": 223 }, { "epoch": 0.028495102404274265, "ewc_loss": 0.003173828125, "ewc_loss_diag": 3.1441450119018555e-06, "ewc_loss_parallel": 2.8461217880249023e-06, "grad_norm": 0.9132827520370483, "learning_rate": 9.453158117846546e-07, "loss": 0.5674, "mean_token_accuracy": 0.8249573707580566, "num_tokens": 8556677.0, "step": 224 }, { "epoch": 0.028622312682864776, "ewc_loss": 0.00311279296875, "ewc_loss_diag": 3.203749656677246e-06, "ewc_loss_parallel": 2.7865171432495117e-06, "grad_norm": 0.9638309478759766, "learning_rate": 9.495548961424333e-07, "loss": 0.6023, "mean_token_accuracy": 0.8178927898406982, "num_tokens": 8594335.0, "step": 225 }, { "epoch": 0.028749522961455286, "ewc_loss": 0.003204345703125, "ewc_loss_diag": 3.2186508178710938e-06, "ewc_loss_parallel": 2.8908252716064453e-06, "grad_norm": 0.939350962638855, "learning_rate": 9.53793980500212e-07, "loss": 0.5235, "mean_token_accuracy": 0.8365694284439087, "num_tokens": 8626259.0, "step": 226 }, { "epoch": 0.028876733240045797, "ewc_loss": 0.0031890869140625, "ewc_loss_diag": 3.2335519790649414e-06, "ewc_loss_parallel": 2.86102294921875e-06, "grad_norm": 0.9552206993103027, "learning_rate": 9.580330648579909e-07, "loss": 0.6842, "mean_token_accuracy": 0.795199990272522, "num_tokens": 8665481.0, "step": 227 }, { "epoch": 0.029003943518636305, "ewc_loss": 0.0032196044921875, "ewc_loss_diag": 3.2633543014526367e-06, "ewc_loss_parallel": 2.905726432800293e-06, "grad_norm": 0.9328243136405945, "learning_rate": 9.622721492157695e-07, "loss": 0.5506, "mean_token_accuracy": 0.8316534757614136, "num_tokens": 8703136.0, "step": 228 }, { "epoch": 0.029131153797226816, "ewc_loss": 0.003204345703125, "ewc_loss_diag": 3.2782554626464844e-06, "ewc_loss_parallel": 2.8759241104125977e-06, "grad_norm": 0.986501932144165, "learning_rate": 9.665112335735482e-07, "loss": 0.5647, "mean_token_accuracy": 0.8235299587249756, "num_tokens": 8734989.0, "step": 229 }, { "epoch": 0.029258364075817327, "ewc_loss": 0.0032806396484375, "ewc_loss_diag": 3.293156623840332e-06, "ewc_loss_parallel": 2.950429916381836e-06, "grad_norm": 0.9441779851913452, "learning_rate": 9.707503179313269e-07, "loss": 0.5309, "mean_token_accuracy": 0.8328566551208496, "num_tokens": 8768145.0, "step": 230 }, { "epoch": 0.029385574354407838, "ewc_loss": 0.0032501220703125, "ewc_loss_diag": 3.3080577850341797e-06, "ewc_loss_parallel": 2.9206275939941406e-06, "grad_norm": 0.9709951877593994, "learning_rate": 9.749894022891056e-07, "loss": 0.584, "mean_token_accuracy": 0.8200896978378296, "num_tokens": 8804668.0, "step": 231 }, { "epoch": 0.029512784632998345, "ewc_loss": 0.003265380859375, "ewc_loss_diag": 3.337860107421875e-06, "ewc_loss_parallel": 2.9355287551879883e-06, "grad_norm": 0.9827685952186584, "learning_rate": 9.792284866468842e-07, "loss": 0.6059, "mean_token_accuracy": 0.8088885545730591, "num_tokens": 8848343.0, "step": 232 }, { "epoch": 0.029639994911588856, "ewc_loss": 0.003265380859375, "ewc_loss_diag": 3.3527612686157227e-06, "ewc_loss_parallel": 2.9355287551879883e-06, "grad_norm": 0.9480843544006348, "learning_rate": 9.834675710046631e-07, "loss": 0.5709, "mean_token_accuracy": 0.8197765946388245, "num_tokens": 8883794.0, "step": 233 }, { "epoch": 0.029767205190179367, "ewc_loss": 0.0032501220703125, "ewc_loss_diag": 3.4123659133911133e-06, "ewc_loss_parallel": 2.9206275939941406e-06, "grad_norm": 0.9356430768966675, "learning_rate": 9.877066553624418e-07, "loss": 0.591, "mean_token_accuracy": 0.8167980909347534, "num_tokens": 8921597.0, "step": 234 }, { "epoch": 0.029894415468769878, "ewc_loss": 0.003265380859375, "ewc_loss_diag": 3.4421682357788086e-06, "ewc_loss_parallel": 2.9206275939941406e-06, "grad_norm": 1.1494804620742798, "learning_rate": 9.919457397202205e-07, "loss": 0.6118, "mean_token_accuracy": 0.8119602203369141, "num_tokens": 8956260.0, "step": 235 }, { "epoch": 0.030021625747360385, "ewc_loss": 0.0035858154296875, "ewc_loss_diag": 3.471970558166504e-06, "ewc_loss_parallel": 3.2335519790649414e-06, "grad_norm": 1.0456427335739136, "learning_rate": 9.961848240779991e-07, "loss": 0.6836, "mean_token_accuracy": 0.795313835144043, "num_tokens": 8992492.0, "step": 236 }, { "epoch": 0.030148836025950896, "ewc_loss": 0.00341796875, "ewc_loss_diag": 3.4868717193603516e-06, "ewc_loss_parallel": 3.069639205932617e-06, "grad_norm": 1.0525200366973877, "learning_rate": 1.000423908435778e-06, "loss": 0.5882, "mean_token_accuracy": 0.8189712166786194, "num_tokens": 9028300.0, "step": 237 }, { "epoch": 0.030276046304541407, "ewc_loss": 0.0034332275390625, "ewc_loss_diag": 3.516674041748047e-06, "ewc_loss_parallel": 3.084540367126465e-06, "grad_norm": 0.9435207843780518, "learning_rate": 1.0046629927935567e-06, "loss": 0.5813, "mean_token_accuracy": 0.820970892906189, "num_tokens": 9066620.0, "step": 238 }, { "epoch": 0.030403256583131918, "ewc_loss": 0.0032806396484375, "ewc_loss_diag": 3.546476364135742e-06, "ewc_loss_parallel": 2.9355287551879883e-06, "grad_norm": 0.9183095097541809, "learning_rate": 1.0089020771513354e-06, "loss": 0.5267, "mean_token_accuracy": 0.837424099445343, "num_tokens": 9105213.0, "step": 239 }, { "epoch": 0.03053046686172243, "ewc_loss": 0.0032958984375, "ewc_loss_diag": 3.5762786865234375e-06, "ewc_loss_parallel": 2.950429916381836e-06, "grad_norm": 0.9788690805435181, "learning_rate": 1.013141161509114e-06, "loss": 0.5873, "mean_token_accuracy": 0.8191753029823303, "num_tokens": 9139673.0, "step": 240 }, { "epoch": 0.030657677140312936, "ewc_loss": 0.003387451171875, "ewc_loss_diag": 3.591179847717285e-06, "ewc_loss_parallel": 3.039836883544922e-06, "grad_norm": 0.9713954329490662, "learning_rate": 1.017380245866893e-06, "loss": 0.5893, "mean_token_accuracy": 0.8149325847625732, "num_tokens": 9172435.0, "step": 241 }, { "epoch": 0.030784887418903447, "ewc_loss": 0.0034332275390625, "ewc_loss_diag": 3.606081008911133e-06, "ewc_loss_parallel": 3.069639205932617e-06, "grad_norm": 0.9760288596153259, "learning_rate": 1.0216193302246716e-06, "loss": 0.6132, "mean_token_accuracy": 0.8057183027267456, "num_tokens": 9206015.0, "step": 242 }, { "epoch": 0.030912097697493958, "ewc_loss": 0.003448486328125, "ewc_loss_diag": 3.635883331298828e-06, "ewc_loss_parallel": 3.084540367126465e-06, "grad_norm": 0.9148854613304138, "learning_rate": 1.0258584145824503e-06, "loss": 0.5489, "mean_token_accuracy": 0.8281267881393433, "num_tokens": 9247224.0, "step": 243 }, { "epoch": 0.03103930797608447, "ewc_loss": 0.003387451171875, "ewc_loss_diag": 3.6507844924926758e-06, "ewc_loss_parallel": 3.0249357223510742e-06, "grad_norm": 0.9505922198295593, "learning_rate": 1.030097498940229e-06, "loss": 0.534, "mean_token_accuracy": 0.8330594897270203, "num_tokens": 9282171.0, "step": 244 }, { "epoch": 0.031166518254674976, "ewc_loss": 0.003448486328125, "ewc_loss_diag": 3.680586814880371e-06, "ewc_loss_parallel": 3.084540367126465e-06, "grad_norm": 0.917728841304779, "learning_rate": 1.0343365832980078e-06, "loss": 0.5858, "mean_token_accuracy": 0.8180699348449707, "num_tokens": 9324078.0, "step": 245 }, { "epoch": 0.03129372853326549, "ewc_loss": 0.0034027099609375, "ewc_loss_diag": 3.680586814880371e-06, "ewc_loss_parallel": 3.039836883544922e-06, "grad_norm": 0.9385110139846802, "learning_rate": 1.0385756676557865e-06, "loss": 0.5815, "mean_token_accuracy": 0.820418119430542, "num_tokens": 9364516.0, "step": 246 }, { "epoch": 0.031420938811855995, "ewc_loss": 0.00347900390625, "ewc_loss_diag": 3.7103891372680664e-06, "ewc_loss_parallel": 3.11434268951416e-06, "grad_norm": 0.9434341192245483, "learning_rate": 1.0428147520135652e-06, "loss": 0.5591, "mean_token_accuracy": 0.8264406323432922, "num_tokens": 9403804.0, "step": 247 }, { "epoch": 0.03154814909044651, "ewc_loss": 0.0035400390625, "ewc_loss_diag": 3.7401914596557617e-06, "ewc_loss_parallel": 3.1739473342895508e-06, "grad_norm": 0.9576502442359924, "learning_rate": 1.0470538363713438e-06, "loss": 0.6018, "mean_token_accuracy": 0.8125132322311401, "num_tokens": 9441589.0, "step": 248 }, { "epoch": 0.031675359369037016, "ewc_loss": 0.00360107421875, "ewc_loss_diag": 3.7401914596557617e-06, "ewc_loss_parallel": 3.2186508178710938e-06, "grad_norm": 0.962683379650116, "learning_rate": 1.0512929207291225e-06, "loss": 0.6023, "mean_token_accuracy": 0.8127244710922241, "num_tokens": 9480240.0, "step": 249 }, { "epoch": 0.03180256964762753, "ewc_loss": 0.0035858154296875, "ewc_loss_diag": 3.7550926208496094e-06, "ewc_loss_parallel": 3.203749656677246e-06, "grad_norm": 1.0495543479919434, "learning_rate": 1.0555320050869012e-06, "loss": 0.5638, "mean_token_accuracy": 0.8213520050048828, "num_tokens": 9512704.0, "step": 250 }, { "epoch": 0.03192977992621804, "ewc_loss": 0.00372314453125, "ewc_loss_diag": 3.769993782043457e-06, "ewc_loss_parallel": 3.337860107421875e-06, "grad_norm": 1.0439343452453613, "learning_rate": 1.05977108944468e-06, "loss": 0.6193, "mean_token_accuracy": 0.805199921131134, "num_tokens": 9545057.0, "step": 251 }, { "epoch": 0.032056990204808546, "ewc_loss": 0.00372314453125, "ewc_loss_diag": 3.7997961044311523e-06, "ewc_loss_parallel": 3.337860107421875e-06, "grad_norm": 0.9471247792243958, "learning_rate": 1.0640101738024587e-06, "loss": 0.5623, "mean_token_accuracy": 0.8265793323516846, "num_tokens": 9585797.0, "step": 252 }, { "epoch": 0.03218420048339906, "ewc_loss": 0.0035247802734375, "ewc_loss_diag": 3.814697265625e-06, "ewc_loss_parallel": 3.1441450119018555e-06, "grad_norm": 0.9649631381034851, "learning_rate": 1.0682492581602374e-06, "loss": 0.523, "mean_token_accuracy": 0.8372804522514343, "num_tokens": 9617890.0, "step": 253 }, { "epoch": 0.03231141076198957, "ewc_loss": 0.00360107421875, "ewc_loss_diag": 3.814697265625e-06, "ewc_loss_parallel": 3.2186508178710938e-06, "grad_norm": 0.9642612338066101, "learning_rate": 1.072488342518016e-06, "loss": 0.5225, "mean_token_accuracy": 0.8333213329315186, "num_tokens": 9652895.0, "step": 254 }, { "epoch": 0.03243862104058008, "ewc_loss": 0.00360107421875, "ewc_loss_diag": 3.874301910400391e-06, "ewc_loss_parallel": 3.2186508178710938e-06, "grad_norm": 0.965857744216919, "learning_rate": 1.076727426875795e-06, "loss": 0.583, "mean_token_accuracy": 0.8174750804901123, "num_tokens": 9691804.0, "step": 255 }, { "epoch": 0.03256583131917059, "ewc_loss": 0.0036468505859375, "ewc_loss_diag": 3.904104232788086e-06, "ewc_loss_parallel": 3.248453140258789e-06, "grad_norm": 0.9495276212692261, "learning_rate": 1.0809665112335736e-06, "loss": 0.608, "mean_token_accuracy": 0.8139078617095947, "num_tokens": 9736178.0, "step": 256 }, { "epoch": 0.0326930415977611, "ewc_loss": 0.0036468505859375, "ewc_loss_diag": 3.904104232788086e-06, "ewc_loss_parallel": 3.248453140258789e-06, "grad_norm": 0.9289876818656921, "learning_rate": 1.0852055955913523e-06, "loss": 0.499, "mean_token_accuracy": 0.843024492263794, "num_tokens": 9773867.0, "step": 257 }, { "epoch": 0.03282025187635161, "ewc_loss": 0.00360107421875, "ewc_loss_diag": 3.904104232788086e-06, "ewc_loss_parallel": 3.203749656677246e-06, "grad_norm": 0.94237220287323, "learning_rate": 1.089444679949131e-06, "loss": 0.6058, "mean_token_accuracy": 0.8123699426651001, "num_tokens": 9813764.0, "step": 258 }, { "epoch": 0.03294746215494212, "ewc_loss": 0.0036773681640625, "ewc_loss_diag": 3.9637088775634766e-06, "ewc_loss_parallel": 3.2782554626464844e-06, "grad_norm": 0.9339504241943359, "learning_rate": 1.0936837643069099e-06, "loss": 0.5605, "mean_token_accuracy": 0.8247318267822266, "num_tokens": 9855447.0, "step": 259 }, { "epoch": 0.03307467243353263, "ewc_loss": 0.003692626953125, "ewc_loss_diag": 3.9637088775634766e-06, "ewc_loss_parallel": 3.293156623840332e-06, "grad_norm": 0.9543271660804749, "learning_rate": 1.0979228486646885e-06, "loss": 0.5209, "mean_token_accuracy": 0.8391008377075195, "num_tokens": 9889806.0, "step": 260 }, { "epoch": 0.03320188271212314, "ewc_loss": 0.0037384033203125, "ewc_loss_diag": 3.993511199951172e-06, "ewc_loss_parallel": 3.337860107421875e-06, "grad_norm": 0.9848282933235168, "learning_rate": 1.1021619330224672e-06, "loss": 0.6613, "mean_token_accuracy": 0.7955621480941772, "num_tokens": 9929257.0, "step": 261 }, { "epoch": 0.03332909299071365, "ewc_loss": 0.0037841796875, "ewc_loss_diag": 3.993511199951172e-06, "ewc_loss_parallel": 3.382563591003418e-06, "grad_norm": 0.9647748470306396, "learning_rate": 1.1064010173802459e-06, "loss": 0.5931, "mean_token_accuracy": 0.8148229122161865, "num_tokens": 9967822.0, "step": 262 }, { "epoch": 0.03345630326930416, "ewc_loss": 0.003753662109375, "ewc_loss_diag": 3.993511199951172e-06, "ewc_loss_parallel": 3.3527612686157227e-06, "grad_norm": 0.9446191787719727, "learning_rate": 1.1106401017380248e-06, "loss": 0.524, "mean_token_accuracy": 0.8352522850036621, "num_tokens": 10004873.0, "step": 263 }, { "epoch": 0.03358351354789467, "ewc_loss": 0.003753662109375, "ewc_loss_diag": 3.993511199951172e-06, "ewc_loss_parallel": 3.3527612686157227e-06, "grad_norm": 1.0413482189178467, "learning_rate": 1.1148791860958034e-06, "loss": 0.6052, "mean_token_accuracy": 0.8086259961128235, "num_tokens": 10033761.0, "step": 264 }, { "epoch": 0.03371072382648518, "ewc_loss": 0.003875732421875, "ewc_loss_diag": 4.023313522338867e-06, "ewc_loss_parallel": 3.471970558166504e-06, "grad_norm": 0.9761093258857727, "learning_rate": 1.1191182704535821e-06, "loss": 0.5922, "mean_token_accuracy": 0.8153913617134094, "num_tokens": 10072586.0, "step": 265 }, { "epoch": 0.03383793410507569, "ewc_loss": 0.0037994384765625, "ewc_loss_diag": 4.023313522338867e-06, "ewc_loss_parallel": 3.3974647521972656e-06, "grad_norm": 0.9455331563949585, "learning_rate": 1.123357354811361e-06, "loss": 0.585, "mean_token_accuracy": 0.8185479640960693, "num_tokens": 10112210.0, "step": 266 }, { "epoch": 0.0339651443836662, "ewc_loss": 0.0037841796875, "ewc_loss_diag": 4.0531158447265625e-06, "ewc_loss_parallel": 3.3676624298095703e-06, "grad_norm": 1.003176212310791, "learning_rate": 1.1275964391691395e-06, "loss": 0.5683, "mean_token_accuracy": 0.8250418901443481, "num_tokens": 10155908.0, "step": 267 }, { "epoch": 0.03409235466225671, "ewc_loss": 0.00384521484375, "ewc_loss_diag": 4.0531158447265625e-06, "ewc_loss_parallel": 3.427267074584961e-06, "grad_norm": 0.9966697096824646, "learning_rate": 1.1318355235269181e-06, "loss": 0.6201, "mean_token_accuracy": 0.8067444562911987, "num_tokens": 10195944.0, "step": 268 }, { "epoch": 0.03421956494084722, "ewc_loss": 0.003875732421875, "ewc_loss_diag": 4.082918167114258e-06, "ewc_loss_parallel": 3.4570693969726562e-06, "grad_norm": 0.9667297005653381, "learning_rate": 1.136074607884697e-06, "loss": 0.6067, "mean_token_accuracy": 0.8124505281448364, "num_tokens": 10233602.0, "step": 269 }, { "epoch": 0.03434677521943773, "ewc_loss": 0.003814697265625, "ewc_loss_diag": 4.082918167114258e-06, "ewc_loss_parallel": 3.3974647521972656e-06, "grad_norm": 0.997513473033905, "learning_rate": 1.1403136922424757e-06, "loss": 0.5602, "mean_token_accuracy": 0.8248183727264404, "num_tokens": 10267001.0, "step": 270 }, { "epoch": 0.03447398549802824, "ewc_loss": 0.0038909912109375, "ewc_loss_diag": 4.112720489501953e-06, "ewc_loss_parallel": 3.471970558166504e-06, "grad_norm": 0.9512168765068054, "learning_rate": 1.1445527766002544e-06, "loss": 0.5689, "mean_token_accuracy": 0.8196677565574646, "num_tokens": 10308264.0, "step": 271 }, { "epoch": 0.03460119577661875, "ewc_loss": 0.0038299560546875, "ewc_loss_diag": 4.1425228118896484e-06, "ewc_loss_parallel": 3.4123659133911133e-06, "grad_norm": 1.0297605991363525, "learning_rate": 1.148791860958033e-06, "loss": 0.6696, "mean_token_accuracy": 0.7941149473190308, "num_tokens": 10343440.0, "step": 272 }, { "epoch": 0.034728406055209264, "ewc_loss": 0.00390625, "ewc_loss_diag": 4.1425228118896484e-06, "ewc_loss_parallel": 3.5017728805541992e-06, "grad_norm": 0.9879862070083618, "learning_rate": 1.153030945315812e-06, "loss": 0.5976, "mean_token_accuracy": 0.8159505128860474, "num_tokens": 10383402.0, "step": 273 }, { "epoch": 0.03485561633379977, "ewc_loss": 0.00390625, "ewc_loss_diag": 4.1425228118896484e-06, "ewc_loss_parallel": 3.5017728805541992e-06, "grad_norm": 0.9631134271621704, "learning_rate": 1.1572700296735906e-06, "loss": 0.573, "mean_token_accuracy": 0.8191885948181152, "num_tokens": 10421700.0, "step": 274 }, { "epoch": 0.03498282661239028, "ewc_loss": 0.0038604736328125, "ewc_loss_diag": 4.172325134277344e-06, "ewc_loss_parallel": 3.4421682357788086e-06, "grad_norm": 1.0031510591506958, "learning_rate": 1.1615091140313693e-06, "loss": 0.6128, "mean_token_accuracy": 0.8114879131317139, "num_tokens": 10460295.0, "step": 275 }, { "epoch": 0.03511003689098079, "ewc_loss": 0.00396728515625, "ewc_loss_diag": 4.231929779052734e-06, "ewc_loss_parallel": 3.546476364135742e-06, "grad_norm": 0.9589861035346985, "learning_rate": 1.165748198389148e-06, "loss": 0.5607, "mean_token_accuracy": 0.8251866102218628, "num_tokens": 10501619.0, "step": 276 }, { "epoch": 0.0352372471695713, "ewc_loss": 0.00390625, "ewc_loss_diag": 4.231929779052734e-06, "ewc_loss_parallel": 3.471970558166504e-06, "grad_norm": 0.9559510946273804, "learning_rate": 1.1699872827469268e-06, "loss": 0.5639, "mean_token_accuracy": 0.8243951201438904, "num_tokens": 10540465.0, "step": 277 }, { "epoch": 0.03536445744816181, "ewc_loss": 0.003936767578125, "ewc_loss_diag": 4.26173210144043e-06, "ewc_loss_parallel": 3.516674041748047e-06, "grad_norm": 1.0408991575241089, "learning_rate": 1.1742263671047055e-06, "loss": 0.5711, "mean_token_accuracy": 0.8219743967056274, "num_tokens": 10580980.0, "step": 278 }, { "epoch": 0.03549166772675232, "ewc_loss": 0.0040283203125, "ewc_loss_diag": 4.291534423828125e-06, "ewc_loss_parallel": 3.606081008911133e-06, "grad_norm": 1.0299376249313354, "learning_rate": 1.1784654514624842e-06, "loss": 0.6435, "mean_token_accuracy": 0.7991958856582642, "num_tokens": 10621332.0, "step": 279 }, { "epoch": 0.03561887800534283, "ewc_loss": 0.003997802734375, "ewc_loss_diag": 4.32133674621582e-06, "ewc_loss_parallel": 3.56137752532959e-06, "grad_norm": 1.0082968473434448, "learning_rate": 1.1827045358202629e-06, "loss": 0.5967, "mean_token_accuracy": 0.8185296058654785, "num_tokens": 10654468.0, "step": 280 }, { "epoch": 0.035746088283933344, "ewc_loss": 0.003997802734375, "ewc_loss_diag": 4.32133674621582e-06, "ewc_loss_parallel": 3.56137752532959e-06, "grad_norm": 0.9627666473388672, "learning_rate": 1.1869436201780417e-06, "loss": 0.5172, "mean_token_accuracy": 0.8386507034301758, "num_tokens": 10692191.0, "step": 281 }, { "epoch": 0.03587329856252385, "ewc_loss": 0.00396728515625, "ewc_loss_diag": 4.32133674621582e-06, "ewc_loss_parallel": 3.546476364135742e-06, "grad_norm": 0.95736163854599, "learning_rate": 1.1911827045358204e-06, "loss": 0.5576, "mean_token_accuracy": 0.8307207822799683, "num_tokens": 10731568.0, "step": 282 }, { "epoch": 0.03600050884111436, "ewc_loss": 0.00396728515625, "ewc_loss_diag": 4.32133674621582e-06, "ewc_loss_parallel": 3.546476364135742e-06, "grad_norm": 0.9879003763198853, "learning_rate": 1.195421788893599e-06, "loss": 0.629, "mean_token_accuracy": 0.8090973496437073, "num_tokens": 10772166.0, "step": 283 }, { "epoch": 0.036127719119704874, "ewc_loss": 0.0040283203125, "ewc_loss_diag": 4.32133674621582e-06, "ewc_loss_parallel": 3.591179847717285e-06, "grad_norm": 0.9856236577033997, "learning_rate": 1.199660873251378e-06, "loss": 0.5711, "mean_token_accuracy": 0.8239172101020813, "num_tokens": 10807527.0, "step": 284 }, { "epoch": 0.03625492939829538, "ewc_loss": 0.0040283203125, "ewc_loss_diag": 4.351139068603516e-06, "ewc_loss_parallel": 3.591179847717285e-06, "grad_norm": 1.0012203454971313, "learning_rate": 1.2038999576091564e-06, "loss": 0.5446, "mean_token_accuracy": 0.8300286531448364, "num_tokens": 10846064.0, "step": 285 }, { "epoch": 0.036382139676885895, "ewc_loss": 0.004058837890625, "ewc_loss_diag": 4.380941390991211e-06, "ewc_loss_parallel": 3.6209821701049805e-06, "grad_norm": 0.9739910960197449, "learning_rate": 1.208139041966935e-06, "loss": 0.5218, "mean_token_accuracy": 0.8396990299224854, "num_tokens": 10884496.0, "step": 286 }, { "epoch": 0.0365093499554764, "ewc_loss": 0.0040283203125, "ewc_loss_diag": 4.380941390991211e-06, "ewc_loss_parallel": 3.591179847717285e-06, "grad_norm": 1.0181853771209717, "learning_rate": 1.212378126324714e-06, "loss": 0.5653, "mean_token_accuracy": 0.8213337659835815, "num_tokens": 10917542.0, "step": 287 }, { "epoch": 0.03663656023406691, "ewc_loss": 0.004180908203125, "ewc_loss_diag": 4.380941390991211e-06, "ewc_loss_parallel": 3.7550926208496094e-06, "grad_norm": 1.0520350933074951, "learning_rate": 1.2166172106824927e-06, "loss": 0.6299, "mean_token_accuracy": 0.8108175992965698, "num_tokens": 10949596.0, "step": 288 }, { "epoch": 0.036763770512657425, "ewc_loss": 0.004241943359375, "ewc_loss_diag": 4.380941390991211e-06, "ewc_loss_parallel": 3.814697265625e-06, "grad_norm": 0.9793835878372192, "learning_rate": 1.2208562950402713e-06, "loss": 0.5356, "mean_token_accuracy": 0.8295024633407593, "num_tokens": 10986143.0, "step": 289 }, { "epoch": 0.03689098079124793, "ewc_loss": 0.004058837890625, "ewc_loss_diag": 4.410743713378906e-06, "ewc_loss_parallel": 3.635883331298828e-06, "grad_norm": 0.993441641330719, "learning_rate": 1.22509537939805e-06, "loss": 0.5743, "mean_token_accuracy": 0.819037914276123, "num_tokens": 11024678.0, "step": 290 }, { "epoch": 0.03701819106983844, "ewc_loss": 0.00421142578125, "ewc_loss_diag": 4.4405460357666016e-06, "ewc_loss_parallel": 3.7550926208496094e-06, "grad_norm": 0.9892035722732544, "learning_rate": 1.2293344637558289e-06, "loss": 0.5911, "mean_token_accuracy": 0.816170334815979, "num_tokens": 11068950.0, "step": 291 }, { "epoch": 0.037145401348428954, "ewc_loss": 0.004150390625, "ewc_loss_diag": 4.500150680541992e-06, "ewc_loss_parallel": 3.7103891372680664e-06, "grad_norm": 0.9811976552009583, "learning_rate": 1.2335735481136076e-06, "loss": 0.5396, "mean_token_accuracy": 0.8270294070243835, "num_tokens": 11108697.0, "step": 292 }, { "epoch": 0.03727261162701946, "ewc_loss": 0.004150390625, "ewc_loss_diag": 4.500150680541992e-06, "ewc_loss_parallel": 3.7103891372680664e-06, "grad_norm": 0.9839898943901062, "learning_rate": 1.2378126324713862e-06, "loss": 0.558, "mean_token_accuracy": 0.8288312554359436, "num_tokens": 11148482.0, "step": 293 }, { "epoch": 0.037399821905609976, "ewc_loss": 0.004241943359375, "ewc_loss_diag": 4.5299530029296875e-06, "ewc_loss_parallel": 3.7997961044311523e-06, "grad_norm": 0.9780330657958984, "learning_rate": 1.242051716829165e-06, "loss": 0.5728, "mean_token_accuracy": 0.8147270083427429, "num_tokens": 11189139.0, "step": 294 }, { "epoch": 0.03752703218420048, "ewc_loss": 0.004241943359375, "ewc_loss_diag": 4.5299530029296875e-06, "ewc_loss_parallel": 3.7997961044311523e-06, "grad_norm": 0.9770030379295349, "learning_rate": 1.2462908011869438e-06, "loss": 0.5583, "mean_token_accuracy": 0.8237560987472534, "num_tokens": 11226484.0, "step": 295 }, { "epoch": 0.03765424246279099, "ewc_loss": 0.004241943359375, "ewc_loss_diag": 4.559755325317383e-06, "ewc_loss_parallel": 3.7997961044311523e-06, "grad_norm": 0.9806024432182312, "learning_rate": 1.2505298855447223e-06, "loss": 0.5129, "mean_token_accuracy": 0.8377929329872131, "num_tokens": 11263806.0, "step": 296 }, { "epoch": 0.037781452741381505, "ewc_loss": 0.004302978515625, "ewc_loss_diag": 4.589557647705078e-06, "ewc_loss_parallel": 3.844499588012695e-06, "grad_norm": 1.0233125686645508, "learning_rate": 1.2547689699025011e-06, "loss": 0.6361, "mean_token_accuracy": 0.8043566942214966, "num_tokens": 11299822.0, "step": 297 }, { "epoch": 0.03790866301997201, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 4.589557647705078e-06, "ewc_loss_parallel": 3.904104232788086e-06, "grad_norm": 1.0816514492034912, "learning_rate": 1.2590080542602798e-06, "loss": 0.5966, "mean_token_accuracy": 0.8137538433074951, "num_tokens": 11330796.0, "step": 298 }, { "epoch": 0.03803587329856253, "ewc_loss": 0.004425048828125, "ewc_loss_diag": 4.649162292480469e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 1.000815510749817, "learning_rate": 1.2632471386180587e-06, "loss": 0.5877, "mean_token_accuracy": 0.8162336945533752, "num_tokens": 11369529.0, "step": 299 }, { "epoch": 0.038163083577153034, "ewc_loss": 0.004302978515625, "ewc_loss_diag": 4.649162292480469e-06, "ewc_loss_parallel": 3.844499588012695e-06, "grad_norm": 1.0141297578811646, "learning_rate": 1.2674862229758372e-06, "loss": 0.5608, "mean_token_accuracy": 0.8316142559051514, "num_tokens": 11405044.0, "step": 300 }, { "epoch": 0.03829029385574354, "ewc_loss": 0.00439453125, "ewc_loss_diag": 4.708766937255859e-06, "ewc_loss_parallel": 3.933906555175781e-06, "grad_norm": 0.9860320091247559, "learning_rate": 1.271725307333616e-06, "loss": 0.5559, "mean_token_accuracy": 0.8276753425598145, "num_tokens": 11445409.0, "step": 301 }, { "epoch": 0.038417504134334056, "ewc_loss": 0.00433349609375, "ewc_loss_diag": 4.708766937255859e-06, "ewc_loss_parallel": 3.874301910400391e-06, "grad_norm": 0.9628744721412659, "learning_rate": 1.2759643916913947e-06, "loss": 0.5504, "mean_token_accuracy": 0.8260682821273804, "num_tokens": 11488285.0, "step": 302 }, { "epoch": 0.03854471441292456, "ewc_loss": 0.004364013671875, "ewc_loss_diag": 4.798173904418945e-06, "ewc_loss_parallel": 3.874301910400391e-06, "grad_norm": 1.0254372358322144, "learning_rate": 1.2802034760491736e-06, "loss": 0.5575, "mean_token_accuracy": 0.8255860209465027, "num_tokens": 11520970.0, "step": 303 }, { "epoch": 0.03867192469151508, "ewc_loss": 0.004425048828125, "ewc_loss_diag": 4.798173904418945e-06, "ewc_loss_parallel": 3.933906555175781e-06, "grad_norm": 0.9881832003593445, "learning_rate": 1.284442560406952e-06, "loss": 0.6154, "mean_token_accuracy": 0.8066660165786743, "num_tokens": 11563590.0, "step": 304 }, { "epoch": 0.038799134970105585, "ewc_loss": 0.00439453125, "ewc_loss_diag": 4.827976226806641e-06, "ewc_loss_parallel": 3.904104232788086e-06, "grad_norm": 1.0600084066390991, "learning_rate": 1.288681644764731e-06, "loss": 0.6556, "mean_token_accuracy": 0.7943137884140015, "num_tokens": 11598629.0, "step": 305 }, { "epoch": 0.03892634524869609, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 4.857778549194336e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 1.0227701663970947, "learning_rate": 1.2929207291225096e-06, "loss": 0.5545, "mean_token_accuracy": 0.8294433355331421, "num_tokens": 11635526.0, "step": 306 }, { "epoch": 0.03905355552728661, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 4.887580871582031e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 1.019207239151001, "learning_rate": 1.2971598134802885e-06, "loss": 0.5728, "mean_token_accuracy": 0.8217974901199341, "num_tokens": 11672974.0, "step": 307 }, { "epoch": 0.039180765805877114, "ewc_loss": 0.004425048828125, "ewc_loss_diag": 4.887580871582031e-06, "ewc_loss_parallel": 3.933906555175781e-06, "grad_norm": 1.041421890258789, "learning_rate": 1.301398897838067e-06, "loss": 0.5934, "mean_token_accuracy": 0.8164749145507812, "num_tokens": 11708886.0, "step": 308 }, { "epoch": 0.03930797608446762, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 4.9173831939697266e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 0.9964346289634705, "learning_rate": 1.3056379821958458e-06, "loss": 0.6027, "mean_token_accuracy": 0.8172144889831543, "num_tokens": 11749217.0, "step": 309 }, { "epoch": 0.039435186363058136, "ewc_loss": 0.00439453125, "ewc_loss_diag": 4.9173831939697266e-06, "ewc_loss_parallel": 3.904104232788086e-06, "grad_norm": 1.0018810033798218, "learning_rate": 1.3098770665536245e-06, "loss": 0.5646, "mean_token_accuracy": 0.825576901435852, "num_tokens": 11788380.0, "step": 310 }, { "epoch": 0.039562396641648644, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 4.9173831939697266e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 1.0562214851379395, "learning_rate": 1.3141161509114034e-06, "loss": 0.5357, "mean_token_accuracy": 0.8305330276489258, "num_tokens": 11820587.0, "step": 311 }, { "epoch": 0.03968960692023916, "ewc_loss": 0.004486083984375, "ewc_loss_diag": 4.9173831939697266e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 1.0198806524276733, "learning_rate": 1.3183552352691819e-06, "loss": 0.5422, "mean_token_accuracy": 0.8311158418655396, "num_tokens": 11854473.0, "step": 312 }, { "epoch": 0.039816817198829665, "ewc_loss": 0.004425048828125, "ewc_loss_diag": 4.9173831939697266e-06, "ewc_loss_parallel": 3.933906555175781e-06, "grad_norm": 0.9831312298774719, "learning_rate": 1.3225943196269605e-06, "loss": 0.5805, "mean_token_accuracy": 0.8233615756034851, "num_tokens": 11894744.0, "step": 313 }, { "epoch": 0.03994402747742017, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 4.9173831939697266e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 0.9834412932395935, "learning_rate": 1.3268334039847394e-06, "loss": 0.5417, "mean_token_accuracy": 0.8303599953651428, "num_tokens": 11932346.0, "step": 314 }, { "epoch": 0.04007123775601069, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 4.976987838745117e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 1.031469702720642, "learning_rate": 1.331072488342518e-06, "loss": 0.6001, "mean_token_accuracy": 0.8100851774215698, "num_tokens": 11967683.0, "step": 315 }, { "epoch": 0.040198448034601195, "ewc_loss": 0.004486083984375, "ewc_loss_diag": 4.976987838745117e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 0.9876194000244141, "learning_rate": 1.335311572700297e-06, "loss": 0.5865, "mean_token_accuracy": 0.8160735368728638, "num_tokens": 12010521.0, "step": 316 }, { "epoch": 0.04032565831319171, "ewc_loss": 0.004425048828125, "ewc_loss_diag": 5.0067901611328125e-06, "ewc_loss_parallel": 3.933906555175781e-06, "grad_norm": 0.9766626954078674, "learning_rate": 1.3395506570580754e-06, "loss": 0.5578, "mean_token_accuracy": 0.8258333206176758, "num_tokens": 12053317.0, "step": 317 }, { "epoch": 0.040452868591782216, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 5.0067901611328125e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 1.1185898780822754, "learning_rate": 1.3437897414158543e-06, "loss": 0.6151, "mean_token_accuracy": 0.8083981871604919, "num_tokens": 12081589.0, "step": 318 }, { "epoch": 0.040580078870372724, "ewc_loss": 0.00457763671875, "ewc_loss_diag": 5.036592483520508e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 1.056647777557373, "learning_rate": 1.348028825773633e-06, "loss": 0.5554, "mean_token_accuracy": 0.824539303779602, "num_tokens": 12114826.0, "step": 319 }, { "epoch": 0.04070728914896324, "ewc_loss": 0.00445556640625, "ewc_loss_diag": 5.036592483520508e-06, "ewc_loss_parallel": 3.9637088775634766e-06, "grad_norm": 1.0086745023727417, "learning_rate": 1.3522679101314119e-06, "loss": 0.5364, "mean_token_accuracy": 0.8315538167953491, "num_tokens": 12150769.0, "step": 320 }, { "epoch": 0.040834499427553746, "ewc_loss": 0.0045166015625, "ewc_loss_diag": 5.066394805908203e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 1.0044281482696533, "learning_rate": 1.3565069944891903e-06, "loss": 0.5579, "mean_token_accuracy": 0.8272332549095154, "num_tokens": 12189035.0, "step": 321 }, { "epoch": 0.04096170970614425, "ewc_loss": 0.0045166015625, "ewc_loss_diag": 5.125999450683594e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 1.0151551961898804, "learning_rate": 1.3607460788469692e-06, "loss": 0.5482, "mean_token_accuracy": 0.8280484676361084, "num_tokens": 12225974.0, "step": 322 }, { "epoch": 0.04108891998473477, "ewc_loss": 0.0045166015625, "ewc_loss_diag": 5.125999450683594e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 1.0175753831863403, "learning_rate": 1.364985163204748e-06, "loss": 0.6164, "mean_token_accuracy": 0.8090301156044006, "num_tokens": 12266971.0, "step": 323 }, { "epoch": 0.041216130263325275, "ewc_loss": 0.0045166015625, "ewc_loss_diag": 5.185604095458984e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 1.054550290107727, "learning_rate": 1.3692242475625268e-06, "loss": 0.5278, "mean_token_accuracy": 0.8399007320404053, "num_tokens": 12306365.0, "step": 324 }, { "epoch": 0.04134334054191579, "ewc_loss": 0.004547119140625, "ewc_loss_diag": 5.185604095458984e-06, "ewc_loss_parallel": 4.023313522338867e-06, "grad_norm": 1.019200325012207, "learning_rate": 1.3734633319203052e-06, "loss": 0.5821, "mean_token_accuracy": 0.8285413384437561, "num_tokens": 12345343.0, "step": 325 }, { "epoch": 0.0414705508205063, "ewc_loss": 0.0045166015625, "ewc_loss_diag": 5.185604095458984e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 1.046457290649414, "learning_rate": 1.3777024162780841e-06, "loss": 0.6475, "mean_token_accuracy": 0.8002352118492126, "num_tokens": 12384532.0, "step": 326 }, { "epoch": 0.041597761099096804, "ewc_loss": 0.004547119140625, "ewc_loss_diag": 5.21540641784668e-06, "ewc_loss_parallel": 4.023313522338867e-06, "grad_norm": 0.9999906420707703, "learning_rate": 1.3819415006358628e-06, "loss": 0.5106, "mean_token_accuracy": 0.8398220539093018, "num_tokens": 12424955.0, "step": 327 }, { "epoch": 0.04172497137768732, "ewc_loss": 0.0045166015625, "ewc_loss_diag": 5.245208740234375e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 0.9566069841384888, "learning_rate": 1.3861805849936417e-06, "loss": 0.5313, "mean_token_accuracy": 0.8360843658447266, "num_tokens": 12469583.0, "step": 328 }, { "epoch": 0.041852181656277826, "ewc_loss": 0.0045166015625, "ewc_loss_diag": 5.304813385009766e-06, "ewc_loss_parallel": 3.993511199951172e-06, "grad_norm": 1.0443503856658936, "learning_rate": 1.3904196693514201e-06, "loss": 0.5773, "mean_token_accuracy": 0.8171480894088745, "num_tokens": 12504997.0, "step": 329 }, { "epoch": 0.04197939193486834, "ewc_loss": 0.004638671875, "ewc_loss_diag": 5.334615707397461e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 1.0255961418151855, "learning_rate": 1.394658753709199e-06, "loss": 0.579, "mean_token_accuracy": 0.8217468857765198, "num_tokens": 12546248.0, "step": 330 }, { "epoch": 0.04210660221345885, "ewc_loss": 0.004547119140625, "ewc_loss_diag": 5.304813385009766e-06, "ewc_loss_parallel": 4.023313522338867e-06, "grad_norm": 1.0488487482070923, "learning_rate": 1.3988978380669777e-06, "loss": 0.5973, "mean_token_accuracy": 0.81427001953125, "num_tokens": 12580741.0, "step": 331 }, { "epoch": 0.042233812492049355, "ewc_loss": 0.004669189453125, "ewc_loss_diag": 5.364418029785156e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 1.0122392177581787, "learning_rate": 1.4031369224247562e-06, "loss": 0.6035, "mean_token_accuracy": 0.8133020401000977, "num_tokens": 12622539.0, "step": 332 }, { "epoch": 0.04236102277063987, "ewc_loss": 0.00457763671875, "ewc_loss_diag": 5.3942203521728516e-06, "ewc_loss_parallel": 4.023313522338867e-06, "grad_norm": 1.0280343294143677, "learning_rate": 1.407376006782535e-06, "loss": 0.5583, "mean_token_accuracy": 0.824722409248352, "num_tokens": 12660137.0, "step": 333 }, { "epoch": 0.04248823304923038, "ewc_loss": 0.004608154296875, "ewc_loss_diag": 5.3942203521728516e-06, "ewc_loss_parallel": 4.0531158447265625e-06, "grad_norm": 1.007444143295288, "learning_rate": 1.4116150911403137e-06, "loss": 0.5843, "mean_token_accuracy": 0.8186426162719727, "num_tokens": 12702341.0, "step": 334 }, { "epoch": 0.04261544332782089, "ewc_loss": 0.004608154296875, "ewc_loss_diag": 5.424022674560547e-06, "ewc_loss_parallel": 4.0531158447265625e-06, "grad_norm": 1.008756160736084, "learning_rate": 1.4158541754980926e-06, "loss": 0.5261, "mean_token_accuracy": 0.8367210030555725, "num_tokens": 12739560.0, "step": 335 }, { "epoch": 0.0427426536064114, "ewc_loss": 0.004608154296875, "ewc_loss_diag": 5.424022674560547e-06, "ewc_loss_parallel": 4.0531158447265625e-06, "grad_norm": 1.0006530284881592, "learning_rate": 1.420093259855871e-06, "loss": 0.563, "mean_token_accuracy": 0.8307936191558838, "num_tokens": 12781534.0, "step": 336 }, { "epoch": 0.042869863885001906, "ewc_loss": 0.004638671875, "ewc_loss_diag": 5.453824996948242e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 1.054871678352356, "learning_rate": 1.42433234421365e-06, "loss": 0.568, "mean_token_accuracy": 0.8226033449172974, "num_tokens": 12815840.0, "step": 337 }, { "epoch": 0.04299707416359242, "ewc_loss": 0.004730224609375, "ewc_loss_diag": 5.453824996948242e-06, "ewc_loss_parallel": 4.172325134277344e-06, "grad_norm": 1.0667345523834229, "learning_rate": 1.4285714285714286e-06, "loss": 0.5763, "mean_token_accuracy": 0.821305513381958, "num_tokens": 12849201.0, "step": 338 }, { "epoch": 0.04312428444218293, "ewc_loss": 0.00469970703125, "ewc_loss_diag": 5.453824996948242e-06, "ewc_loss_parallel": 4.1425228118896484e-06, "grad_norm": 1.059046983718872, "learning_rate": 1.4328105129292075e-06, "loss": 0.5801, "mean_token_accuracy": 0.8212238550186157, "num_tokens": 12888490.0, "step": 339 }, { "epoch": 0.043251494720773435, "ewc_loss": 0.00469970703125, "ewc_loss_diag": 5.453824996948242e-06, "ewc_loss_parallel": 4.1425228118896484e-06, "grad_norm": 1.0190353393554688, "learning_rate": 1.437049597286986e-06, "loss": 0.5546, "mean_token_accuracy": 0.8275439143180847, "num_tokens": 12927554.0, "step": 340 }, { "epoch": 0.04337870499936395, "ewc_loss": 0.004638671875, "ewc_loss_diag": 5.513429641723633e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 1.0728602409362793, "learning_rate": 1.4412886816447649e-06, "loss": 0.6007, "mean_token_accuracy": 0.8124215006828308, "num_tokens": 12960801.0, "step": 341 }, { "epoch": 0.04350591527795446, "ewc_loss": 0.0047607421875, "ewc_loss_diag": 5.513429641723633e-06, "ewc_loss_parallel": 4.202127456665039e-06, "grad_norm": 1.0654247999191284, "learning_rate": 1.4455277660025435e-06, "loss": 0.5193, "mean_token_accuracy": 0.8353451490402222, "num_tokens": 12993119.0, "step": 342 }, { "epoch": 0.04363312555654497, "ewc_loss": 0.00469970703125, "ewc_loss_diag": 5.543231964111328e-06, "ewc_loss_parallel": 4.1425228118896484e-06, "grad_norm": 1.0089482069015503, "learning_rate": 1.4497668503603224e-06, "loss": 0.5794, "mean_token_accuracy": 0.8202893137931824, "num_tokens": 13033966.0, "step": 343 }, { "epoch": 0.04376033583513548, "ewc_loss": 0.004638671875, "ewc_loss_diag": 5.543231964111328e-06, "ewc_loss_parallel": 4.082918167114258e-06, "grad_norm": 1.0832229852676392, "learning_rate": 1.4540059347181009e-06, "loss": 0.5567, "mean_token_accuracy": 0.8272922039031982, "num_tokens": 13063649.0, "step": 344 }, { "epoch": 0.043887546113725986, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 5.543231964111328e-06, "ewc_loss_parallel": 4.26173210144043e-06, "grad_norm": 1.0313531160354614, "learning_rate": 1.4582450190758798e-06, "loss": 0.61, "mean_token_accuracy": 0.8105862140655518, "num_tokens": 13106926.0, "step": 345 }, { "epoch": 0.0440147563923165, "ewc_loss": 0.004669189453125, "ewc_loss_diag": 5.5730342864990234e-06, "ewc_loss_parallel": 4.112720489501953e-06, "grad_norm": 1.0295499563217163, "learning_rate": 1.4624841034336584e-06, "loss": 0.5553, "mean_token_accuracy": 0.8272304534912109, "num_tokens": 13142961.0, "step": 346 }, { "epoch": 0.04414196667090701, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 5.5730342864990234e-06, "ewc_loss_parallel": 4.26173210144043e-06, "grad_norm": 1.0249204635620117, "learning_rate": 1.4667231877914373e-06, "loss": 0.5365, "mean_token_accuracy": 0.8323636054992676, "num_tokens": 13183544.0, "step": 347 }, { "epoch": 0.04426917694949752, "ewc_loss": 0.0047607421875, "ewc_loss_diag": 5.5730342864990234e-06, "ewc_loss_parallel": 4.202127456665039e-06, "grad_norm": 1.021447777748108, "learning_rate": 1.470962272149216e-06, "loss": 0.553, "mean_token_accuracy": 0.826596736907959, "num_tokens": 13222609.0, "step": 348 }, { "epoch": 0.04439638722808803, "ewc_loss": 0.004791259765625, "ewc_loss_diag": 5.602836608886719e-06, "ewc_loss_parallel": 4.231929779052734e-06, "grad_norm": 1.0536093711853027, "learning_rate": 1.4752013565069945e-06, "loss": 0.5516, "mean_token_accuracy": 0.8287138342857361, "num_tokens": 13258267.0, "step": 349 }, { "epoch": 0.04452359750667854, "ewc_loss": 0.0048828125, "ewc_loss_diag": 5.602836608886719e-06, "ewc_loss_parallel": 4.32133674621582e-06, "grad_norm": 1.0608974695205688, "learning_rate": 1.4794404408647733e-06, "loss": 0.5583, "mean_token_accuracy": 0.827157735824585, "num_tokens": 13294681.0, "step": 350 }, { "epoch": 0.04465080778526905, "ewc_loss": 0.0048828125, "ewc_loss_diag": 5.602836608886719e-06, "ewc_loss_parallel": 4.32133674621582e-06, "grad_norm": 1.0376877784729004, "learning_rate": 1.483679525222552e-06, "loss": 0.506, "mean_token_accuracy": 0.8382989764213562, "num_tokens": 13335469.0, "step": 351 }, { "epoch": 0.04477801806385956, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 5.632638931274414e-06, "ewc_loss_parallel": 4.26173210144043e-06, "grad_norm": 1.0232793092727661, "learning_rate": 1.487918609580331e-06, "loss": 0.4856, "mean_token_accuracy": 0.8427115678787231, "num_tokens": 13372465.0, "step": 352 }, { "epoch": 0.04490522834245007, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 5.662441253662109e-06, "ewc_loss_parallel": 4.26173210144043e-06, "grad_norm": 1.0489872694015503, "learning_rate": 1.4921576939381094e-06, "loss": 0.535, "mean_token_accuracy": 0.8348541855812073, "num_tokens": 13409218.0, "step": 353 }, { "epoch": 0.04503243862104058, "ewc_loss": 0.00494384765625, "ewc_loss_diag": 5.662441253662109e-06, "ewc_loss_parallel": 4.380941390991211e-06, "grad_norm": 1.0209739208221436, "learning_rate": 1.4963967782958882e-06, "loss": 0.5034, "mean_token_accuracy": 0.8421172499656677, "num_tokens": 13452548.0, "step": 354 }, { "epoch": 0.04515964889963109, "ewc_loss": 0.0048828125, "ewc_loss_diag": 5.692243576049805e-06, "ewc_loss_parallel": 4.291534423828125e-06, "grad_norm": 1.0831658840179443, "learning_rate": 1.500635862653667e-06, "loss": 0.5909, "mean_token_accuracy": 0.8180058598518372, "num_tokens": 13488325.0, "step": 355 }, { "epoch": 0.0452868591782216, "ewc_loss": 0.005035400390625, "ewc_loss_diag": 5.7220458984375e-06, "ewc_loss_parallel": 4.470348358154297e-06, "grad_norm": 1.0539379119873047, "learning_rate": 1.5048749470114458e-06, "loss": 0.5911, "mean_token_accuracy": 0.817585825920105, "num_tokens": 13528203.0, "step": 356 }, { "epoch": 0.04541406945681211, "ewc_loss": 0.004974365234375, "ewc_loss_diag": 5.751848220825195e-06, "ewc_loss_parallel": 4.380941390991211e-06, "grad_norm": 1.019164800643921, "learning_rate": 1.5091140313692243e-06, "loss": 0.5181, "mean_token_accuracy": 0.8365663290023804, "num_tokens": 13567242.0, "step": 357 }, { "epoch": 0.04554127973540262, "ewc_loss": 0.004913330078125, "ewc_loss_diag": 5.781650543212891e-06, "ewc_loss_parallel": 4.32133674621582e-06, "grad_norm": 1.037137746810913, "learning_rate": 1.5133531157270031e-06, "loss": 0.5693, "mean_token_accuracy": 0.818928599357605, "num_tokens": 13604521.0, "step": 358 }, { "epoch": 0.04566849001399313, "ewc_loss": 0.004974365234375, "ewc_loss_diag": 5.781650543212891e-06, "ewc_loss_parallel": 4.380941390991211e-06, "grad_norm": 1.0416858196258545, "learning_rate": 1.5175922000847818e-06, "loss": 0.5306, "mean_token_accuracy": 0.8352954983711243, "num_tokens": 13642787.0, "step": 359 }, { "epoch": 0.04579570029258364, "ewc_loss": 0.0050048828125, "ewc_loss_diag": 5.811452865600586e-06, "ewc_loss_parallel": 4.410743713378906e-06, "grad_norm": 1.0306755304336548, "learning_rate": 1.5218312844425607e-06, "loss": 0.5551, "mean_token_accuracy": 0.8205715417861938, "num_tokens": 13684113.0, "step": 360 }, { "epoch": 0.045922910571174154, "ewc_loss": 0.0050048828125, "ewc_loss_diag": 5.811452865600586e-06, "ewc_loss_parallel": 4.410743713378906e-06, "grad_norm": 1.0477356910705566, "learning_rate": 1.5260703688003392e-06, "loss": 0.4859, "mean_token_accuracy": 0.842883288860321, "num_tokens": 13719208.0, "step": 361 }, { "epoch": 0.04605012084976466, "ewc_loss": 0.005035400390625, "ewc_loss_diag": 5.811452865600586e-06, "ewc_loss_parallel": 4.4405460357666016e-06, "grad_norm": 1.0326073169708252, "learning_rate": 1.530309453158118e-06, "loss": 0.507, "mean_token_accuracy": 0.8403787612915039, "num_tokens": 13760159.0, "step": 362 }, { "epoch": 0.04617733112835517, "ewc_loss": 0.0050048828125, "ewc_loss_diag": 5.811452865600586e-06, "ewc_loss_parallel": 4.410743713378906e-06, "grad_norm": 1.0648952722549438, "learning_rate": 1.5345485375158967e-06, "loss": 0.5943, "mean_token_accuracy": 0.8146148920059204, "num_tokens": 13798375.0, "step": 363 }, { "epoch": 0.04630454140694568, "ewc_loss": 0.00506591796875, "ewc_loss_diag": 5.811452865600586e-06, "ewc_loss_parallel": 4.500150680541992e-06, "grad_norm": 1.0242228507995605, "learning_rate": 1.5387876218736756e-06, "loss": 0.472, "mean_token_accuracy": 0.849988579750061, "num_tokens": 13841924.0, "step": 364 }, { "epoch": 0.04643175168553619, "ewc_loss": 0.004974365234375, "ewc_loss_diag": 5.841255187988281e-06, "ewc_loss_parallel": 4.380941390991211e-06, "grad_norm": 1.0779664516448975, "learning_rate": 1.543026706231454e-06, "loss": 0.5822, "mean_token_accuracy": 0.8214449882507324, "num_tokens": 13881436.0, "step": 365 }, { "epoch": 0.0465589619641267, "ewc_loss": 0.005126953125, "ewc_loss_diag": 5.8710575103759766e-06, "ewc_loss_parallel": 4.559755325317383e-06, "grad_norm": 1.0972425937652588, "learning_rate": 1.547265790589233e-06, "loss": 0.5948, "mean_token_accuracy": 0.8148970603942871, "num_tokens": 13916295.0, "step": 366 }, { "epoch": 0.04668617224271721, "ewc_loss": 0.00506591796875, "ewc_loss_diag": 5.900859832763672e-06, "ewc_loss_parallel": 4.500150680541992e-06, "grad_norm": 1.5212345123291016, "learning_rate": 1.5515048749470116e-06, "loss": 0.5589, "mean_token_accuracy": 0.8242814540863037, "num_tokens": 13950189.0, "step": 367 }, { "epoch": 0.04681338252130772, "ewc_loss": 0.0057373046875, "ewc_loss_diag": 5.9604644775390625e-06, "ewc_loss_parallel": 5.155801773071289e-06, "grad_norm": 1.1056019067764282, "learning_rate": 1.55574395930479e-06, "loss": 0.5664, "mean_token_accuracy": 0.8237841129302979, "num_tokens": 13988422.0, "step": 368 }, { "epoch": 0.046940592799898234, "ewc_loss": 0.00482177734375, "ewc_loss_diag": 5.9604644775390625e-06, "ewc_loss_parallel": 4.202127456665039e-06, "grad_norm": 0.9899879693984985, "learning_rate": 1.559983043662569e-06, "loss": 0.5628, "mean_token_accuracy": 0.8246673941612244, "num_tokens": 14029199.0, "step": 369 }, { "epoch": 0.04706780307848874, "ewc_loss": 0.005096435546875, "ewc_loss_diag": 5.990266799926758e-06, "ewc_loss_parallel": 4.500150680541992e-06, "grad_norm": 1.0774012804031372, "learning_rate": 1.5642221280203476e-06, "loss": 0.5635, "mean_token_accuracy": 0.8262421488761902, "num_tokens": 14067976.0, "step": 370 }, { "epoch": 0.04719501335707925, "ewc_loss": 0.00518798828125, "ewc_loss_diag": 5.990266799926758e-06, "ewc_loss_parallel": 4.589557647705078e-06, "grad_norm": 1.0603188276290894, "learning_rate": 1.5684612123781265e-06, "loss": 0.4869, "mean_token_accuracy": 0.847169041633606, "num_tokens": 14102858.0, "step": 371 }, { "epoch": 0.04732222363566976, "ewc_loss": 0.00506591796875, "ewc_loss_diag": 6.020069122314453e-06, "ewc_loss_parallel": 4.470348358154297e-06, "grad_norm": 1.0780341625213623, "learning_rate": 1.572700296735905e-06, "loss": 0.5668, "mean_token_accuracy": 0.8287129402160645, "num_tokens": 14138560.0, "step": 372 }, { "epoch": 0.04744943391426027, "ewc_loss": 0.00518798828125, "ewc_loss_diag": 6.079673767089844e-06, "ewc_loss_parallel": 4.589557647705078e-06, "grad_norm": 1.117595911026001, "learning_rate": 1.5769393810936839e-06, "loss": 0.5557, "mean_token_accuracy": 0.8268039226531982, "num_tokens": 14168007.0, "step": 373 }, { "epoch": 0.047576644192850785, "ewc_loss": 0.00531005859375, "ewc_loss_diag": 6.109476089477539e-06, "ewc_loss_parallel": 4.708766937255859e-06, "grad_norm": 1.0714653730392456, "learning_rate": 1.5811784654514625e-06, "loss": 0.4851, "mean_token_accuracy": 0.8486257791519165, "num_tokens": 14201779.0, "step": 374 }, { "epoch": 0.04770385447144129, "ewc_loss": 0.005218505859375, "ewc_loss_diag": 6.109476089477539e-06, "ewc_loss_parallel": 4.6193599700927734e-06, "grad_norm": 1.0372508764266968, "learning_rate": 1.5854175498092414e-06, "loss": 0.6138, "mean_token_accuracy": 0.809286892414093, "num_tokens": 14245917.0, "step": 375 }, { "epoch": 0.0478310647500318, "ewc_loss": 0.005218505859375, "ewc_loss_diag": 6.139278411865234e-06, "ewc_loss_parallel": 4.6193599700927734e-06, "grad_norm": 1.0633444786071777, "learning_rate": 1.5896566341670199e-06, "loss": 0.5712, "mean_token_accuracy": 0.8194001913070679, "num_tokens": 14284032.0, "step": 376 }, { "epoch": 0.047958275028622314, "ewc_loss": 0.0052490234375, "ewc_loss_diag": 6.139278411865234e-06, "ewc_loss_parallel": 4.649162292480469e-06, "grad_norm": 1.1186360120773315, "learning_rate": 1.5938957185247988e-06, "loss": 0.6205, "mean_token_accuracy": 0.8041995763778687, "num_tokens": 14322077.0, "step": 377 }, { "epoch": 0.04808548530721282, "ewc_loss": 0.005340576171875, "ewc_loss_diag": 6.139278411865234e-06, "ewc_loss_parallel": 4.738569259643555e-06, "grad_norm": 1.0773769617080688, "learning_rate": 1.5981348028825774e-06, "loss": 0.5698, "mean_token_accuracy": 0.8214694261550903, "num_tokens": 14358031.0, "step": 378 }, { "epoch": 0.048212695585803336, "ewc_loss": 0.005279541015625, "ewc_loss_diag": 6.16908073425293e-06, "ewc_loss_parallel": 4.678964614868164e-06, "grad_norm": 1.1478543281555176, "learning_rate": 1.6023738872403563e-06, "loss": 0.6591, "mean_token_accuracy": 0.7939860224723816, "num_tokens": 14390230.0, "step": 379 }, { "epoch": 0.048339905864393844, "ewc_loss": 0.005401611328125, "ewc_loss_diag": 6.198883056640625e-06, "ewc_loss_parallel": 4.798173904418945e-06, "grad_norm": 1.0686001777648926, "learning_rate": 1.6066129715981348e-06, "loss": 0.5112, "mean_token_accuracy": 0.8395437002182007, "num_tokens": 14427394.0, "step": 380 }, { "epoch": 0.04846711614298435, "ewc_loss": 0.005218505859375, "ewc_loss_diag": 6.22868537902832e-06, "ewc_loss_parallel": 4.6193599700927734e-06, "grad_norm": 1.0265631675720215, "learning_rate": 1.6108520559559137e-06, "loss": 0.5074, "mean_token_accuracy": 0.8399186730384827, "num_tokens": 14466606.0, "step": 381 }, { "epoch": 0.048594326421574865, "ewc_loss": 0.005218505859375, "ewc_loss_diag": 6.22868537902832e-06, "ewc_loss_parallel": 4.6193599700927734e-06, "grad_norm": 1.146774172782898, "learning_rate": 1.6150911403136924e-06, "loss": 0.5735, "mean_token_accuracy": 0.8240435123443604, "num_tokens": 14494795.0, "step": 382 }, { "epoch": 0.04872153670016537, "ewc_loss": 0.00543212890625, "ewc_loss_diag": 6.22868537902832e-06, "ewc_loss_parallel": 4.827976226806641e-06, "grad_norm": 1.0590065717697144, "learning_rate": 1.6193302246714712e-06, "loss": 0.5753, "mean_token_accuracy": 0.8224904537200928, "num_tokens": 14537400.0, "step": 383 }, { "epoch": 0.04884874697875588, "ewc_loss": 0.005340576171875, "ewc_loss_diag": 6.288290023803711e-06, "ewc_loss_parallel": 4.708766937255859e-06, "grad_norm": 1.1802064180374146, "learning_rate": 1.62356930902925e-06, "loss": 0.5935, "mean_token_accuracy": 0.822938084602356, "num_tokens": 14575215.0, "step": 384 }, { "epoch": 0.048975957257346395, "ewc_loss": 0.005584716796875, "ewc_loss_diag": 6.288290023803711e-06, "ewc_loss_parallel": 4.947185516357422e-06, "grad_norm": 1.1353377103805542, "learning_rate": 1.6278083933870286e-06, "loss": 0.617, "mean_token_accuracy": 0.8070507049560547, "num_tokens": 14608486.0, "step": 385 }, { "epoch": 0.0491031675359369, "ewc_loss": 0.00543212890625, "ewc_loss_diag": 6.318092346191406e-06, "ewc_loss_parallel": 4.798173904418945e-06, "grad_norm": 1.0702499151229858, "learning_rate": 1.6320474777448073e-06, "loss": 0.5458, "mean_token_accuracy": 0.8283847570419312, "num_tokens": 14644583.0, "step": 386 }, { "epoch": 0.049230377814527417, "ewc_loss": 0.005340576171875, "ewc_loss_diag": 6.318092346191406e-06, "ewc_loss_parallel": 4.708766937255859e-06, "grad_norm": 1.0743052959442139, "learning_rate": 1.636286562102586e-06, "loss": 0.5692, "mean_token_accuracy": 0.8200781345367432, "num_tokens": 14682976.0, "step": 387 }, { "epoch": 0.049357588093117924, "ewc_loss": 0.00537109375, "ewc_loss_diag": 6.3478946685791016e-06, "ewc_loss_parallel": 4.738569259643555e-06, "grad_norm": 1.092366337776184, "learning_rate": 1.6405256464603648e-06, "loss": 0.541, "mean_token_accuracy": 0.8261769413948059, "num_tokens": 14717926.0, "step": 388 }, { "epoch": 0.04948479837170843, "ewc_loss": 0.00543212890625, "ewc_loss_diag": 6.3478946685791016e-06, "ewc_loss_parallel": 4.798173904418945e-06, "grad_norm": 1.1261590719223022, "learning_rate": 1.6447647308181433e-06, "loss": 0.5815, "mean_token_accuracy": 0.818154513835907, "num_tokens": 14750484.0, "step": 389 }, { "epoch": 0.049612008650298946, "ewc_loss": 0.00543212890625, "ewc_loss_diag": 6.377696990966797e-06, "ewc_loss_parallel": 4.798173904418945e-06, "grad_norm": 1.0855087041854858, "learning_rate": 1.6490038151759222e-06, "loss": 0.599, "mean_token_accuracy": 0.8145620822906494, "num_tokens": 14790044.0, "step": 390 }, { "epoch": 0.04973921892888945, "ewc_loss": 0.005401611328125, "ewc_loss_diag": 6.377696990966797e-06, "ewc_loss_parallel": 4.76837158203125e-06, "grad_norm": 1.0868537425994873, "learning_rate": 1.6532428995337008e-06, "loss": 0.5204, "mean_token_accuracy": 0.8346907496452332, "num_tokens": 14828768.0, "step": 391 }, { "epoch": 0.04986642920747997, "ewc_loss": 0.005462646484375, "ewc_loss_diag": 6.467103958129883e-06, "ewc_loss_parallel": 4.827976226806641e-06, "grad_norm": 1.0715357065200806, "learning_rate": 1.6574819838914797e-06, "loss": 0.5086, "mean_token_accuracy": 0.8444570899009705, "num_tokens": 14865848.0, "step": 392 }, { "epoch": 0.049993639486070475, "ewc_loss": 0.005401611328125, "ewc_loss_diag": 6.467103958129883e-06, "ewc_loss_parallel": 4.76837158203125e-06, "grad_norm": 1.0750012397766113, "learning_rate": 1.6617210682492582e-06, "loss": 0.5457, "mean_token_accuracy": 0.8283039331436157, "num_tokens": 14903973.0, "step": 393 }, { "epoch": 0.05012084976466098, "ewc_loss": 0.005462646484375, "ewc_loss_diag": 6.496906280517578e-06, "ewc_loss_parallel": 4.827976226806641e-06, "grad_norm": 1.0894657373428345, "learning_rate": 1.665960152607037e-06, "loss": 0.5806, "mean_token_accuracy": 0.820654571056366, "num_tokens": 14944958.0, "step": 394 }, { "epoch": 0.0502480600432515, "ewc_loss": 0.005584716796875, "ewc_loss_diag": 6.496906280517578e-06, "ewc_loss_parallel": 4.947185516357422e-06, "grad_norm": 1.0883409976959229, "learning_rate": 1.6701992369648157e-06, "loss": 0.5717, "mean_token_accuracy": 0.8200066685676575, "num_tokens": 14985924.0, "step": 395 }, { "epoch": 0.050375270321842004, "ewc_loss": 0.005523681640625, "ewc_loss_diag": 6.496906280517578e-06, "ewc_loss_parallel": 4.887580871582031e-06, "grad_norm": 1.0951515436172485, "learning_rate": 1.6744383213225946e-06, "loss": 0.5373, "mean_token_accuracy": 0.8291347026824951, "num_tokens": 15022346.0, "step": 396 }, { "epoch": 0.05050248060043251, "ewc_loss": 0.005645751953125, "ewc_loss_diag": 6.5267086029052734e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 1.065243124961853, "learning_rate": 1.678677405680373e-06, "loss": 0.489, "mean_token_accuracy": 0.8431533575057983, "num_tokens": 15058656.0, "step": 397 }, { "epoch": 0.050629690879023026, "ewc_loss": 0.005523681640625, "ewc_loss_diag": 6.5267086029052734e-06, "ewc_loss_parallel": 4.887580871582031e-06, "grad_norm": 1.073792576789856, "learning_rate": 1.682916490038152e-06, "loss": 0.5238, "mean_token_accuracy": 0.832179605960846, "num_tokens": 15094165.0, "step": 398 }, { "epoch": 0.05075690115761353, "ewc_loss": 0.00567626953125, "ewc_loss_diag": 6.556510925292969e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 1.0476583242416382, "learning_rate": 1.6871555743959306e-06, "loss": 0.505, "mean_token_accuracy": 0.8385089635848999, "num_tokens": 15136979.0, "step": 399 }, { "epoch": 0.05088411143620405, "ewc_loss": 0.005523681640625, "ewc_loss_diag": 6.586313247680664e-06, "ewc_loss_parallel": 4.857778549194336e-06, "grad_norm": 1.0647368431091309, "learning_rate": 1.6913946587537095e-06, "loss": 0.5307, "mean_token_accuracy": 0.8329227566719055, "num_tokens": 15178765.0, "step": 400 }, { "epoch": 0.051011321714794555, "ewc_loss": 0.00567626953125, "ewc_loss_diag": 6.616115570068359e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 1.0953251123428345, "learning_rate": 1.695633743111488e-06, "loss": 0.5324, "mean_token_accuracy": 0.8345491290092468, "num_tokens": 15215892.0, "step": 401 }, { "epoch": 0.05113853199338506, "ewc_loss": 0.005706787109375, "ewc_loss_diag": 6.645917892456055e-06, "ewc_loss_parallel": 5.036592483520508e-06, "grad_norm": 1.1190946102142334, "learning_rate": 1.6998728274692669e-06, "loss": 0.5041, "mean_token_accuracy": 0.8432697057723999, "num_tokens": 15248693.0, "step": 402 }, { "epoch": 0.05126574227197558, "ewc_loss": 0.0057373046875, "ewc_loss_diag": 6.645917892456055e-06, "ewc_loss_parallel": 5.066394805908203e-06, "grad_norm": 1.0815939903259277, "learning_rate": 1.7041119118270455e-06, "loss": 0.529, "mean_token_accuracy": 0.8348143100738525, "num_tokens": 15289476.0, "step": 403 }, { "epoch": 0.051392952550566084, "ewc_loss": 0.00567626953125, "ewc_loss_diag": 6.705522537231445e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 1.118682861328125, "learning_rate": 1.708350996184824e-06, "loss": 0.5665, "mean_token_accuracy": 0.8217738270759583, "num_tokens": 15323032.0, "step": 404 }, { "epoch": 0.0515201628291566, "ewc_loss": 0.0057373046875, "ewc_loss_diag": 6.705522537231445e-06, "ewc_loss_parallel": 5.066394805908203e-06, "grad_norm": 1.0778601169586182, "learning_rate": 1.7125900805426029e-06, "loss": 0.5967, "mean_token_accuracy": 0.813615620136261, "num_tokens": 15369155.0, "step": 405 }, { "epoch": 0.051647373107747106, "ewc_loss": 0.005645751953125, "ewc_loss_diag": 6.735324859619141e-06, "ewc_loss_parallel": 4.976987838745117e-06, "grad_norm": 1.1169284582138062, "learning_rate": 1.7168291649003816e-06, "loss": 0.5138, "mean_token_accuracy": 0.8343896269798279, "num_tokens": 15402623.0, "step": 406 }, { "epoch": 0.051774583386337614, "ewc_loss": 0.0057373046875, "ewc_loss_diag": 6.735324859619141e-06, "ewc_loss_parallel": 5.066394805908203e-06, "grad_norm": 1.1012468338012695, "learning_rate": 1.7210682492581604e-06, "loss": 0.5524, "mean_token_accuracy": 0.8277585506439209, "num_tokens": 15440485.0, "step": 407 }, { "epoch": 0.05190179366492813, "ewc_loss": 0.00567626953125, "ewc_loss_diag": 6.735324859619141e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 1.0906946659088135, "learning_rate": 1.725307333615939e-06, "loss": 0.5226, "mean_token_accuracy": 0.8342708349227905, "num_tokens": 15479641.0, "step": 408 }, { "epoch": 0.052029003943518635, "ewc_loss": 0.00567626953125, "ewc_loss_diag": 6.735324859619141e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 1.104824185371399, "learning_rate": 1.7295464179737178e-06, "loss": 0.5642, "mean_token_accuracy": 0.8244210481643677, "num_tokens": 15519085.0, "step": 409 }, { "epoch": 0.05215621422210915, "ewc_loss": 0.005706787109375, "ewc_loss_diag": 6.765127182006836e-06, "ewc_loss_parallel": 5.036592483520508e-06, "grad_norm": 1.0822315216064453, "learning_rate": 1.7337855023314965e-06, "loss": 0.4984, "mean_token_accuracy": 0.8425167202949524, "num_tokens": 15557642.0, "step": 410 }, { "epoch": 0.05228342450069966, "ewc_loss": 0.00567626953125, "ewc_loss_diag": 6.765127182006836e-06, "ewc_loss_parallel": 5.0067901611328125e-06, "grad_norm": 1.0908095836639404, "learning_rate": 1.7380245866892753e-06, "loss": 0.4987, "mean_token_accuracy": 0.8440595269203186, "num_tokens": 15593178.0, "step": 411 }, { "epoch": 0.052410634779290165, "ewc_loss": 0.0057373046875, "ewc_loss_diag": 6.8247318267822266e-06, "ewc_loss_parallel": 5.066394805908203e-06, "grad_norm": 1.0976417064666748, "learning_rate": 1.7422636710470538e-06, "loss": 0.5536, "mean_token_accuracy": 0.8254343271255493, "num_tokens": 15630167.0, "step": 412 }, { "epoch": 0.05253784505788068, "ewc_loss": 0.0057373046875, "ewc_loss_diag": 6.854534149169922e-06, "ewc_loss_parallel": 5.066394805908203e-06, "grad_norm": 1.092288851737976, "learning_rate": 1.7465027554048327e-06, "loss": 0.5844, "mean_token_accuracy": 0.819188117980957, "num_tokens": 15669380.0, "step": 413 }, { "epoch": 0.05266505533647119, "ewc_loss": 0.0057373046875, "ewc_loss_diag": 6.884336471557617e-06, "ewc_loss_parallel": 5.036592483520508e-06, "grad_norm": 1.1319494247436523, "learning_rate": 1.7507418397626114e-06, "loss": 0.5261, "mean_token_accuracy": 0.8357117772102356, "num_tokens": 15701791.0, "step": 414 }, { "epoch": 0.052792265615061694, "ewc_loss": 0.00579833984375, "ewc_loss_diag": 6.884336471557617e-06, "ewc_loss_parallel": 5.125999450683594e-06, "grad_norm": 1.091526985168457, "learning_rate": 1.7549809241203902e-06, "loss": 0.5015, "mean_token_accuracy": 0.8407172560691833, "num_tokens": 15738223.0, "step": 415 }, { "epoch": 0.05291947589365221, "ewc_loss": 0.005767822265625, "ewc_loss_diag": 6.943941116333008e-06, "ewc_loss_parallel": 5.066394805908203e-06, "grad_norm": 1.1161727905273438, "learning_rate": 1.759220008478169e-06, "loss": 0.583, "mean_token_accuracy": 0.8192429542541504, "num_tokens": 15773159.0, "step": 416 }, { "epoch": 0.053046686172242716, "ewc_loss": 0.005889892578125, "ewc_loss_diag": 6.943941116333008e-06, "ewc_loss_parallel": 5.185604095458984e-06, "grad_norm": 1.099105715751648, "learning_rate": 1.7634590928359476e-06, "loss": 0.549, "mean_token_accuracy": 0.8284467458724976, "num_tokens": 15813132.0, "step": 417 }, { "epoch": 0.05317389645083323, "ewc_loss": 0.005859375, "ewc_loss_diag": 6.943941116333008e-06, "ewc_loss_parallel": 5.155801773071289e-06, "grad_norm": 1.0908229351043701, "learning_rate": 1.7676981771937263e-06, "loss": 0.5299, "mean_token_accuracy": 0.8313829898834229, "num_tokens": 15849752.0, "step": 418 }, { "epoch": 0.05330110672942374, "ewc_loss": 0.00592041015625, "ewc_loss_diag": 6.943941116333008e-06, "ewc_loss_parallel": 5.21540641784668e-06, "grad_norm": 1.124315619468689, "learning_rate": 1.7719372615515052e-06, "loss": 0.5094, "mean_token_accuracy": 0.841381311416626, "num_tokens": 15885331.0, "step": 419 }, { "epoch": 0.053428317008014245, "ewc_loss": 0.005950927734375, "ewc_loss_diag": 6.973743438720703e-06, "ewc_loss_parallel": 5.245208740234375e-06, "grad_norm": 1.1490256786346436, "learning_rate": 1.7761763459092838e-06, "loss": 0.5502, "mean_token_accuracy": 0.8316095471382141, "num_tokens": 15924649.0, "step": 420 }, { "epoch": 0.05355552728660476, "ewc_loss": 0.0059814453125, "ewc_loss_diag": 7.0035457611083984e-06, "ewc_loss_parallel": 5.27501106262207e-06, "grad_norm": 1.1248985528945923, "learning_rate": 1.7804154302670625e-06, "loss": 0.5197, "mean_token_accuracy": 0.8353770971298218, "num_tokens": 15959342.0, "step": 421 }, { "epoch": 0.05368273756519527, "ewc_loss": 0.0059814453125, "ewc_loss_diag": 7.033348083496094e-06, "ewc_loss_parallel": 5.27501106262207e-06, "grad_norm": 1.0853327512741089, "learning_rate": 1.7846545146248412e-06, "loss": 0.507, "mean_token_accuracy": 0.8395007252693176, "num_tokens": 15998479.0, "step": 422 }, { "epoch": 0.05380994784378578, "ewc_loss": 0.005889892578125, "ewc_loss_diag": 7.063150405883789e-06, "ewc_loss_parallel": 5.185604095458984e-06, "grad_norm": 1.1213886737823486, "learning_rate": 1.7888935989826198e-06, "loss": 0.5504, "mean_token_accuracy": 0.8256288170814514, "num_tokens": 16034197.0, "step": 423 }, { "epoch": 0.05393715812237629, "ewc_loss": 0.0059814453125, "ewc_loss_diag": 7.063150405883789e-06, "ewc_loss_parallel": 5.27501106262207e-06, "grad_norm": 1.126273274421692, "learning_rate": 1.7931326833403987e-06, "loss": 0.5637, "mean_token_accuracy": 0.8215456008911133, "num_tokens": 16071448.0, "step": 424 }, { "epoch": 0.054064368400966796, "ewc_loss": 0.00604248046875, "ewc_loss_diag": 7.063150405883789e-06, "ewc_loss_parallel": 5.334615707397461e-06, "grad_norm": 1.0594338178634644, "learning_rate": 1.7973717676981772e-06, "loss": 0.4879, "mean_token_accuracy": 0.8483268022537231, "num_tokens": 16116007.0, "step": 425 }, { "epoch": 0.05419157867955731, "ewc_loss": 0.005889892578125, "ewc_loss_diag": 7.092952728271484e-06, "ewc_loss_parallel": 5.185604095458984e-06, "grad_norm": 1.0803993940353394, "learning_rate": 1.801610852055956e-06, "loss": 0.5236, "mean_token_accuracy": 0.8370506763458252, "num_tokens": 16158070.0, "step": 426 }, { "epoch": 0.05431878895814782, "ewc_loss": 0.006011962890625, "ewc_loss_diag": 7.092952728271484e-06, "ewc_loss_parallel": 5.304813385009766e-06, "grad_norm": 1.1145654916763306, "learning_rate": 1.8058499364137347e-06, "loss": 0.5639, "mean_token_accuracy": 0.8263896703720093, "num_tokens": 16196310.0, "step": 427 }, { "epoch": 0.054445999236738325, "ewc_loss": 0.006072998046875, "ewc_loss_diag": 7.092952728271484e-06, "ewc_loss_parallel": 5.364418029785156e-06, "grad_norm": 1.1858123540878296, "learning_rate": 1.8100890207715136e-06, "loss": 0.6109, "mean_token_accuracy": 0.8066779375076294, "num_tokens": 16237927.0, "step": 428 }, { "epoch": 0.05457320951532884, "ewc_loss": 0.0062255859375, "ewc_loss_diag": 7.152557373046875e-06, "ewc_loss_parallel": 5.4836273193359375e-06, "grad_norm": 1.1290242671966553, "learning_rate": 1.814328105129292e-06, "loss": 0.5652, "mean_token_accuracy": 0.8209969401359558, "num_tokens": 16274865.0, "step": 429 }, { "epoch": 0.05470041979391935, "ewc_loss": 0.00604248046875, "ewc_loss_diag": 7.18235969543457e-06, "ewc_loss_parallel": 5.304813385009766e-06, "grad_norm": 1.1833868026733398, "learning_rate": 1.818567189487071e-06, "loss": 0.542, "mean_token_accuracy": 0.8249776363372803, "num_tokens": 16302503.0, "step": 430 }, { "epoch": 0.05482763007250986, "ewc_loss": 0.00628662109375, "ewc_loss_diag": 7.18235969543457e-06, "ewc_loss_parallel": 5.543231964111328e-06, "grad_norm": 1.1296911239624023, "learning_rate": 1.8228062738448496e-06, "loss": 0.5966, "mean_token_accuracy": 0.814661979675293, "num_tokens": 16342472.0, "step": 431 }, { "epoch": 0.05495484035110037, "ewc_loss": 0.00604248046875, "ewc_loss_diag": 7.212162017822266e-06, "ewc_loss_parallel": 5.304813385009766e-06, "grad_norm": 1.1073509454727173, "learning_rate": 1.8270453582026285e-06, "loss": 0.5564, "mean_token_accuracy": 0.8248463869094849, "num_tokens": 16378423.0, "step": 432 }, { "epoch": 0.055082050629690876, "ewc_loss": 0.00604248046875, "ewc_loss_diag": 7.212162017822266e-06, "ewc_loss_parallel": 5.304813385009766e-06, "grad_norm": 1.1021630764007568, "learning_rate": 1.831284442560407e-06, "loss": 0.5257, "mean_token_accuracy": 0.835792064666748, "num_tokens": 16418430.0, "step": 433 }, { "epoch": 0.05520926090828139, "ewc_loss": 0.006011962890625, "ewc_loss_diag": 7.212162017822266e-06, "ewc_loss_parallel": 5.27501106262207e-06, "grad_norm": 1.1097413301467896, "learning_rate": 1.8355235269181859e-06, "loss": 0.5381, "mean_token_accuracy": 0.830958366394043, "num_tokens": 16460984.0, "step": 434 }, { "epoch": 0.0553364711868719, "ewc_loss": 0.006134033203125, "ewc_loss_diag": 7.241964340209961e-06, "ewc_loss_parallel": 5.3942203521728516e-06, "grad_norm": 1.1402498483657837, "learning_rate": 1.8397626112759646e-06, "loss": 0.5162, "mean_token_accuracy": 0.8374260663986206, "num_tokens": 16493460.0, "step": 435 }, { "epoch": 0.05546368146546241, "ewc_loss": 0.006256103515625, "ewc_loss_diag": 7.241964340209961e-06, "ewc_loss_parallel": 5.513429641723633e-06, "grad_norm": 1.1232085227966309, "learning_rate": 1.8440016956337434e-06, "loss": 0.4942, "mean_token_accuracy": 0.8424553871154785, "num_tokens": 16528998.0, "step": 436 }, { "epoch": 0.05559089174405292, "ewc_loss": 0.006195068359375, "ewc_loss_diag": 7.271766662597656e-06, "ewc_loss_parallel": 5.453824996948242e-06, "grad_norm": 1.130642294883728, "learning_rate": 1.848240779991522e-06, "loss": 0.5196, "mean_token_accuracy": 0.8356980085372925, "num_tokens": 16564041.0, "step": 437 }, { "epoch": 0.05571810202264343, "ewc_loss": 0.006256103515625, "ewc_loss_diag": 7.271766662597656e-06, "ewc_loss_parallel": 5.513429641723633e-06, "grad_norm": 1.1764568090438843, "learning_rate": 1.8524798643493008e-06, "loss": 0.593, "mean_token_accuracy": 0.8126650452613831, "num_tokens": 16597967.0, "step": 438 }, { "epoch": 0.05584531230123394, "ewc_loss": 0.006378173828125, "ewc_loss_diag": 7.271766662597656e-06, "ewc_loss_parallel": 5.632638931274414e-06, "grad_norm": 1.1336714029312134, "learning_rate": 1.8567189487070795e-06, "loss": 0.5773, "mean_token_accuracy": 0.8214571475982666, "num_tokens": 16636756.0, "step": 439 }, { "epoch": 0.05597252257982445, "ewc_loss": 0.0062255859375, "ewc_loss_diag": 7.271766662597656e-06, "ewc_loss_parallel": 5.4836273193359375e-06, "grad_norm": 1.0954055786132812, "learning_rate": 1.860958033064858e-06, "loss": 0.4899, "mean_token_accuracy": 0.843441367149353, "num_tokens": 16675811.0, "step": 440 }, { "epoch": 0.05609973285841496, "ewc_loss": 0.0062255859375, "ewc_loss_diag": 7.331371307373047e-06, "ewc_loss_parallel": 5.4836273193359375e-06, "grad_norm": 1.1451342105865479, "learning_rate": 1.8651971174226368e-06, "loss": 0.5738, "mean_token_accuracy": 0.8231765031814575, "num_tokens": 16713714.0, "step": 441 }, { "epoch": 0.05622694313700547, "ewc_loss": 0.00634765625, "ewc_loss_diag": 7.331371307373047e-06, "ewc_loss_parallel": 5.602836608886719e-06, "grad_norm": 1.1185109615325928, "learning_rate": 1.8694362017804155e-06, "loss": 0.5296, "mean_token_accuracy": 0.8326228857040405, "num_tokens": 16754317.0, "step": 442 }, { "epoch": 0.05635415341559598, "ewc_loss": 0.0062255859375, "ewc_loss_diag": 7.331371307373047e-06, "ewc_loss_parallel": 5.4836273193359375e-06, "grad_norm": 1.1094735860824585, "learning_rate": 1.8736752861381944e-06, "loss": 0.4909, "mean_token_accuracy": 0.8453400135040283, "num_tokens": 16793189.0, "step": 443 }, { "epoch": 0.05648136369418649, "ewc_loss": 0.00628662109375, "ewc_loss_diag": 7.331371307373047e-06, "ewc_loss_parallel": 5.543231964111328e-06, "grad_norm": 1.1407040357589722, "learning_rate": 1.8779143704959728e-06, "loss": 0.5004, "mean_token_accuracy": 0.8410964608192444, "num_tokens": 16825046.0, "step": 444 }, { "epoch": 0.056608573972777, "ewc_loss": 0.006378173828125, "ewc_loss_diag": 7.331371307373047e-06, "ewc_loss_parallel": 5.632638931274414e-06, "grad_norm": 1.169202446937561, "learning_rate": 1.8821534548537517e-06, "loss": 0.5629, "mean_token_accuracy": 0.8243550062179565, "num_tokens": 16857769.0, "step": 445 }, { "epoch": 0.05673578425136751, "ewc_loss": 0.0064697265625, "ewc_loss_diag": 7.331371307373047e-06, "ewc_loss_parallel": 5.7220458984375e-06, "grad_norm": 1.1251438856124878, "learning_rate": 1.8863925392115304e-06, "loss": 0.5212, "mean_token_accuracy": 0.8370568752288818, "num_tokens": 16900878.0, "step": 446 }, { "epoch": 0.05686299452995802, "ewc_loss": 0.00634765625, "ewc_loss_diag": 7.331371307373047e-06, "ewc_loss_parallel": 5.602836608886719e-06, "grad_norm": 1.10400390625, "learning_rate": 1.8906316235693093e-06, "loss": 0.5356, "mean_token_accuracy": 0.8298655152320862, "num_tokens": 16942832.0, "step": 447 }, { "epoch": 0.05699020480854853, "ewc_loss": 0.00634765625, "ewc_loss_diag": 7.3909759521484375e-06, "ewc_loss_parallel": 5.602836608886719e-06, "grad_norm": 1.1534947156906128, "learning_rate": 1.8948707079270877e-06, "loss": 0.5389, "mean_token_accuracy": 0.8240237832069397, "num_tokens": 16975609.0, "step": 448 }, { "epoch": 0.057117415087139044, "ewc_loss": 0.0064697265625, "ewc_loss_diag": 7.3909759521484375e-06, "ewc_loss_parallel": 5.751848220825195e-06, "grad_norm": 1.1321812868118286, "learning_rate": 1.8991097922848666e-06, "loss": 0.51, "mean_token_accuracy": 0.8385940194129944, "num_tokens": 17012620.0, "step": 449 }, { "epoch": 0.05724462536572955, "ewc_loss": 0.00640869140625, "ewc_loss_diag": 7.3909759521484375e-06, "ewc_loss_parallel": 5.662441253662109e-06, "grad_norm": 1.1305547952651978, "learning_rate": 1.9033488766426453e-06, "loss": 0.5546, "mean_token_accuracy": 0.8226971626281738, "num_tokens": 17053820.0, "step": 450 }, { "epoch": 0.05737183564432006, "ewc_loss": 0.006500244140625, "ewc_loss_diag": 7.3909759521484375e-06, "ewc_loss_parallel": 5.781650543212891e-06, "grad_norm": 1.1393544673919678, "learning_rate": 1.907587961000424e-06, "loss": 0.5083, "mean_token_accuracy": 0.8426564931869507, "num_tokens": 17091885.0, "step": 451 }, { "epoch": 0.05749904592291057, "ewc_loss": 0.0064697265625, "ewc_loss_diag": 7.420778274536133e-06, "ewc_loss_parallel": 5.751848220825195e-06, "grad_norm": 1.12924325466156, "learning_rate": 1.911827045358203e-06, "loss": 0.5551, "mean_token_accuracy": 0.8265621662139893, "num_tokens": 17133432.0, "step": 452 }, { "epoch": 0.05762625620150108, "ewc_loss": 0.0064697265625, "ewc_loss_diag": 7.4803829193115234e-06, "ewc_loss_parallel": 5.7220458984375e-06, "grad_norm": 1.1357510089874268, "learning_rate": 1.9160661297159817e-06, "loss": 0.5769, "mean_token_accuracy": 0.8215678930282593, "num_tokens": 17171845.0, "step": 453 }, { "epoch": 0.057753466480091595, "ewc_loss": 0.00653076171875, "ewc_loss_diag": 7.4803829193115234e-06, "ewc_loss_parallel": 5.781650543212891e-06, "grad_norm": 1.1624828577041626, "learning_rate": 1.92030521407376e-06, "loss": 0.5664, "mean_token_accuracy": 0.8200667500495911, "num_tokens": 17206137.0, "step": 454 }, { "epoch": 0.0578806767586821, "ewc_loss": 0.006591796875, "ewc_loss_diag": 7.510185241699219e-06, "ewc_loss_parallel": 5.841255187988281e-06, "grad_norm": 1.1215829849243164, "learning_rate": 1.924544298431539e-06, "loss": 0.5464, "mean_token_accuracy": 0.8318749070167542, "num_tokens": 17247619.0, "step": 455 }, { "epoch": 0.05800788703727261, "ewc_loss": 0.006500244140625, "ewc_loss_diag": 7.510185241699219e-06, "ewc_loss_parallel": 5.751848220825195e-06, "grad_norm": 1.1805340051651, "learning_rate": 1.9287833827893175e-06, "loss": 0.5683, "mean_token_accuracy": 0.8211472034454346, "num_tokens": 17285053.0, "step": 456 }, { "epoch": 0.058135097315863124, "ewc_loss": 0.0067138671875, "ewc_loss_diag": 7.510185241699219e-06, "ewc_loss_parallel": 5.9604644775390625e-06, "grad_norm": 1.1747032403945923, "learning_rate": 1.9330224671470964e-06, "loss": 0.5426, "mean_token_accuracy": 0.827914834022522, "num_tokens": 17320303.0, "step": 457 }, { "epoch": 0.05826230759445363, "ewc_loss": 0.006561279296875, "ewc_loss_diag": 7.569789886474609e-06, "ewc_loss_parallel": 5.811452865600586e-06, "grad_norm": 1.197214126586914, "learning_rate": 1.9372615515048753e-06, "loss": 0.5936, "mean_token_accuracy": 0.8189075589179993, "num_tokens": 17354102.0, "step": 458 }, { "epoch": 0.05838951787304414, "ewc_loss": 0.0067138671875, "ewc_loss_diag": 7.62939453125e-06, "ewc_loss_parallel": 5.9604644775390625e-06, "grad_norm": 1.1569063663482666, "learning_rate": 1.9415006358626538e-06, "loss": 0.5333, "mean_token_accuracy": 0.8324663639068604, "num_tokens": 17389958.0, "step": 459 }, { "epoch": 0.05851672815163465, "ewc_loss": 0.006561279296875, "ewc_loss_diag": 7.62939453125e-06, "ewc_loss_parallel": 5.811452865600586e-06, "grad_norm": 1.1812694072723389, "learning_rate": 1.9457397202204326e-06, "loss": 0.5203, "mean_token_accuracy": 0.8358300924301147, "num_tokens": 17421488.0, "step": 460 }, { "epoch": 0.05864393843022516, "ewc_loss": 0.0067138671875, "ewc_loss_diag": 7.68899917602539e-06, "ewc_loss_parallel": 5.9604644775390625e-06, "grad_norm": 1.149168610572815, "learning_rate": 1.949978804578211e-06, "loss": 0.4831, "mean_token_accuracy": 0.8487021923065186, "num_tokens": 17455172.0, "step": 461 }, { "epoch": 0.058771148708815675, "ewc_loss": 0.006561279296875, "ewc_loss_diag": 7.748603820800781e-06, "ewc_loss_parallel": 5.811452865600586e-06, "grad_norm": 1.1391464471817017, "learning_rate": 1.95421788893599e-06, "loss": 0.5811, "mean_token_accuracy": 0.8170015811920166, "num_tokens": 17494798.0, "step": 462 }, { "epoch": 0.05889835898740618, "ewc_loss": 0.006744384765625, "ewc_loss_diag": 7.808208465576172e-06, "ewc_loss_parallel": 5.9604644775390625e-06, "grad_norm": 1.233253836631775, "learning_rate": 1.9584569732937684e-06, "loss": 0.5453, "mean_token_accuracy": 0.8337513208389282, "num_tokens": 17527619.0, "step": 463 }, { "epoch": 0.05902556926599669, "ewc_loss": 0.006927490234375, "ewc_loss_diag": 7.808208465576172e-06, "ewc_loss_parallel": 6.139278411865234e-06, "grad_norm": 1.1310021877288818, "learning_rate": 1.9626960576515473e-06, "loss": 0.4647, "mean_token_accuracy": 0.8525805473327637, "num_tokens": 17564299.0, "step": 464 }, { "epoch": 0.059152779544587204, "ewc_loss": 0.006591796875, "ewc_loss_diag": 7.867813110351562e-06, "ewc_loss_parallel": 5.811452865600586e-06, "grad_norm": 1.172690749168396, "learning_rate": 1.9669351420093262e-06, "loss": 0.5795, "mean_token_accuracy": 0.8203551769256592, "num_tokens": 17598333.0, "step": 465 }, { "epoch": 0.05927998982317771, "ewc_loss": 0.006927490234375, "ewc_loss_diag": 7.867813110351562e-06, "ewc_loss_parallel": 6.139278411865234e-06, "grad_norm": 1.1414475440979004, "learning_rate": 1.971174226367105e-06, "loss": 0.5626, "mean_token_accuracy": 0.8268122673034668, "num_tokens": 17642106.0, "step": 466 }, { "epoch": 0.059407200101768226, "ewc_loss": 0.0067138671875, "ewc_loss_diag": 7.867813110351562e-06, "ewc_loss_parallel": 5.930662155151367e-06, "grad_norm": 1.1488441228866577, "learning_rate": 1.9754133107248836e-06, "loss": 0.5044, "mean_token_accuracy": 0.8368871808052063, "num_tokens": 17676412.0, "step": 467 }, { "epoch": 0.059534410380358734, "ewc_loss": 0.006927490234375, "ewc_loss_diag": 7.867813110351562e-06, "ewc_loss_parallel": 6.139278411865234e-06, "grad_norm": 1.176627516746521, "learning_rate": 1.9796523950826624e-06, "loss": 0.5867, "mean_token_accuracy": 0.8183290958404541, "num_tokens": 17717079.0, "step": 468 }, { "epoch": 0.05966162065894924, "ewc_loss": 0.0068359375, "ewc_loss_diag": 7.867813110351562e-06, "ewc_loss_parallel": 6.0498714447021484e-06, "grad_norm": 1.1557457447052002, "learning_rate": 1.983891479440441e-06, "loss": 0.4768, "mean_token_accuracy": 0.8484256267547607, "num_tokens": 17753603.0, "step": 469 }, { "epoch": 0.059788830937539755, "ewc_loss": 0.006805419921875, "ewc_loss_diag": 7.867813110351562e-06, "ewc_loss_parallel": 6.020069122314453e-06, "grad_norm": 1.135987401008606, "learning_rate": 1.98813056379822e-06, "loss": 0.5425, "mean_token_accuracy": 0.8305092453956604, "num_tokens": 17795960.0, "step": 470 }, { "epoch": 0.05991604121613026, "ewc_loss": 0.006927490234375, "ewc_loss_diag": 7.927417755126953e-06, "ewc_loss_parallel": 6.139278411865234e-06, "grad_norm": 1.199168086051941, "learning_rate": 1.9923696481559983e-06, "loss": 0.5602, "mean_token_accuracy": 0.8260580897331238, "num_tokens": 17829298.0, "step": 471 }, { "epoch": 0.06004325149472077, "ewc_loss": 0.0069580078125, "ewc_loss_diag": 7.927417755126953e-06, "ewc_loss_parallel": 6.16908073425293e-06, "grad_norm": 1.160233974456787, "learning_rate": 1.996608732513777e-06, "loss": 0.5959, "mean_token_accuracy": 0.8153480291366577, "num_tokens": 17872027.0, "step": 472 }, { "epoch": 0.060170461773311285, "ewc_loss": 0.006866455078125, "ewc_loss_diag": 7.987022399902344e-06, "ewc_loss_parallel": 6.079673767089844e-06, "grad_norm": 1.1298184394836426, "learning_rate": 2.000847816871556e-06, "loss": 0.483, "mean_token_accuracy": 0.8443681001663208, "num_tokens": 17911846.0, "step": 473 }, { "epoch": 0.06029767205190179, "ewc_loss": 0.006927490234375, "ewc_loss_diag": 7.987022399902344e-06, "ewc_loss_parallel": 6.139278411865234e-06, "grad_norm": 1.1696845293045044, "learning_rate": 2.005086901229335e-06, "loss": 0.479, "mean_token_accuracy": 0.8464686274528503, "num_tokens": 17946095.0, "step": 474 }, { "epoch": 0.060424882330492306, "ewc_loss": 0.0069580078125, "ewc_loss_diag": 7.987022399902344e-06, "ewc_loss_parallel": 6.16908073425293e-06, "grad_norm": 1.1755188703536987, "learning_rate": 2.0093259855871134e-06, "loss": 0.5096, "mean_token_accuracy": 0.8384293913841248, "num_tokens": 17983232.0, "step": 475 }, { "epoch": 0.060552092609082814, "ewc_loss": 0.0069580078125, "ewc_loss_diag": 7.987022399902344e-06, "ewc_loss_parallel": 6.16908073425293e-06, "grad_norm": 1.1735399961471558, "learning_rate": 2.0135650699448923e-06, "loss": 0.599, "mean_token_accuracy": 0.8107380867004395, "num_tokens": 18022455.0, "step": 476 }, { "epoch": 0.06067930288767332, "ewc_loss": 0.006988525390625, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.198883056640625e-06, "grad_norm": 1.1809797286987305, "learning_rate": 2.0178041543026707e-06, "loss": 0.564, "mean_token_accuracy": 0.8241074681282043, "num_tokens": 18059844.0, "step": 477 }, { "epoch": 0.060806513166263836, "ewc_loss": 0.0069580078125, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.16908073425293e-06, "grad_norm": 1.1803064346313477, "learning_rate": 2.022043238660449e-06, "loss": 0.6142, "mean_token_accuracy": 0.8084216117858887, "num_tokens": 18099569.0, "step": 478 }, { "epoch": 0.06093372344485434, "ewc_loss": 0.006927490234375, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.139278411865234e-06, "grad_norm": 1.188612461090088, "learning_rate": 2.026282323018228e-06, "loss": 0.5754, "mean_token_accuracy": 0.8195894360542297, "num_tokens": 18134747.0, "step": 479 }, { "epoch": 0.06106093372344486, "ewc_loss": 0.006988525390625, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.198883056640625e-06, "grad_norm": 1.16542649269104, "learning_rate": 2.030521407376007e-06, "loss": 0.5831, "mean_token_accuracy": 0.8212276697158813, "num_tokens": 18173989.0, "step": 480 }, { "epoch": 0.061188144002035365, "ewc_loss": 0.0069580078125, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.16908073425293e-06, "grad_norm": 1.2197186946868896, "learning_rate": 2.034760491733786e-06, "loss": 0.5553, "mean_token_accuracy": 0.8243454694747925, "num_tokens": 18205065.0, "step": 481 }, { "epoch": 0.06131535428062587, "ewc_loss": 0.007080078125, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.288290023803711e-06, "grad_norm": 1.1568208932876587, "learning_rate": 2.0389995760915643e-06, "loss": 0.5039, "mean_token_accuracy": 0.8413658738136292, "num_tokens": 18242999.0, "step": 482 }, { "epoch": 0.06144256455921639, "ewc_loss": 0.006927490234375, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.139278411865234e-06, "grad_norm": 1.1712092161178589, "learning_rate": 2.043238660449343e-06, "loss": 0.5868, "mean_token_accuracy": 0.8168911933898926, "num_tokens": 18280642.0, "step": 483 }, { "epoch": 0.061569774837806894, "ewc_loss": 0.007110595703125, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.318092346191406e-06, "grad_norm": 1.1930147409439087, "learning_rate": 2.0474777448071216e-06, "loss": 0.5101, "mean_token_accuracy": 0.8407557010650635, "num_tokens": 18317894.0, "step": 484 }, { "epoch": 0.0616969851163974, "ewc_loss": 0.007080078125, "ewc_loss_diag": 8.046627044677734e-06, "ewc_loss_parallel": 6.288290023803711e-06, "grad_norm": 1.185319423675537, "learning_rate": 2.0517168291649005e-06, "loss": 0.5387, "mean_token_accuracy": 0.8301477432250977, "num_tokens": 18350481.0, "step": 485 }, { "epoch": 0.061824195394987916, "ewc_loss": 0.007080078125, "ewc_loss_diag": 8.106231689453125e-06, "ewc_loss_parallel": 6.288290023803711e-06, "grad_norm": 1.2061781883239746, "learning_rate": 2.0559559135226794e-06, "loss": 0.5469, "mean_token_accuracy": 0.8239802122116089, "num_tokens": 18384390.0, "step": 486 }, { "epoch": 0.06195140567357842, "ewc_loss": 0.0072021484375, "ewc_loss_diag": 8.106231689453125e-06, "ewc_loss_parallel": 6.407499313354492e-06, "grad_norm": 1.1504662036895752, "learning_rate": 2.060194997880458e-06, "loss": 0.5067, "mean_token_accuracy": 0.8376708030700684, "num_tokens": 18423958.0, "step": 487 }, { "epoch": 0.06207861595216894, "ewc_loss": 0.007080078125, "ewc_loss_diag": 8.106231689453125e-06, "ewc_loss_parallel": 6.288290023803711e-06, "grad_norm": 1.1609691381454468, "learning_rate": 2.0644340822382368e-06, "loss": 0.4714, "mean_token_accuracy": 0.8502973914146423, "num_tokens": 18459622.0, "step": 488 }, { "epoch": 0.062205826230759445, "ewc_loss": 0.00714111328125, "ewc_loss_diag": 8.106231689453125e-06, "ewc_loss_parallel": 6.3478946685791016e-06, "grad_norm": 1.1852715015411377, "learning_rate": 2.0686731665960156e-06, "loss": 0.5285, "mean_token_accuracy": 0.8365540504455566, "num_tokens": 18495568.0, "step": 489 }, { "epoch": 0.06233303650934995, "ewc_loss": 0.007232666015625, "ewc_loss_diag": 8.165836334228516e-06, "ewc_loss_parallel": 6.407499313354492e-06, "grad_norm": 1.1640570163726807, "learning_rate": 2.072912250953794e-06, "loss": 0.4953, "mean_token_accuracy": 0.8475339412689209, "num_tokens": 18534440.0, "step": 490 }, { "epoch": 0.06246024678794047, "ewc_loss": 0.007232666015625, "ewc_loss_diag": 8.165836334228516e-06, "ewc_loss_parallel": 6.407499313354492e-06, "grad_norm": 1.1847578287124634, "learning_rate": 2.077151335311573e-06, "loss": 0.5309, "mean_token_accuracy": 0.8314359188079834, "num_tokens": 18570583.0, "step": 491 }, { "epoch": 0.06258745706653097, "ewc_loss": 0.00726318359375, "ewc_loss_diag": 8.165836334228516e-06, "ewc_loss_parallel": 6.4373016357421875e-06, "grad_norm": 1.1746845245361328, "learning_rate": 2.0813904196693514e-06, "loss": 0.5095, "mean_token_accuracy": 0.8398951888084412, "num_tokens": 18611801.0, "step": 492 }, { "epoch": 0.06271466734512149, "ewc_loss": 0.007232666015625, "ewc_loss_diag": 8.225440979003906e-06, "ewc_loss_parallel": 6.407499313354492e-06, "grad_norm": 1.1995837688446045, "learning_rate": 2.0856295040271303e-06, "loss": 0.5667, "mean_token_accuracy": 0.8191808462142944, "num_tokens": 18646985.0, "step": 493 }, { "epoch": 0.06284187762371199, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 8.225440979003906e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.18357515335083, "learning_rate": 2.0898685883849092e-06, "loss": 0.5465, "mean_token_accuracy": 0.83231520652771, "num_tokens": 18688818.0, "step": 494 }, { "epoch": 0.0629690879023025, "ewc_loss": 0.00726318359375, "ewc_loss_diag": 8.225440979003906e-06, "ewc_loss_parallel": 6.4373016357421875e-06, "grad_norm": 1.175123929977417, "learning_rate": 2.0941076727426877e-06, "loss": 0.5266, "mean_token_accuracy": 0.8363713026046753, "num_tokens": 18727208.0, "step": 495 }, { "epoch": 0.06309629818089302, "ewc_loss": 0.007293701171875, "ewc_loss_diag": 8.225440979003906e-06, "ewc_loss_parallel": 6.467103958129883e-06, "grad_norm": 1.1449522972106934, "learning_rate": 2.0983467571004666e-06, "loss": 0.5115, "mean_token_accuracy": 0.8382413387298584, "num_tokens": 18772174.0, "step": 496 }, { "epoch": 0.06322350845948353, "ewc_loss": 0.0072021484375, "ewc_loss_diag": 8.225440979003906e-06, "ewc_loss_parallel": 6.377696990966797e-06, "grad_norm": 1.1870312690734863, "learning_rate": 2.102585841458245e-06, "loss": 0.5465, "mean_token_accuracy": 0.8286551237106323, "num_tokens": 18808111.0, "step": 497 }, { "epoch": 0.06335071873807403, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 8.225440979003906e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.2021898031234741, "learning_rate": 2.106824925816024e-06, "loss": 0.5032, "mean_token_accuracy": 0.8350262641906738, "num_tokens": 18841912.0, "step": 498 }, { "epoch": 0.06347792901666455, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 8.225440979003906e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.1970134973526, "learning_rate": 2.1110640101738024e-06, "loss": 0.5132, "mean_token_accuracy": 0.839556097984314, "num_tokens": 18877694.0, "step": 499 }, { "epoch": 0.06360513929525506, "ewc_loss": 0.00732421875, "ewc_loss_diag": 8.344650268554688e-06, "ewc_loss_parallel": 6.496906280517578e-06, "grad_norm": 1.191758394241333, "learning_rate": 2.1153030945315813e-06, "loss": 0.5179, "mean_token_accuracy": 0.8365835547447205, "num_tokens": 18912253.0, "step": 500 }, { "epoch": 0.06373234957384556, "ewc_loss": 0.007354736328125, "ewc_loss_diag": 8.344650268554688e-06, "ewc_loss_parallel": 6.5267086029052734e-06, "grad_norm": 1.2048200368881226, "learning_rate": 2.11954217888936e-06, "loss": 0.5194, "mean_token_accuracy": 0.8379867672920227, "num_tokens": 18945700.0, "step": 501 }, { "epoch": 0.06385955985243608, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 8.344650268554688e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.1838700771331787, "learning_rate": 2.123781263247139e-06, "loss": 0.5442, "mean_token_accuracy": 0.8285140991210938, "num_tokens": 18985260.0, "step": 502 }, { "epoch": 0.06398677013102659, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 8.344650268554688e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.1906933784484863, "learning_rate": 2.1280203476049175e-06, "loss": 0.6149, "mean_token_accuracy": 0.8054941296577454, "num_tokens": 19025364.0, "step": 503 }, { "epoch": 0.06411398040961709, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 8.344650268554688e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.185957431793213, "learning_rate": 2.1322594319626964e-06, "loss": 0.5509, "mean_token_accuracy": 0.8261926174163818, "num_tokens": 19064262.0, "step": 504 }, { "epoch": 0.0642411906882076, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 8.404254913330078e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.249908447265625, "learning_rate": 2.136498516320475e-06, "loss": 0.553, "mean_token_accuracy": 0.8288304805755615, "num_tokens": 19094780.0, "step": 505 }, { "epoch": 0.06436840096679812, "ewc_loss": 0.00750732421875, "ewc_loss_diag": 8.404254913330078e-06, "ewc_loss_parallel": 6.67572021484375e-06, "grad_norm": 1.1995986700057983, "learning_rate": 2.1407376006782537e-06, "loss": 0.4916, "mean_token_accuracy": 0.8440693616867065, "num_tokens": 19131166.0, "step": 506 }, { "epoch": 0.06449561124538863, "ewc_loss": 0.00738525390625, "ewc_loss_diag": 8.404254913330078e-06, "ewc_loss_parallel": 6.5267086029052734e-06, "grad_norm": 1.18338143825531, "learning_rate": 2.144976685036032e-06, "loss": 0.5054, "mean_token_accuracy": 0.8420299887657166, "num_tokens": 19167371.0, "step": 507 }, { "epoch": 0.06462282152397913, "ewc_loss": 0.00750732421875, "ewc_loss_diag": 8.404254913330078e-06, "ewc_loss_parallel": 6.645917892456055e-06, "grad_norm": 1.1722745895385742, "learning_rate": 2.149215769393811e-06, "loss": 0.5556, "mean_token_accuracy": 0.8257688283920288, "num_tokens": 19206751.0, "step": 508 }, { "epoch": 0.06475003180256965, "ewc_loss": 0.007415771484375, "ewc_loss_diag": 8.463859558105469e-06, "ewc_loss_parallel": 6.556510925292969e-06, "grad_norm": 1.194821834564209, "learning_rate": 2.15345485375159e-06, "loss": 0.5109, "mean_token_accuracy": 0.8386257886886597, "num_tokens": 19243227.0, "step": 509 }, { "epoch": 0.06487724208116016, "ewc_loss": 0.00750732421875, "ewc_loss_diag": 8.463859558105469e-06, "ewc_loss_parallel": 6.645917892456055e-06, "grad_norm": 1.1981306076049805, "learning_rate": 2.157693938109369e-06, "loss": 0.5632, "mean_token_accuracy": 0.8264349699020386, "num_tokens": 19285050.0, "step": 510 }, { "epoch": 0.06500445235975066, "ewc_loss": 0.007537841796875, "ewc_loss_diag": 8.52346420288086e-06, "ewc_loss_parallel": 6.67572021484375e-06, "grad_norm": 1.2152884006500244, "learning_rate": 2.1619330224671473e-06, "loss": 0.5475, "mean_token_accuracy": 0.8283106088638306, "num_tokens": 19328780.0, "step": 511 }, { "epoch": 0.06513166263834118, "ewc_loss": 0.00750732421875, "ewc_loss_diag": 8.52346420288086e-06, "ewc_loss_parallel": 6.645917892456055e-06, "grad_norm": 1.1757853031158447, "learning_rate": 2.166172106824926e-06, "loss": 0.5435, "mean_token_accuracy": 0.8323721885681152, "num_tokens": 19373011.0, "step": 512 }, { "epoch": 0.06525887291693169, "ewc_loss": 0.007476806640625, "ewc_loss_diag": 8.52346420288086e-06, "ewc_loss_parallel": 6.616115570068359e-06, "grad_norm": 1.2071826457977295, "learning_rate": 2.1704111911827046e-06, "loss": 0.5338, "mean_token_accuracy": 0.8316461443901062, "num_tokens": 19409641.0, "step": 513 }, { "epoch": 0.0653860831955222, "ewc_loss": 0.007598876953125, "ewc_loss_diag": 8.52346420288086e-06, "ewc_loss_parallel": 6.735324859619141e-06, "grad_norm": 1.1903421878814697, "learning_rate": 2.174650275540483e-06, "loss": 0.5746, "mean_token_accuracy": 0.8235828876495361, "num_tokens": 19451285.0, "step": 514 }, { "epoch": 0.06551329347411271, "ewc_loss": 0.00750732421875, "ewc_loss_diag": 8.52346420288086e-06, "ewc_loss_parallel": 6.645917892456055e-06, "grad_norm": 1.2116763591766357, "learning_rate": 2.178889359898262e-06, "loss": 0.5716, "mean_token_accuracy": 0.8246315717697144, "num_tokens": 19487584.0, "step": 515 }, { "epoch": 0.06564050375270322, "ewc_loss": 0.00762939453125, "ewc_loss_diag": 8.52346420288086e-06, "ewc_loss_parallel": 6.765127182006836e-06, "grad_norm": 1.1910333633422852, "learning_rate": 2.183128444256041e-06, "loss": 0.5804, "mean_token_accuracy": 0.8223904371261597, "num_tokens": 19531200.0, "step": 516 }, { "epoch": 0.06576771403129372, "ewc_loss": 0.00750732421875, "ewc_loss_diag": 8.58306884765625e-06, "ewc_loss_parallel": 6.645917892456055e-06, "grad_norm": 1.189095139503479, "learning_rate": 2.1873675286138197e-06, "loss": 0.5795, "mean_token_accuracy": 0.81584632396698, "num_tokens": 19573126.0, "step": 517 }, { "epoch": 0.06589492430988424, "ewc_loss": 0.007568359375, "ewc_loss_diag": 8.58306884765625e-06, "ewc_loss_parallel": 6.705522537231445e-06, "grad_norm": 1.1953901052474976, "learning_rate": 2.191606612971598e-06, "loss": 0.5939, "mean_token_accuracy": 0.8144067525863647, "num_tokens": 19613499.0, "step": 518 }, { "epoch": 0.06602213458847475, "ewc_loss": 0.007568359375, "ewc_loss_diag": 8.64267349243164e-06, "ewc_loss_parallel": 6.705522537231445e-06, "grad_norm": 1.2041699886322021, "learning_rate": 2.195845697329377e-06, "loss": 0.5834, "mean_token_accuracy": 0.8129454851150513, "num_tokens": 19652459.0, "step": 519 }, { "epoch": 0.06614934486706527, "ewc_loss": 0.00762939453125, "ewc_loss_diag": 8.64267349243164e-06, "ewc_loss_parallel": 6.765127182006836e-06, "grad_norm": 1.1916661262512207, "learning_rate": 2.2000847816871556e-06, "loss": 0.4646, "mean_token_accuracy": 0.8511629104614258, "num_tokens": 19691874.0, "step": 520 }, { "epoch": 0.06627655514565577, "ewc_loss": 0.007598876953125, "ewc_loss_diag": 8.64267349243164e-06, "ewc_loss_parallel": 6.735324859619141e-06, "grad_norm": 1.190144658088684, "learning_rate": 2.2043238660449344e-06, "loss": 0.5138, "mean_token_accuracy": 0.8401572704315186, "num_tokens": 19731424.0, "step": 521 }, { "epoch": 0.06640376542424628, "ewc_loss": 0.007568359375, "ewc_loss_diag": 8.64267349243164e-06, "ewc_loss_parallel": 6.705522537231445e-06, "grad_norm": 1.1942182779312134, "learning_rate": 2.2085629504027133e-06, "loss": 0.5071, "mean_token_accuracy": 0.8380334377288818, "num_tokens": 19768530.0, "step": 522 }, { "epoch": 0.0665309757028368, "ewc_loss": 0.007598876953125, "ewc_loss_diag": 8.64267349243164e-06, "ewc_loss_parallel": 6.735324859619141e-06, "grad_norm": 1.2087528705596924, "learning_rate": 2.2128020347604918e-06, "loss": 0.5419, "mean_token_accuracy": 0.8297892212867737, "num_tokens": 19807836.0, "step": 523 }, { "epoch": 0.0666581859814273, "ewc_loss": 0.00762939453125, "ewc_loss_diag": 8.702278137207031e-06, "ewc_loss_parallel": 6.765127182006836e-06, "grad_norm": 1.1761219501495361, "learning_rate": 2.2170411191182707e-06, "loss": 0.5213, "mean_token_accuracy": 0.83586585521698, "num_tokens": 19854338.0, "step": 524 }, { "epoch": 0.06678539626001781, "ewc_loss": 0.007568359375, "ewc_loss_diag": 8.702278137207031e-06, "ewc_loss_parallel": 6.705522537231445e-06, "grad_norm": 1.2404251098632812, "learning_rate": 2.2212802034760496e-06, "loss": 0.5393, "mean_token_accuracy": 0.8286007642745972, "num_tokens": 19888469.0, "step": 525 }, { "epoch": 0.06691260653860832, "ewc_loss": 0.00775146484375, "ewc_loss_diag": 8.702278137207031e-06, "ewc_loss_parallel": 6.854534149169922e-06, "grad_norm": 1.2723616361618042, "learning_rate": 2.225519287833828e-06, "loss": 0.59, "mean_token_accuracy": 0.8183718919754028, "num_tokens": 19925137.0, "step": 526 }, { "epoch": 0.06703981681719882, "ewc_loss": 0.00775146484375, "ewc_loss_diag": 8.702278137207031e-06, "ewc_loss_parallel": 6.854534149169922e-06, "grad_norm": 1.222827434539795, "learning_rate": 2.229758372191607e-06, "loss": 0.5479, "mean_token_accuracy": 0.8278987407684326, "num_tokens": 19959658.0, "step": 527 }, { "epoch": 0.06716702709578934, "ewc_loss": 0.0076904296875, "ewc_loss_diag": 8.702278137207031e-06, "ewc_loss_parallel": 6.794929504394531e-06, "grad_norm": 1.1976062059402466, "learning_rate": 2.2339974565493854e-06, "loss": 0.506, "mean_token_accuracy": 0.8404070734977722, "num_tokens": 19998702.0, "step": 528 }, { "epoch": 0.06729423737437985, "ewc_loss": 0.0076904296875, "ewc_loss_diag": 8.761882781982422e-06, "ewc_loss_parallel": 6.794929504394531e-06, "grad_norm": 1.2256393432617188, "learning_rate": 2.2382365409071642e-06, "loss": 0.5769, "mean_token_accuracy": 0.8229581117630005, "num_tokens": 20038406.0, "step": 529 }, { "epoch": 0.06742144765297035, "ewc_loss": 0.007781982421875, "ewc_loss_diag": 8.761882781982422e-06, "ewc_loss_parallel": 6.884336471557617e-06, "grad_norm": 1.2335413694381714, "learning_rate": 2.242475625264943e-06, "loss": 0.5321, "mean_token_accuracy": 0.8309231996536255, "num_tokens": 20073185.0, "step": 530 }, { "epoch": 0.06754865793156087, "ewc_loss": 0.00775146484375, "ewc_loss_diag": 8.821487426757812e-06, "ewc_loss_parallel": 6.854534149169922e-06, "grad_norm": 1.2252925634384155, "learning_rate": 2.246714709622722e-06, "loss": 0.514, "mean_token_accuracy": 0.8392977118492126, "num_tokens": 20106828.0, "step": 531 }, { "epoch": 0.06767586821015138, "ewc_loss": 0.0078125, "ewc_loss_diag": 8.821487426757812e-06, "ewc_loss_parallel": 6.9141387939453125e-06, "grad_norm": 1.2595847845077515, "learning_rate": 2.2509537939805005e-06, "loss": 0.5168, "mean_token_accuracy": 0.8353023529052734, "num_tokens": 20137541.0, "step": 532 }, { "epoch": 0.0678030784887419, "ewc_loss": 0.00787353515625, "ewc_loss_diag": 8.821487426757812e-06, "ewc_loss_parallel": 6.973743438720703e-06, "grad_norm": 1.2170428037643433, "learning_rate": 2.255192878338279e-06, "loss": 0.5296, "mean_token_accuracy": 0.8318475484848022, "num_tokens": 20174452.0, "step": 533 }, { "epoch": 0.0679302887673324, "ewc_loss": 0.007781982421875, "ewc_loss_diag": 8.821487426757812e-06, "ewc_loss_parallel": 6.884336471557617e-06, "grad_norm": 1.2173137664794922, "learning_rate": 2.259431962696058e-06, "loss": 0.5461, "mean_token_accuracy": 0.8311675786972046, "num_tokens": 20212195.0, "step": 534 }, { "epoch": 0.06805749904592291, "ewc_loss": 0.00787353515625, "ewc_loss_diag": 8.821487426757812e-06, "ewc_loss_parallel": 6.973743438720703e-06, "grad_norm": 1.2055062055587769, "learning_rate": 2.2636710470538363e-06, "loss": 0.4598, "mean_token_accuracy": 0.8476264476776123, "num_tokens": 20248344.0, "step": 535 }, { "epoch": 0.06818470932451343, "ewc_loss": 0.00787353515625, "ewc_loss_diag": 8.881092071533203e-06, "ewc_loss_parallel": 6.943941116333008e-06, "grad_norm": 1.1969199180603027, "learning_rate": 2.267910131411615e-06, "loss": 0.5045, "mean_token_accuracy": 0.8409780263900757, "num_tokens": 20289874.0, "step": 536 }, { "epoch": 0.06831191960310393, "ewc_loss": 0.0079345703125, "ewc_loss_diag": 8.881092071533203e-06, "ewc_loss_parallel": 7.033348083496094e-06, "grad_norm": 1.2696924209594727, "learning_rate": 2.272149215769394e-06, "loss": 0.516, "mean_token_accuracy": 0.8338481783866882, "num_tokens": 20319582.0, "step": 537 }, { "epoch": 0.06843912988169444, "ewc_loss": 0.00799560546875, "ewc_loss_diag": 8.940696716308594e-06, "ewc_loss_parallel": 7.12275505065918e-06, "grad_norm": 1.224892258644104, "learning_rate": 2.276388300127173e-06, "loss": 0.4948, "mean_token_accuracy": 0.8395019769668579, "num_tokens": 20356944.0, "step": 538 }, { "epoch": 0.06856634016028496, "ewc_loss": 0.0079345703125, "ewc_loss_diag": 9.000301361083984e-06, "ewc_loss_parallel": 7.033348083496094e-06, "grad_norm": 1.196871280670166, "learning_rate": 2.2806273844849514e-06, "loss": 0.5474, "mean_token_accuracy": 0.8246937394142151, "num_tokens": 20398099.0, "step": 539 }, { "epoch": 0.06869355043887546, "ewc_loss": 0.0079345703125, "ewc_loss_diag": 9.000301361083984e-06, "ewc_loss_parallel": 7.063150405883789e-06, "grad_norm": 1.2347036600112915, "learning_rate": 2.2848664688427303e-06, "loss": 0.5194, "mean_token_accuracy": 0.8337950110435486, "num_tokens": 20435155.0, "step": 540 }, { "epoch": 0.06882076071746597, "ewc_loss": 0.00799560546875, "ewc_loss_diag": 9.000301361083984e-06, "ewc_loss_parallel": 7.12275505065918e-06, "grad_norm": 1.2858504056930542, "learning_rate": 2.2891055532005087e-06, "loss": 0.5778, "mean_token_accuracy": 0.8208141326904297, "num_tokens": 20468369.0, "step": 541 }, { "epoch": 0.06894797099605648, "ewc_loss": 0.00811767578125, "ewc_loss_diag": 9.000301361083984e-06, "ewc_loss_parallel": 7.212162017822266e-06, "grad_norm": 1.2793105840682983, "learning_rate": 2.2933446375582876e-06, "loss": 0.5531, "mean_token_accuracy": 0.8303726315498352, "num_tokens": 20500303.0, "step": 542 }, { "epoch": 0.06907518127464699, "ewc_loss": 0.008056640625, "ewc_loss_diag": 9.059906005859375e-06, "ewc_loss_parallel": 7.18235969543457e-06, "grad_norm": 1.2163785696029663, "learning_rate": 2.297583721916066e-06, "loss": 0.5055, "mean_token_accuracy": 0.8428269028663635, "num_tokens": 20541290.0, "step": 543 }, { "epoch": 0.0692023915532375, "ewc_loss": 0.0079345703125, "ewc_loss_diag": 9.059906005859375e-06, "ewc_loss_parallel": 7.033348083496094e-06, "grad_norm": 1.2157491445541382, "learning_rate": 2.301822806273845e-06, "loss": 0.5253, "mean_token_accuracy": 0.8357862234115601, "num_tokens": 20583486.0, "step": 544 }, { "epoch": 0.06932960183182801, "ewc_loss": 0.00799560546875, "ewc_loss_diag": 9.059906005859375e-06, "ewc_loss_parallel": 7.092952728271484e-06, "grad_norm": 1.2636739015579224, "learning_rate": 2.306061890631624e-06, "loss": 0.5772, "mean_token_accuracy": 0.822844922542572, "num_tokens": 20618655.0, "step": 545 }, { "epoch": 0.06945681211041853, "ewc_loss": 0.00811767578125, "ewc_loss_diag": 9.059906005859375e-06, "ewc_loss_parallel": 7.212162017822266e-06, "grad_norm": 1.2309256792068481, "learning_rate": 2.3103009749894027e-06, "loss": 0.5303, "mean_token_accuracy": 0.8335146307945251, "num_tokens": 20655733.0, "step": 546 }, { "epoch": 0.06958402238900903, "ewc_loss": 0.00799560546875, "ewc_loss_diag": 9.059906005859375e-06, "ewc_loss_parallel": 7.12275505065918e-06, "grad_norm": 1.2255239486694336, "learning_rate": 2.314540059347181e-06, "loss": 0.5585, "mean_token_accuracy": 0.8253429532051086, "num_tokens": 20699338.0, "step": 547 }, { "epoch": 0.06971123266759954, "ewc_loss": 0.00799560546875, "ewc_loss_diag": 9.059906005859375e-06, "ewc_loss_parallel": 7.12275505065918e-06, "grad_norm": 1.2709230184555054, "learning_rate": 2.31877914370496e-06, "loss": 0.5829, "mean_token_accuracy": 0.817919135093689, "num_tokens": 20731841.0, "step": 548 }, { "epoch": 0.06983844294619006, "ewc_loss": 0.0081787109375, "ewc_loss_diag": 9.179115295410156e-06, "ewc_loss_parallel": 7.271766662597656e-06, "grad_norm": 1.2717305421829224, "learning_rate": 2.3230182280627385e-06, "loss": 0.5045, "mean_token_accuracy": 0.8413888216018677, "num_tokens": 20761097.0, "step": 549 }, { "epoch": 0.06996565322478056, "ewc_loss": 0.00811767578125, "ewc_loss_diag": 9.179115295410156e-06, "ewc_loss_parallel": 7.18235969543457e-06, "grad_norm": 1.221205472946167, "learning_rate": 2.3272573124205174e-06, "loss": 0.4797, "mean_token_accuracy": 0.8470667600631714, "num_tokens": 20797013.0, "step": 550 }, { "epoch": 0.07009286350337107, "ewc_loss": 0.008056640625, "ewc_loss_diag": 9.179115295410156e-06, "ewc_loss_parallel": 7.152557373046875e-06, "grad_norm": 1.2322288751602173, "learning_rate": 2.331496396778296e-06, "loss": 0.5156, "mean_token_accuracy": 0.8342181444168091, "num_tokens": 20833727.0, "step": 551 }, { "epoch": 0.07022007378196159, "ewc_loss": 0.0081787109375, "ewc_loss_diag": 9.179115295410156e-06, "ewc_loss_parallel": 7.271766662597656e-06, "grad_norm": 1.2135733366012573, "learning_rate": 2.3357354811360748e-06, "loss": 0.5449, "mean_token_accuracy": 0.8313148021697998, "num_tokens": 20877587.0, "step": 552 }, { "epoch": 0.07034728406055209, "ewc_loss": 0.00811767578125, "ewc_loss_diag": 9.179115295410156e-06, "ewc_loss_parallel": 7.212162017822266e-06, "grad_norm": 1.2832129001617432, "learning_rate": 2.3399745654938537e-06, "loss": 0.5538, "mean_token_accuracy": 0.825904130935669, "num_tokens": 20909623.0, "step": 553 }, { "epoch": 0.0704744943391426, "ewc_loss": 0.00836181640625, "ewc_loss_diag": 9.357929229736328e-06, "ewc_loss_parallel": 7.420778274536133e-06, "grad_norm": 1.2566643953323364, "learning_rate": 2.344213649851632e-06, "loss": 0.5353, "mean_token_accuracy": 0.8326981067657471, "num_tokens": 20946348.0, "step": 554 }, { "epoch": 0.07060170461773312, "ewc_loss": 0.00823974609375, "ewc_loss_diag": 9.417533874511719e-06, "ewc_loss_parallel": 7.3015689849853516e-06, "grad_norm": 1.2313501834869385, "learning_rate": 2.348452734209411e-06, "loss": 0.6162, "mean_token_accuracy": 0.8106650114059448, "num_tokens": 20989831.0, "step": 555 }, { "epoch": 0.07072891489632362, "ewc_loss": 0.0081787109375, "ewc_loss_diag": 9.417533874511719e-06, "ewc_loss_parallel": 7.271766662597656e-06, "grad_norm": 1.2547166347503662, "learning_rate": 2.3526918185671895e-06, "loss": 0.564, "mean_token_accuracy": 0.8208783864974976, "num_tokens": 21027007.0, "step": 556 }, { "epoch": 0.07085612517491413, "ewc_loss": 0.00830078125, "ewc_loss_diag": 9.47713851928711e-06, "ewc_loss_parallel": 7.3909759521484375e-06, "grad_norm": 1.2224271297454834, "learning_rate": 2.3569309029249684e-06, "loss": 0.5126, "mean_token_accuracy": 0.8375529646873474, "num_tokens": 21066488.0, "step": 557 }, { "epoch": 0.07098333545350465, "ewc_loss": 0.00830078125, "ewc_loss_diag": 9.59634780883789e-06, "ewc_loss_parallel": 7.331371307373047e-06, "grad_norm": 1.2344733476638794, "learning_rate": 2.3611699872827472e-06, "loss": 0.6023, "mean_token_accuracy": 0.8123571872711182, "num_tokens": 21109049.0, "step": 558 }, { "epoch": 0.07111054573209516, "ewc_loss": 0.00830078125, "ewc_loss_diag": 9.59634780883789e-06, "ewc_loss_parallel": 7.361173629760742e-06, "grad_norm": 1.266397476196289, "learning_rate": 2.3654090716405257e-06, "loss": 0.5452, "mean_token_accuracy": 0.8284667730331421, "num_tokens": 21146661.0, "step": 559 }, { "epoch": 0.07123775601068566, "ewc_loss": 0.0084228515625, "ewc_loss_diag": 9.59634780883789e-06, "ewc_loss_parallel": 7.450580596923828e-06, "grad_norm": 1.2737219333648682, "learning_rate": 2.3696481559983046e-06, "loss": 0.5625, "mean_token_accuracy": 0.821319580078125, "num_tokens": 21179568.0, "step": 560 }, { "epoch": 0.07136496628927617, "ewc_loss": 0.00830078125, "ewc_loss_diag": 9.59634780883789e-06, "ewc_loss_parallel": 7.331371307373047e-06, "grad_norm": 1.3158884048461914, "learning_rate": 2.3738872403560835e-06, "loss": 0.565, "mean_token_accuracy": 0.822073221206665, "num_tokens": 21216535.0, "step": 561 }, { "epoch": 0.07149217656786669, "ewc_loss": 0.0084228515625, "ewc_loss_diag": 9.59634780883789e-06, "ewc_loss_parallel": 7.4803829193115234e-06, "grad_norm": 1.2539188861846924, "learning_rate": 2.378126324713862e-06, "loss": 0.5262, "mean_token_accuracy": 0.8322795629501343, "num_tokens": 21253887.0, "step": 562 }, { "epoch": 0.07161938684645719, "ewc_loss": 0.00823974609375, "ewc_loss_diag": 9.655952453613281e-06, "ewc_loss_parallel": 7.271766662597656e-06, "grad_norm": 1.2457983493804932, "learning_rate": 2.382365409071641e-06, "loss": 0.5307, "mean_token_accuracy": 0.8335877656936646, "num_tokens": 21289813.0, "step": 563 }, { "epoch": 0.0717465971250477, "ewc_loss": 0.00836181640625, "ewc_loss_diag": 9.715557098388672e-06, "ewc_loss_parallel": 7.3909759521484375e-06, "grad_norm": 1.2349185943603516, "learning_rate": 2.3866044934294193e-06, "loss": 0.4831, "mean_token_accuracy": 0.845986008644104, "num_tokens": 21326121.0, "step": 564 }, { "epoch": 0.07187380740363822, "ewc_loss": 0.00836181640625, "ewc_loss_diag": 9.715557098388672e-06, "ewc_loss_parallel": 7.3909759521484375e-06, "grad_norm": 1.2598564624786377, "learning_rate": 2.390843577787198e-06, "loss": 0.5769, "mean_token_accuracy": 0.8184329867362976, "num_tokens": 21363644.0, "step": 565 }, { "epoch": 0.07200101768222872, "ewc_loss": 0.0084228515625, "ewc_loss_diag": 9.715557098388672e-06, "ewc_loss_parallel": 7.450580596923828e-06, "grad_norm": 1.2095763683319092, "learning_rate": 2.395082662144977e-06, "loss": 0.4628, "mean_token_accuracy": 0.8540121912956238, "num_tokens": 21407119.0, "step": 566 }, { "epoch": 0.07212822796081923, "ewc_loss": 0.00836181640625, "ewc_loss_diag": 9.715557098388672e-06, "ewc_loss_parallel": 7.420778274536133e-06, "grad_norm": 1.2498199939727783, "learning_rate": 2.399321746502756e-06, "loss": 0.5372, "mean_token_accuracy": 0.8328130841255188, "num_tokens": 21447648.0, "step": 567 }, { "epoch": 0.07225543823940975, "ewc_loss": 0.008544921875, "ewc_loss_diag": 9.775161743164062e-06, "ewc_loss_parallel": 7.599592208862305e-06, "grad_norm": 1.2754359245300293, "learning_rate": 2.4035608308605344e-06, "loss": 0.5788, "mean_token_accuracy": 0.8267598748207092, "num_tokens": 21488335.0, "step": 568 }, { "epoch": 0.07238264851800025, "ewc_loss": 0.008544921875, "ewc_loss_diag": 9.775161743164062e-06, "ewc_loss_parallel": 7.599592208862305e-06, "grad_norm": 1.2830442190170288, "learning_rate": 2.407799915218313e-06, "loss": 0.5811, "mean_token_accuracy": 0.8159433603286743, "num_tokens": 21525987.0, "step": 569 }, { "epoch": 0.07250985879659076, "ewc_loss": 0.008544921875, "ewc_loss_diag": 9.775161743164062e-06, "ewc_loss_parallel": 7.569789886474609e-06, "grad_norm": 1.2264047861099243, "learning_rate": 2.4120389995760917e-06, "loss": 0.4726, "mean_token_accuracy": 0.8533998727798462, "num_tokens": 21564671.0, "step": 570 }, { "epoch": 0.07263706907518128, "ewc_loss": 0.0084228515625, "ewc_loss_diag": 9.834766387939453e-06, "ewc_loss_parallel": 7.420778274536133e-06, "grad_norm": 1.2573593854904175, "learning_rate": 2.41627808393387e-06, "loss": 0.5191, "mean_token_accuracy": 0.8344061374664307, "num_tokens": 21608590.0, "step": 571 }, { "epoch": 0.07276427935377179, "ewc_loss": 0.00860595703125, "ewc_loss_diag": 9.834766387939453e-06, "ewc_loss_parallel": 7.599592208862305e-06, "grad_norm": 1.2870553731918335, "learning_rate": 2.420517168291649e-06, "loss": 0.5402, "mean_token_accuracy": 0.8341999053955078, "num_tokens": 21641684.0, "step": 572 }, { "epoch": 0.07289148963236229, "ewc_loss": 0.00860595703125, "ewc_loss_diag": 9.894371032714844e-06, "ewc_loss_parallel": 7.599592208862305e-06, "grad_norm": 1.2703187465667725, "learning_rate": 2.424756252649428e-06, "loss": 0.481, "mean_token_accuracy": 0.8462316393852234, "num_tokens": 21674923.0, "step": 573 }, { "epoch": 0.0730186999109528, "ewc_loss": 0.008544921875, "ewc_loss_diag": 9.894371032714844e-06, "ewc_loss_parallel": 7.539987564086914e-06, "grad_norm": 1.2625012397766113, "learning_rate": 2.428995337007207e-06, "loss": 0.5664, "mean_token_accuracy": 0.8249819278717041, "num_tokens": 21714006.0, "step": 574 }, { "epoch": 0.07314591018954332, "ewc_loss": 0.008544921875, "ewc_loss_diag": 9.894371032714844e-06, "ewc_loss_parallel": 7.539987564086914e-06, "grad_norm": 1.2957985401153564, "learning_rate": 2.4332344213649853e-06, "loss": 0.5187, "mean_token_accuracy": 0.841001033782959, "num_tokens": 21750580.0, "step": 575 }, { "epoch": 0.07327312046813382, "ewc_loss": 0.00872802734375, "ewc_loss_diag": 9.894371032714844e-06, "ewc_loss_parallel": 7.748603820800781e-06, "grad_norm": 1.2567129135131836, "learning_rate": 2.437473505722764e-06, "loss": 0.4938, "mean_token_accuracy": 0.8424704670906067, "num_tokens": 21787400.0, "step": 576 }, { "epoch": 0.07340033074672433, "ewc_loss": 0.00848388671875, "ewc_loss_diag": 1.0013580322265625e-05, "ewc_loss_parallel": 7.510185241699219e-06, "grad_norm": 1.2621386051177979, "learning_rate": 2.4417125900805427e-06, "loss": 0.5092, "mean_token_accuracy": 0.8338542580604553, "num_tokens": 21823159.0, "step": 577 }, { "epoch": 0.07352754102531485, "ewc_loss": 0.0087890625, "ewc_loss_diag": 1.0013580322265625e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2808507680892944, "learning_rate": 2.4459516744383215e-06, "loss": 0.5484, "mean_token_accuracy": 0.8293551206588745, "num_tokens": 21858067.0, "step": 578 }, { "epoch": 0.07365475130390535, "ewc_loss": 0.0086669921875, "ewc_loss_diag": 1.0073184967041016e-05, "ewc_loss_parallel": 7.68899917602539e-06, "grad_norm": 1.266114592552185, "learning_rate": 2.4501907587961e-06, "loss": 0.5081, "mean_token_accuracy": 0.8372532725334167, "num_tokens": 21893332.0, "step": 579 }, { "epoch": 0.07378196158249586, "ewc_loss": 0.0087890625, "ewc_loss_diag": 1.0073184967041016e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2845460176467896, "learning_rate": 2.454429843153879e-06, "loss": 0.62, "mean_token_accuracy": 0.8068498969078064, "num_tokens": 21932696.0, "step": 580 }, { "epoch": 0.07390917186108638, "ewc_loss": 0.0087890625, "ewc_loss_diag": 1.0073184967041016e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.301231861114502, "learning_rate": 2.4586689275116578e-06, "loss": 0.5326, "mean_token_accuracy": 0.8257341384887695, "num_tokens": 21964825.0, "step": 581 }, { "epoch": 0.07403638213967688, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0132789611816406e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2952595949172974, "learning_rate": 2.4629080118694367e-06, "loss": 0.4941, "mean_token_accuracy": 0.8430588245391846, "num_tokens": 22003396.0, "step": 582 }, { "epoch": 0.0741635924182674, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0132789611816406e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2421247959136963, "learning_rate": 2.467147096227215e-06, "loss": 0.5014, "mean_token_accuracy": 0.8420153856277466, "num_tokens": 22045107.0, "step": 583 }, { "epoch": 0.07429080269685791, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0192394256591797e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2564489841461182, "learning_rate": 2.471386180584994e-06, "loss": 0.4763, "mean_token_accuracy": 0.8493943214416504, "num_tokens": 22082754.0, "step": 584 }, { "epoch": 0.07441801297544842, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0192394256591797e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2601518630981445, "learning_rate": 2.4756252649427725e-06, "loss": 0.4961, "mean_token_accuracy": 0.8419217467308044, "num_tokens": 22122501.0, "step": 585 }, { "epoch": 0.07454522325403892, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0192394256591797e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.272335171699524, "learning_rate": 2.4798643493005513e-06, "loss": 0.5336, "mean_token_accuracy": 0.8340659141540527, "num_tokens": 22158110.0, "step": 586 }, { "epoch": 0.07467243353262944, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0192394256591797e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.289158582687378, "learning_rate": 2.48410343365833e-06, "loss": 0.4839, "mean_token_accuracy": 0.8399389982223511, "num_tokens": 22190097.0, "step": 587 }, { "epoch": 0.07479964381121995, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0192394256591797e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2894573211669922, "learning_rate": 2.4883425180161087e-06, "loss": 0.559, "mean_token_accuracy": 0.825067400932312, "num_tokens": 22227568.0, "step": 588 }, { "epoch": 0.07492685408981045, "ewc_loss": 0.0089111328125, "ewc_loss_diag": 1.0251998901367188e-05, "ewc_loss_parallel": 7.867813110351562e-06, "grad_norm": 1.2888051271438599, "learning_rate": 2.4925816023738876e-06, "loss": 0.5212, "mean_token_accuracy": 0.8384571075439453, "num_tokens": 22265376.0, "step": 589 }, { "epoch": 0.07505406436840097, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0251998901367188e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.279945969581604, "learning_rate": 2.496820686731666e-06, "loss": 0.5448, "mean_token_accuracy": 0.8311300873756409, "num_tokens": 22310398.0, "step": 590 }, { "epoch": 0.07518127464699148, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0251998901367188e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2970596551895142, "learning_rate": 2.5010597710894445e-06, "loss": 0.4503, "mean_token_accuracy": 0.8564403653144836, "num_tokens": 22341413.0, "step": 591 }, { "epoch": 0.07530848492558198, "ewc_loss": 0.0089111328125, "ewc_loss_diag": 1.0371208190917969e-05, "ewc_loss_parallel": 7.867813110351562e-06, "grad_norm": 1.2646945714950562, "learning_rate": 2.5052988554472234e-06, "loss": 0.5251, "mean_token_accuracy": 0.8342794179916382, "num_tokens": 22383228.0, "step": 592 }, { "epoch": 0.0754356952041725, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.0371208190917969e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2779942750930786, "learning_rate": 2.5095379398050023e-06, "loss": 0.5447, "mean_token_accuracy": 0.8300644755363464, "num_tokens": 22423750.0, "step": 593 }, { "epoch": 0.07556290548276301, "ewc_loss": 0.0089111328125, "ewc_loss_diag": 1.0371208190917969e-05, "ewc_loss_parallel": 7.867813110351562e-06, "grad_norm": 1.312804102897644, "learning_rate": 2.513777024162781e-06, "loss": 0.5014, "mean_token_accuracy": 0.8411520719528198, "num_tokens": 22454717.0, "step": 594 }, { "epoch": 0.07569011576135352, "ewc_loss": 0.00897216796875, "ewc_loss_diag": 1.043081283569336e-05, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.267782211303711, "learning_rate": 2.5180161085205596e-06, "loss": 0.501, "mean_token_accuracy": 0.842558741569519, "num_tokens": 22492798.0, "step": 595 }, { "epoch": 0.07581732603994402, "ewc_loss": 0.00885009765625, "ewc_loss_diag": 1.043081283569336e-05, "ewc_loss_parallel": 7.808208465576172e-06, "grad_norm": 1.2827143669128418, "learning_rate": 2.5222551928783385e-06, "loss": 0.5444, "mean_token_accuracy": 0.8298354148864746, "num_tokens": 22530774.0, "step": 596 }, { "epoch": 0.07594453631853454, "ewc_loss": 0.00897216796875, "ewc_loss_diag": 1.043081283569336e-05, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.2863270044326782, "learning_rate": 2.5264942772361174e-06, "loss": 0.524, "mean_token_accuracy": 0.8350898027420044, "num_tokens": 22566355.0, "step": 597 }, { "epoch": 0.07607174659712505, "ewc_loss": 0.0089111328125, "ewc_loss_diag": 1.043081283569336e-05, "ewc_loss_parallel": 7.867813110351562e-06, "grad_norm": 1.2683539390563965, "learning_rate": 2.5307333615938963e-06, "loss": 0.5092, "mean_token_accuracy": 0.8414084911346436, "num_tokens": 22603633.0, "step": 598 }, { "epoch": 0.07619895687571555, "ewc_loss": 0.00897216796875, "ewc_loss_diag": 1.049041748046875e-05, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.2920130491256714, "learning_rate": 2.5349724459516743e-06, "loss": 0.5383, "mean_token_accuracy": 0.8335409164428711, "num_tokens": 22641481.0, "step": 599 }, { "epoch": 0.07632616715430607, "ewc_loss": 0.00897216796875, "ewc_loss_diag": 1.049041748046875e-05, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.303625226020813, "learning_rate": 2.539211530309453e-06, "loss": 0.4816, "mean_token_accuracy": 0.8471276164054871, "num_tokens": 22672853.0, "step": 600 }, { "epoch": 0.07645337743289658, "ewc_loss": 0.009033203125, "ewc_loss_diag": 1.049041748046875e-05, "ewc_loss_parallel": 7.987022399902344e-06, "grad_norm": 1.31417715549469, "learning_rate": 2.543450614667232e-06, "loss": 0.5549, "mean_token_accuracy": 0.824052095413208, "num_tokens": 22704851.0, "step": 601 }, { "epoch": 0.07658058771148708, "ewc_loss": 0.00897216796875, "ewc_loss_diag": 1.049041748046875e-05, "ewc_loss_parallel": 7.927417755126953e-06, "grad_norm": 1.3655709028244019, "learning_rate": 2.547689699025011e-06, "loss": 0.5737, "mean_token_accuracy": 0.8194698095321655, "num_tokens": 22742578.0, "step": 602 }, { "epoch": 0.0767077979900776, "ewc_loss": 0.009033203125, "ewc_loss_diag": 1.0609626770019531e-05, "ewc_loss_parallel": 7.987022399902344e-06, "grad_norm": 1.2579861879348755, "learning_rate": 2.5519287833827894e-06, "loss": 0.5159, "mean_token_accuracy": 0.8339260816574097, "num_tokens": 22783960.0, "step": 603 }, { "epoch": 0.07683500826866811, "ewc_loss": 0.0089111328125, "ewc_loss_diag": 1.0609626770019531e-05, "ewc_loss_parallel": 7.867813110351562e-06, "grad_norm": 1.313786268234253, "learning_rate": 2.5561678677405683e-06, "loss": 0.4777, "mean_token_accuracy": 0.845673680305481, "num_tokens": 22813643.0, "step": 604 }, { "epoch": 0.07696221854725861, "ewc_loss": 0.00909423828125, "ewc_loss_diag": 1.0609626770019531e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3019623756408691, "learning_rate": 2.560406952098347e-06, "loss": 0.5683, "mean_token_accuracy": 0.8196635246276855, "num_tokens": 22850290.0, "step": 605 }, { "epoch": 0.07708942882584913, "ewc_loss": 0.009033203125, "ewc_loss_diag": 1.0609626770019531e-05, "ewc_loss_parallel": 7.987022399902344e-06, "grad_norm": 1.2848889827728271, "learning_rate": 2.5646460364561252e-06, "loss": 0.5216, "mean_token_accuracy": 0.8326948881149292, "num_tokens": 22894009.0, "step": 606 }, { "epoch": 0.07721663910443964, "ewc_loss": 0.009033203125, "ewc_loss_diag": 1.0609626770019531e-05, "ewc_loss_parallel": 7.987022399902344e-06, "grad_norm": 1.3165175914764404, "learning_rate": 2.568885120813904e-06, "loss": 0.5218, "mean_token_accuracy": 0.8346577286720276, "num_tokens": 22931418.0, "step": 607 }, { "epoch": 0.07734384938303016, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0669231414794922e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.2830877304077148, "learning_rate": 2.573124205171683e-06, "loss": 0.511, "mean_token_accuracy": 0.8412131667137146, "num_tokens": 22970074.0, "step": 608 }, { "epoch": 0.07747105966162066, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0669231414794922e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.2959418296813965, "learning_rate": 2.577363289529462e-06, "loss": 0.5452, "mean_token_accuracy": 0.8311561346054077, "num_tokens": 23008185.0, "step": 609 }, { "epoch": 0.07759826994021117, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0728836059570312e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.2962664365768433, "learning_rate": 2.5816023738872403e-06, "loss": 0.4634, "mean_token_accuracy": 0.8514856100082397, "num_tokens": 23039667.0, "step": 610 }, { "epoch": 0.07772548021880168, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0728836059570312e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3141131401062012, "learning_rate": 2.5858414582450192e-06, "loss": 0.5313, "mean_token_accuracy": 0.833642840385437, "num_tokens": 23075012.0, "step": 611 }, { "epoch": 0.07785269049739219, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0848045349121094e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3044228553771973, "learning_rate": 2.590080542602798e-06, "loss": 0.4639, "mean_token_accuracy": 0.8507577180862427, "num_tokens": 23108397.0, "step": 612 }, { "epoch": 0.0779799007759827, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0848045349121094e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.2950853109359741, "learning_rate": 2.594319626960577e-06, "loss": 0.5403, "mean_token_accuracy": 0.8349791765213013, "num_tokens": 23146419.0, "step": 613 }, { "epoch": 0.07810711105457321, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0848045349121094e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.2641351222991943, "learning_rate": 2.598558711318355e-06, "loss": 0.5349, "mean_token_accuracy": 0.8329520225524902, "num_tokens": 23191750.0, "step": 614 }, { "epoch": 0.07823432133316371, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0967254638671875e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3112733364105225, "learning_rate": 2.602797795676134e-06, "loss": 0.5453, "mean_token_accuracy": 0.8286596536636353, "num_tokens": 23230199.0, "step": 615 }, { "epoch": 0.07836153161175423, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0967254638671875e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3151302337646484, "learning_rate": 2.607036880033913e-06, "loss": 0.5332, "mean_token_accuracy": 0.8258152604103088, "num_tokens": 23264184.0, "step": 616 }, { "epoch": 0.07848874189034474, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 1.0967254638671875e-05, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.3164985179901123, "learning_rate": 2.6112759643916917e-06, "loss": 0.4893, "mean_token_accuracy": 0.8459063768386841, "num_tokens": 23295543.0, "step": 617 }, { "epoch": 0.07861595216893524, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 1.0967254638671875e-05, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.321543574333191, "learning_rate": 2.61551504874947e-06, "loss": 0.5152, "mean_token_accuracy": 0.8400744795799255, "num_tokens": 23328148.0, "step": 618 }, { "epoch": 0.07874316244752576, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.0967254638671875e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.2744100093841553, "learning_rate": 2.619754133107249e-06, "loss": 0.5036, "mean_token_accuracy": 0.8401517868041992, "num_tokens": 23369441.0, "step": 619 }, { "epoch": 0.07887037272611627, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.1026859283447266e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3399211168289185, "learning_rate": 2.623993217465028e-06, "loss": 0.5626, "mean_token_accuracy": 0.8214758634567261, "num_tokens": 23402977.0, "step": 620 }, { "epoch": 0.07899758300470679, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 1.1086463928222656e-05, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.3127052783966064, "learning_rate": 2.628232301822807e-06, "loss": 0.5384, "mean_token_accuracy": 0.8339815139770508, "num_tokens": 23441988.0, "step": 621 }, { "epoch": 0.07912479328329729, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.1086463928222656e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3038349151611328, "learning_rate": 2.6324713861805853e-06, "loss": 0.5675, "mean_token_accuracy": 0.8217333555221558, "num_tokens": 23480481.0, "step": 622 }, { "epoch": 0.0792520035618878, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 1.1086463928222656e-05, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.354404091835022, "learning_rate": 2.6367104705383637e-06, "loss": 0.5645, "mean_token_accuracy": 0.8203246593475342, "num_tokens": 23515188.0, "step": 623 }, { "epoch": 0.07937921384047832, "ewc_loss": 0.00921630859375, "ewc_loss_diag": 1.1086463928222656e-05, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.294900894165039, "learning_rate": 2.6409495548961426e-06, "loss": 0.5512, "mean_token_accuracy": 0.8283141851425171, "num_tokens": 23557777.0, "step": 624 }, { "epoch": 0.07950642411906882, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.1205673217773438e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3490928411483765, "learning_rate": 2.645188639253921e-06, "loss": 0.5395, "mean_token_accuracy": 0.8352138996124268, "num_tokens": 23588357.0, "step": 625 }, { "epoch": 0.07963363439765933, "ewc_loss": 0.00933837890625, "ewc_loss_diag": 1.1205673217773438e-05, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.2965664863586426, "learning_rate": 2.6494277236117e-06, "loss": 0.5012, "mean_token_accuracy": 0.8380728960037231, "num_tokens": 23628656.0, "step": 626 }, { "epoch": 0.07976084467624985, "ewc_loss": 0.0091552734375, "ewc_loss_diag": 1.1205673217773438e-05, "ewc_loss_parallel": 8.046627044677734e-06, "grad_norm": 1.3311687707901, "learning_rate": 2.653666807969479e-06, "loss": 0.6001, "mean_token_accuracy": 0.8125864863395691, "num_tokens": 23666390.0, "step": 627 }, { "epoch": 0.07988805495484035, "ewc_loss": 0.00933837890625, "ewc_loss_diag": 1.1205673217773438e-05, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.3232450485229492, "learning_rate": 2.6579058923272577e-06, "loss": 0.4589, "mean_token_accuracy": 0.8566313982009888, "num_tokens": 23699620.0, "step": 628 }, { "epoch": 0.08001526523343086, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1265277862548828e-05, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.3181533813476562, "learning_rate": 2.662144976685036e-06, "loss": 0.5402, "mean_token_accuracy": 0.8311901688575745, "num_tokens": 23735883.0, "step": 629 }, { "epoch": 0.08014247551202137, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1265277862548828e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3233319520950317, "learning_rate": 2.666384061042815e-06, "loss": 0.5, "mean_token_accuracy": 0.8419243693351746, "num_tokens": 23769918.0, "step": 630 }, { "epoch": 0.08026968579061187, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1265277862548828e-05, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.290507197380066, "learning_rate": 2.670623145400594e-06, "loss": 0.5086, "mean_token_accuracy": 0.838974118232727, "num_tokens": 23815402.0, "step": 631 }, { "epoch": 0.08039689606920239, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1265277862548828e-05, "ewc_loss_parallel": 8.106231689453125e-06, "grad_norm": 1.2913419008255005, "learning_rate": 2.6748622297583724e-06, "loss": 0.5055, "mean_token_accuracy": 0.8407370448112488, "num_tokens": 23858205.0, "step": 632 }, { "epoch": 0.0805241063477929, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1265277862548828e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3355541229248047, "learning_rate": 2.679101314116151e-06, "loss": 0.5195, "mean_token_accuracy": 0.8397536277770996, "num_tokens": 23893506.0, "step": 633 }, { "epoch": 0.08065131662638342, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1265277862548828e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.336893081665039, "learning_rate": 2.6833403984739298e-06, "loss": 0.5105, "mean_token_accuracy": 0.8356460332870483, "num_tokens": 23929457.0, "step": 634 }, { "epoch": 0.08077852690497392, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1324882507324219e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.2886745929718018, "learning_rate": 2.6875794828317086e-06, "loss": 0.5055, "mean_token_accuracy": 0.8436717391014099, "num_tokens": 23974577.0, "step": 635 }, { "epoch": 0.08090573718356443, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1324882507324219e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3670969009399414, "learning_rate": 2.6918185671894875e-06, "loss": 0.5822, "mean_token_accuracy": 0.8194076418876648, "num_tokens": 24007490.0, "step": 636 }, { "epoch": 0.08103294746215495, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1324882507324219e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.332839846611023, "learning_rate": 2.696057651547266e-06, "loss": 0.6203, "mean_token_accuracy": 0.8059439659118652, "num_tokens": 24048058.0, "step": 637 }, { "epoch": 0.08116015774074545, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1324882507324219e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.2892367839813232, "learning_rate": 2.700296735905045e-06, "loss": 0.4864, "mean_token_accuracy": 0.8452781438827515, "num_tokens": 24089776.0, "step": 638 }, { "epoch": 0.08128736801933596, "ewc_loss": 0.00927734375, "ewc_loss_diag": 1.1324882507324219e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3108055591583252, "learning_rate": 2.7045358202628238e-06, "loss": 0.537, "mean_token_accuracy": 0.8267415761947632, "num_tokens": 24135730.0, "step": 639 }, { "epoch": 0.08141457829792648, "ewc_loss": 0.00933837890625, "ewc_loss_diag": 1.138448715209961e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3447543382644653, "learning_rate": 2.7087749046206026e-06, "loss": 0.4924, "mean_token_accuracy": 0.8440179228782654, "num_tokens": 24168513.0, "step": 640 }, { "epoch": 0.08154178857651698, "ewc_loss": 0.00933837890625, "ewc_loss_diag": 1.138448715209961e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3377509117126465, "learning_rate": 2.7130139889783807e-06, "loss": 0.5555, "mean_token_accuracy": 0.8252231478691101, "num_tokens": 24206555.0, "step": 641 }, { "epoch": 0.08166899885510749, "ewc_loss": 0.0093994140625, "ewc_loss_diag": 1.138448715209961e-05, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.2900173664093018, "learning_rate": 2.7172530733361596e-06, "loss": 0.4631, "mean_token_accuracy": 0.8528172969818115, "num_tokens": 24248919.0, "step": 642 }, { "epoch": 0.081796209133698, "ewc_loss": 0.0093994140625, "ewc_loss_diag": 1.150369644165039e-05, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.2986204624176025, "learning_rate": 2.7214921576939385e-06, "loss": 0.4601, "mean_token_accuracy": 0.8548880219459534, "num_tokens": 24290493.0, "step": 643 }, { "epoch": 0.0819234194122885, "ewc_loss": 0.0093994140625, "ewc_loss_diag": 1.150369644165039e-05, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.3297462463378906, "learning_rate": 2.725731242051717e-06, "loss": 0.5153, "mean_token_accuracy": 0.839145302772522, "num_tokens": 24327968.0, "step": 644 }, { "epoch": 0.08205062969087902, "ewc_loss": 0.0093994140625, "ewc_loss_diag": 1.150369644165039e-05, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.32379949092865, "learning_rate": 2.729970326409496e-06, "loss": 0.5475, "mean_token_accuracy": 0.8297858238220215, "num_tokens": 24370019.0, "step": 645 }, { "epoch": 0.08217783996946953, "ewc_loss": 0.00933837890625, "ewc_loss_diag": 1.150369644165039e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3556734323501587, "learning_rate": 2.7342094107672747e-06, "loss": 0.5174, "mean_token_accuracy": 0.8397536873817444, "num_tokens": 24403911.0, "step": 646 }, { "epoch": 0.08230505024806005, "ewc_loss": 0.00933837890625, "ewc_loss_diag": 1.1563301086425781e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3341903686523438, "learning_rate": 2.7384484951250536e-06, "loss": 0.5359, "mean_token_accuracy": 0.83384108543396, "num_tokens": 24445039.0, "step": 647 }, { "epoch": 0.08243226052665055, "ewc_loss": 0.00946044921875, "ewc_loss_diag": 1.1563301086425781e-05, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.370648980140686, "learning_rate": 2.7426875794828316e-06, "loss": 0.5359, "mean_token_accuracy": 0.828570544719696, "num_tokens": 24479543.0, "step": 648 }, { "epoch": 0.08255947080524106, "ewc_loss": 0.0093994140625, "ewc_loss_diag": 1.1563301086425781e-05, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.4145311117172241, "learning_rate": 2.7469266638406105e-06, "loss": 0.4951, "mean_token_accuracy": 0.840997576713562, "num_tokens": 24517483.0, "step": 649 }, { "epoch": 0.08268668108383158, "ewc_loss": 0.00946044921875, "ewc_loss_diag": 1.1563301086425781e-05, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.3205583095550537, "learning_rate": 2.7511657481983894e-06, "loss": 0.5273, "mean_token_accuracy": 0.8311447501182556, "num_tokens": 24557702.0, "step": 650 }, { "epoch": 0.08281389136242208, "ewc_loss": 0.00933837890625, "ewc_loss_diag": 1.1563301086425781e-05, "ewc_loss_parallel": 8.165836334228516e-06, "grad_norm": 1.3896939754486084, "learning_rate": 2.7554048325561683e-06, "loss": 0.5302, "mean_token_accuracy": 0.8343129754066467, "num_tokens": 24595145.0, "step": 651 }, { "epoch": 0.0829411016410126, "ewc_loss": 0.00946044921875, "ewc_loss_diag": 1.1682510375976562e-05, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.3327161073684692, "learning_rate": 2.7596439169139467e-06, "loss": 0.5197, "mean_token_accuracy": 0.8363298773765564, "num_tokens": 24631564.0, "step": 652 }, { "epoch": 0.08306831191960311, "ewc_loss": 0.00946044921875, "ewc_loss_diag": 1.1682510375976562e-05, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.3467804193496704, "learning_rate": 2.7638830012717256e-06, "loss": 0.5737, "mean_token_accuracy": 0.8206771016120911, "num_tokens": 24668563.0, "step": 653 }, { "epoch": 0.08319552219819361, "ewc_loss": 0.0093994140625, "ewc_loss_diag": 1.1682510375976562e-05, "ewc_loss_parallel": 8.225440979003906e-06, "grad_norm": 1.357519507408142, "learning_rate": 2.7681220856295045e-06, "loss": 0.5595, "mean_token_accuracy": 0.8226890563964844, "num_tokens": 24703662.0, "step": 654 }, { "epoch": 0.08332273247678412, "ewc_loss": 0.00958251953125, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 8.404254913330078e-06, "grad_norm": 1.3408005237579346, "learning_rate": 2.7723611699872834e-06, "loss": 0.5134, "mean_token_accuracy": 0.841331422328949, "num_tokens": 24739504.0, "step": 655 }, { "epoch": 0.08344994275537464, "ewc_loss": 0.00958251953125, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 8.404254913330078e-06, "grad_norm": 1.3328981399536133, "learning_rate": 2.7766002543450614e-06, "loss": 0.5224, "mean_token_accuracy": 0.8317545652389526, "num_tokens": 24781884.0, "step": 656 }, { "epoch": 0.08357715303396514, "ewc_loss": 0.00958251953125, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 8.404254913330078e-06, "grad_norm": 1.3290513753890991, "learning_rate": 2.7808393387028403e-06, "loss": 0.5765, "mean_token_accuracy": 0.8180402517318726, "num_tokens": 24823964.0, "step": 657 }, { "epoch": 0.08370436331255565, "ewc_loss": 0.00946044921875, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.4763331413269043, "learning_rate": 2.785078423060619e-06, "loss": 0.5192, "mean_token_accuracy": 0.8367252945899963, "num_tokens": 24859552.0, "step": 658 }, { "epoch": 0.08383157359114617, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 8.702278137207031e-06, "grad_norm": 1.3671795129776, "learning_rate": 2.789317507418398e-06, "loss": 0.5388, "mean_token_accuracy": 0.8298553824424744, "num_tokens": 24897580.0, "step": 659 }, { "epoch": 0.08395878386973668, "ewc_loss": 0.009521484375, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 8.344650268554688e-06, "grad_norm": 1.3254790306091309, "learning_rate": 2.7935565917761765e-06, "loss": 0.4866, "mean_token_accuracy": 0.8428757190704346, "num_tokens": 24934616.0, "step": 660 }, { "epoch": 0.08408599414832718, "ewc_loss": 0.009521484375, "ewc_loss_diag": 1.1742115020751953e-05, "ewc_loss_parallel": 8.344650268554688e-06, "grad_norm": 1.3689466714859009, "learning_rate": 2.7977956761339554e-06, "loss": 0.5585, "mean_token_accuracy": 0.8240944743156433, "num_tokens": 24972149.0, "step": 661 }, { "epoch": 0.0842132044269177, "ewc_loss": 0.00970458984375, "ewc_loss_diag": 1.1801719665527344e-05, "ewc_loss_parallel": 8.52346420288086e-06, "grad_norm": 1.3392781019210815, "learning_rate": 2.8020347604917343e-06, "loss": 0.4401, "mean_token_accuracy": 0.8613722324371338, "num_tokens": 25008682.0, "step": 662 }, { "epoch": 0.08434041470550821, "ewc_loss": 0.00946044921875, "ewc_loss_diag": 1.1861324310302734e-05, "ewc_loss_parallel": 8.285045623779297e-06, "grad_norm": 1.3275686502456665, "learning_rate": 2.8062738448495123e-06, "loss": 0.4658, "mean_token_accuracy": 0.849697470664978, "num_tokens": 25048454.0, "step": 663 }, { "epoch": 0.08446762498409871, "ewc_loss": 0.00970458984375, "ewc_loss_diag": 1.1861324310302734e-05, "ewc_loss_parallel": 8.52346420288086e-06, "grad_norm": 1.3757046461105347, "learning_rate": 2.8105129292072912e-06, "loss": 0.5428, "mean_token_accuracy": 0.8236916065216064, "num_tokens": 25086772.0, "step": 664 }, { "epoch": 0.08459483526268922, "ewc_loss": 0.00970458984375, "ewc_loss_diag": 1.1861324310302734e-05, "ewc_loss_parallel": 8.52346420288086e-06, "grad_norm": 1.347700834274292, "learning_rate": 2.81475201356507e-06, "loss": 0.4917, "mean_token_accuracy": 0.8415341377258301, "num_tokens": 25122215.0, "step": 665 }, { "epoch": 0.08472204554127974, "ewc_loss": 0.00970458984375, "ewc_loss_diag": 1.1861324310302734e-05, "ewc_loss_parallel": 8.52346420288086e-06, "grad_norm": 1.356866478919983, "learning_rate": 2.818991097922849e-06, "loss": 0.5343, "mean_token_accuracy": 0.8348706364631653, "num_tokens": 25159966.0, "step": 666 }, { "epoch": 0.08484925581987024, "ewc_loss": 0.009765625, "ewc_loss_diag": 1.1920928955078125e-05, "ewc_loss_parallel": 8.52346420288086e-06, "grad_norm": 1.3770662546157837, "learning_rate": 2.8232301822806274e-06, "loss": 0.4887, "mean_token_accuracy": 0.8473790884017944, "num_tokens": 25191929.0, "step": 667 }, { "epoch": 0.08497646609846075, "ewc_loss": 0.009765625, "ewc_loss_diag": 1.1920928955078125e-05, "ewc_loss_parallel": 8.52346420288086e-06, "grad_norm": 1.3428865671157837, "learning_rate": 2.8274692666384063e-06, "loss": 0.5098, "mean_token_accuracy": 0.8354113698005676, "num_tokens": 25233385.0, "step": 668 }, { "epoch": 0.08510367637705127, "ewc_loss": 0.009765625, "ewc_loss_diag": 1.1920928955078125e-05, "ewc_loss_parallel": 8.58306884765625e-06, "grad_norm": 1.3318637609481812, "learning_rate": 2.8317083509961852e-06, "loss": 0.5297, "mean_token_accuracy": 0.8338056206703186, "num_tokens": 25276943.0, "step": 669 }, { "epoch": 0.08523088665564178, "ewc_loss": 0.00970458984375, "ewc_loss_diag": 1.1980533599853516e-05, "ewc_loss_parallel": 8.463859558105469e-06, "grad_norm": 1.3481789827346802, "learning_rate": 2.835947435353964e-06, "loss": 0.5074, "mean_token_accuracy": 0.8395369648933411, "num_tokens": 25316552.0, "step": 670 }, { "epoch": 0.08535809693423228, "ewc_loss": 0.00982666015625, "ewc_loss_diag": 1.1980533599853516e-05, "ewc_loss_parallel": 8.58306884765625e-06, "grad_norm": 1.3763515949249268, "learning_rate": 2.840186519711742e-06, "loss": 0.562, "mean_token_accuracy": 0.829578697681427, "num_tokens": 25357517.0, "step": 671 }, { "epoch": 0.0854853072128228, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 1.2040138244628906e-05, "ewc_loss_parallel": 8.64267349243164e-06, "grad_norm": 1.3882544040679932, "learning_rate": 2.844425604069521e-06, "loss": 0.5529, "mean_token_accuracy": 0.8267140984535217, "num_tokens": 25392693.0, "step": 672 }, { "epoch": 0.08561251749141331, "ewc_loss": 0.009765625, "ewc_loss_diag": 1.2040138244628906e-05, "ewc_loss_parallel": 8.52346420288086e-06, "grad_norm": 1.3590807914733887, "learning_rate": 2.8486646884273e-06, "loss": 0.5153, "mean_token_accuracy": 0.837028980255127, "num_tokens": 25430623.0, "step": 673 }, { "epoch": 0.08573972777000381, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 1.2040138244628906e-05, "ewc_loss_parallel": 8.64267349243164e-06, "grad_norm": 1.348572850227356, "learning_rate": 2.852903772785079e-06, "loss": 0.4549, "mean_token_accuracy": 0.8554097414016724, "num_tokens": 25465043.0, "step": 674 }, { "epoch": 0.08586693804859433, "ewc_loss": 0.00982666015625, "ewc_loss_diag": 1.2040138244628906e-05, "ewc_loss_parallel": 8.58306884765625e-06, "grad_norm": 1.3413145542144775, "learning_rate": 2.8571428571428573e-06, "loss": 0.4779, "mean_token_accuracy": 0.8492962121963501, "num_tokens": 25503483.0, "step": 675 }, { "epoch": 0.08599414832718484, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 1.2099742889404297e-05, "ewc_loss_parallel": 8.64267349243164e-06, "grad_norm": 1.346285104751587, "learning_rate": 2.861381941500636e-06, "loss": 0.5304, "mean_token_accuracy": 0.8332939147949219, "num_tokens": 25547079.0, "step": 676 }, { "epoch": 0.08612135860577534, "ewc_loss": 0.00994873046875, "ewc_loss_diag": 1.2159347534179688e-05, "ewc_loss_parallel": 8.702278137207031e-06, "grad_norm": 1.3601645231246948, "learning_rate": 2.865621025858415e-06, "loss": 0.5664, "mean_token_accuracy": 0.8204900622367859, "num_tokens": 25586494.0, "step": 677 }, { "epoch": 0.08624856888436586, "ewc_loss": 0.010009765625, "ewc_loss_diag": 1.2159347534179688e-05, "ewc_loss_parallel": 8.761882781982422e-06, "grad_norm": 1.3978190422058105, "learning_rate": 2.869860110216193e-06, "loss": 0.5659, "mean_token_accuracy": 0.8206943273544312, "num_tokens": 25618998.0, "step": 678 }, { "epoch": 0.08637577916295637, "ewc_loss": 0.010009765625, "ewc_loss_diag": 1.2159347534179688e-05, "ewc_loss_parallel": 8.761882781982422e-06, "grad_norm": 1.339869737625122, "learning_rate": 2.874099194573972e-06, "loss": 0.507, "mean_token_accuracy": 0.8399567604064941, "num_tokens": 25663147.0, "step": 679 }, { "epoch": 0.08650298944154687, "ewc_loss": 0.0098876953125, "ewc_loss_diag": 1.2159347534179688e-05, "ewc_loss_parallel": 8.64267349243164e-06, "grad_norm": 1.3883957862854004, "learning_rate": 2.878338278931751e-06, "loss": 0.5555, "mean_token_accuracy": 0.8301737904548645, "num_tokens": 25696830.0, "step": 680 }, { "epoch": 0.08663019972013739, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.2159347534179688e-05, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.3662148714065552, "learning_rate": 2.8825773632895297e-06, "loss": 0.484, "mean_token_accuracy": 0.8509290218353271, "num_tokens": 25732828.0, "step": 681 }, { "epoch": 0.0867574099987279, "ewc_loss": 0.010009765625, "ewc_loss_diag": 1.2159347534179688e-05, "ewc_loss_parallel": 8.761882781982422e-06, "grad_norm": 1.3521766662597656, "learning_rate": 2.886816447647308e-06, "loss": 0.474, "mean_token_accuracy": 0.8485822677612305, "num_tokens": 25773545.0, "step": 682 }, { "epoch": 0.08688462027731841, "ewc_loss": 0.010009765625, "ewc_loss_diag": 1.2278556823730469e-05, "ewc_loss_parallel": 8.761882781982422e-06, "grad_norm": 1.3493350744247437, "learning_rate": 2.891055532005087e-06, "loss": 0.5123, "mean_token_accuracy": 0.8381400108337402, "num_tokens": 25814792.0, "step": 683 }, { "epoch": 0.08701183055590891, "ewc_loss": 0.01007080078125, "ewc_loss_diag": 1.233816146850586e-05, "ewc_loss_parallel": 8.821487426757812e-06, "grad_norm": 1.377364993095398, "learning_rate": 2.895294616362866e-06, "loss": 0.5276, "mean_token_accuracy": 0.8318020701408386, "num_tokens": 25850713.0, "step": 684 }, { "epoch": 0.08713904083449943, "ewc_loss": 0.01007080078125, "ewc_loss_diag": 1.233816146850586e-05, "ewc_loss_parallel": 8.821487426757812e-06, "grad_norm": 1.3667105436325073, "learning_rate": 2.899533700720645e-06, "loss": 0.4974, "mean_token_accuracy": 0.8381832838058472, "num_tokens": 25888320.0, "step": 685 }, { "epoch": 0.08726625111308994, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.3722361326217651, "learning_rate": 2.9037727850784233e-06, "loss": 0.5403, "mean_token_accuracy": 0.8292741775512695, "num_tokens": 25924941.0, "step": 686 }, { "epoch": 0.08739346139168044, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.363520860671997, "learning_rate": 2.9080118694362018e-06, "loss": 0.5064, "mean_token_accuracy": 0.8401975631713867, "num_tokens": 25964051.0, "step": 687 }, { "epoch": 0.08752067167027096, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.3497505187988281, "learning_rate": 2.9122509537939806e-06, "loss": 0.5092, "mean_token_accuracy": 0.8399658799171448, "num_tokens": 26007300.0, "step": 688 }, { "epoch": 0.08764788194886147, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 8.881092071533203e-06, "grad_norm": 1.394123911857605, "learning_rate": 2.9164900381517595e-06, "loss": 0.5661, "mean_token_accuracy": 0.8184263706207275, "num_tokens": 26045833.0, "step": 689 }, { "epoch": 0.08777509222745197, "ewc_loss": 0.01025390625, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 9.059906005859375e-06, "grad_norm": 1.3535411357879639, "learning_rate": 2.920729122509538e-06, "loss": 0.5584, "mean_token_accuracy": 0.8240064382553101, "num_tokens": 26092277.0, "step": 690 }, { "epoch": 0.08790230250604249, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.3725162744522095, "learning_rate": 2.924968206867317e-06, "loss": 0.5115, "mean_token_accuracy": 0.8353688716888428, "num_tokens": 26130082.0, "step": 691 }, { "epoch": 0.088029512784633, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.3680893182754517, "learning_rate": 2.9292072912250958e-06, "loss": 0.4873, "mean_token_accuracy": 0.8472784757614136, "num_tokens": 26170087.0, "step": 692 }, { "epoch": 0.0881567230632235, "ewc_loss": 0.01025390625, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 9.059906005859375e-06, "grad_norm": 1.4112869501113892, "learning_rate": 2.9334463755828746e-06, "loss": 0.5668, "mean_token_accuracy": 0.8209081292152405, "num_tokens": 26207645.0, "step": 693 }, { "epoch": 0.08828393334181402, "ewc_loss": 0.01019287109375, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 9.000301361083984e-06, "grad_norm": 1.382108449935913, "learning_rate": 2.937685459940653e-06, "loss": 0.4897, "mean_token_accuracy": 0.8419673442840576, "num_tokens": 26242417.0, "step": 694 }, { "epoch": 0.08841114362040453, "ewc_loss": 0.01025390625, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 9.059906005859375e-06, "grad_norm": 1.4093886613845825, "learning_rate": 2.941924544298432e-06, "loss": 0.5572, "mean_token_accuracy": 0.8254843950271606, "num_tokens": 26275427.0, "step": 695 }, { "epoch": 0.08853835389899505, "ewc_loss": 0.01025390625, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 9.059906005859375e-06, "grad_norm": 1.3910473585128784, "learning_rate": 2.9461636286562104e-06, "loss": 0.5098, "mean_token_accuracy": 0.8383721113204956, "num_tokens": 26308692.0, "step": 696 }, { "epoch": 0.08866556417758555, "ewc_loss": 0.0101318359375, "ewc_loss_diag": 1.245737075805664e-05, "ewc_loss_parallel": 8.940696716308594e-06, "grad_norm": 1.3426227569580078, "learning_rate": 2.950402713013989e-06, "loss": 0.5036, "mean_token_accuracy": 0.8407533168792725, "num_tokens": 26352155.0, "step": 697 }, { "epoch": 0.08879277445617606, "ewc_loss": 0.01025390625, "ewc_loss_diag": 1.2516975402832031e-05, "ewc_loss_parallel": 9.059906005859375e-06, "grad_norm": 1.4203529357910156, "learning_rate": 2.954641797371768e-06, "loss": 0.546, "mean_token_accuracy": 0.8291397094726562, "num_tokens": 26392410.0, "step": 698 }, { "epoch": 0.08891998473476657, "ewc_loss": 0.0103759765625, "ewc_loss_diag": 1.2576580047607422e-05, "ewc_loss_parallel": 9.119510650634766e-06, "grad_norm": 1.4132317304611206, "learning_rate": 2.9588808817295467e-06, "loss": 0.4729, "mean_token_accuracy": 0.8484185934066772, "num_tokens": 26430120.0, "step": 699 }, { "epoch": 0.08904719501335707, "ewc_loss": 0.0103759765625, "ewc_loss_diag": 1.2516975402832031e-05, "ewc_loss_parallel": 9.119510650634766e-06, "grad_norm": 1.3893558979034424, "learning_rate": 2.9631199660873256e-06, "loss": 0.5355, "mean_token_accuracy": 0.8287944197654724, "num_tokens": 26466159.0, "step": 700 }, { "epoch": 0.08917440529194759, "ewc_loss": 0.01031494140625, "ewc_loss_diag": 1.2576580047607422e-05, "ewc_loss_parallel": 9.059906005859375e-06, "grad_norm": 1.3637306690216064, "learning_rate": 2.967359050445104e-06, "loss": 0.4562, "mean_token_accuracy": 0.8564428091049194, "num_tokens": 26502382.0, "step": 701 }, { "epoch": 0.0893016155705381, "ewc_loss": 0.01031494140625, "ewc_loss_diag": 1.2576580047607422e-05, "ewc_loss_parallel": 9.059906005859375e-06, "grad_norm": 1.3958916664123535, "learning_rate": 2.971598134802883e-06, "loss": 0.4942, "mean_token_accuracy": 0.8445004224777222, "num_tokens": 26540911.0, "step": 702 }, { "epoch": 0.0894288258491286, "ewc_loss": 0.01043701171875, "ewc_loss_diag": 1.2636184692382812e-05, "ewc_loss_parallel": 9.179115295410156e-06, "grad_norm": 1.3915393352508545, "learning_rate": 2.975837219160662e-06, "loss": 0.5562, "mean_token_accuracy": 0.8240852355957031, "num_tokens": 26580081.0, "step": 703 }, { "epoch": 0.08955603612771912, "ewc_loss": 0.010498046875, "ewc_loss_diag": 1.2695789337158203e-05, "ewc_loss_parallel": 9.238719940185547e-06, "grad_norm": 1.3843592405319214, "learning_rate": 2.9800763035184403e-06, "loss": 0.5219, "mean_token_accuracy": 0.8349214196205139, "num_tokens": 26622033.0, "step": 704 }, { "epoch": 0.08968324640630963, "ewc_loss": 0.010498046875, "ewc_loss_diag": 1.2695789337158203e-05, "ewc_loss_parallel": 9.238719940185547e-06, "grad_norm": 1.3983185291290283, "learning_rate": 2.9843153878762187e-06, "loss": 0.5221, "mean_token_accuracy": 0.8352874517440796, "num_tokens": 26657042.0, "step": 705 }, { "epoch": 0.08981045668490013, "ewc_loss": 0.010498046875, "ewc_loss_diag": 1.2695789337158203e-05, "ewc_loss_parallel": 9.238719940185547e-06, "grad_norm": 1.4040273427963257, "learning_rate": 2.9885544722339976e-06, "loss": 0.5338, "mean_token_accuracy": 0.8322582244873047, "num_tokens": 26696668.0, "step": 706 }, { "epoch": 0.08993766696349065, "ewc_loss": 0.010498046875, "ewc_loss_diag": 1.2814998626708984e-05, "ewc_loss_parallel": 9.238719940185547e-06, "grad_norm": 1.3900209665298462, "learning_rate": 2.9927935565917765e-06, "loss": 0.5366, "mean_token_accuracy": 0.8331068754196167, "num_tokens": 26736550.0, "step": 707 }, { "epoch": 0.09006487724208116, "ewc_loss": 0.010498046875, "ewc_loss_diag": 1.2814998626708984e-05, "ewc_loss_parallel": 9.238719940185547e-06, "grad_norm": 1.3881350755691528, "learning_rate": 2.9970326409495554e-06, "loss": 0.543, "mean_token_accuracy": 0.8304710388183594, "num_tokens": 26776311.0, "step": 708 }, { "epoch": 0.09019208752067168, "ewc_loss": 0.010498046875, "ewc_loss_diag": 1.2814998626708984e-05, "ewc_loss_parallel": 9.238719940185547e-06, "grad_norm": 1.3891053199768066, "learning_rate": 3.001271725307334e-06, "loss": 0.4826, "mean_token_accuracy": 0.8468530774116516, "num_tokens": 26812094.0, "step": 709 }, { "epoch": 0.09031929779926218, "ewc_loss": 0.01055908203125, "ewc_loss_diag": 1.2814998626708984e-05, "ewc_loss_parallel": 9.298324584960938e-06, "grad_norm": 1.4314701557159424, "learning_rate": 3.0055108096651127e-06, "loss": 0.5175, "mean_token_accuracy": 0.8382667899131775, "num_tokens": 26846317.0, "step": 710 }, { "epoch": 0.09044650807785269, "ewc_loss": 0.0106201171875, "ewc_loss_diag": 1.2874603271484375e-05, "ewc_loss_parallel": 9.357929229736328e-06, "grad_norm": 1.3661513328552246, "learning_rate": 3.0097498940228916e-06, "loss": 0.4587, "mean_token_accuracy": 0.8530986905097961, "num_tokens": 26885247.0, "step": 711 }, { "epoch": 0.0905737183564432, "ewc_loss": 0.01055908203125, "ewc_loss_diag": 1.2874603271484375e-05, "ewc_loss_parallel": 9.298324584960938e-06, "grad_norm": 1.3979400396347046, "learning_rate": 3.0139889783806705e-06, "loss": 0.5225, "mean_token_accuracy": 0.8362640738487244, "num_tokens": 26925578.0, "step": 712 }, { "epoch": 0.0907009286350337, "ewc_loss": 0.0106201171875, "ewc_loss_diag": 1.2874603271484375e-05, "ewc_loss_parallel": 9.357929229736328e-06, "grad_norm": 1.4080811738967896, "learning_rate": 3.0182280627384485e-06, "loss": 0.5467, "mean_token_accuracy": 0.8305092453956604, "num_tokens": 26961924.0, "step": 713 }, { "epoch": 0.09082813891362422, "ewc_loss": 0.0106201171875, "ewc_loss_diag": 1.3053417205810547e-05, "ewc_loss_parallel": 9.357929229736328e-06, "grad_norm": 1.3815723657608032, "learning_rate": 3.0224671470962274e-06, "loss": 0.4764, "mean_token_accuracy": 0.8468190431594849, "num_tokens": 26999433.0, "step": 714 }, { "epoch": 0.09095534919221474, "ewc_loss": 0.01068115234375, "ewc_loss_diag": 1.3053417205810547e-05, "ewc_loss_parallel": 9.417533874511719e-06, "grad_norm": 1.372553825378418, "learning_rate": 3.0267062314540063e-06, "loss": 0.4989, "mean_token_accuracy": 0.8417925834655762, "num_tokens": 27044427.0, "step": 715 }, { "epoch": 0.09108255947080524, "ewc_loss": 0.0106201171875, "ewc_loss_diag": 1.3053417205810547e-05, "ewc_loss_parallel": 9.357929229736328e-06, "grad_norm": 1.3941216468811035, "learning_rate": 3.0309453158117847e-06, "loss": 0.5063, "mean_token_accuracy": 0.8391506671905518, "num_tokens": 27080645.0, "step": 716 }, { "epoch": 0.09120976974939575, "ewc_loss": 0.0107421875, "ewc_loss_diag": 1.3053417205810547e-05, "ewc_loss_parallel": 9.47713851928711e-06, "grad_norm": 1.4153587818145752, "learning_rate": 3.0351844001695636e-06, "loss": 0.5734, "mean_token_accuracy": 0.8214269876480103, "num_tokens": 27118213.0, "step": 717 }, { "epoch": 0.09133698002798626, "ewc_loss": 0.0108642578125, "ewc_loss_diag": 1.3113021850585938e-05, "ewc_loss_parallel": 9.5367431640625e-06, "grad_norm": 1.3956472873687744, "learning_rate": 3.0394234845273425e-06, "loss": 0.4911, "mean_token_accuracy": 0.8447152972221375, "num_tokens": 27156827.0, "step": 718 }, { "epoch": 0.09146419030657676, "ewc_loss": 0.01068115234375, "ewc_loss_diag": 1.3172626495361328e-05, "ewc_loss_parallel": 9.357929229736328e-06, "grad_norm": 1.4204890727996826, "learning_rate": 3.0436625688851214e-06, "loss": 0.5245, "mean_token_accuracy": 0.8298917412757874, "num_tokens": 27192016.0, "step": 719 }, { "epoch": 0.09159140058516728, "ewc_loss": 0.0108642578125, "ewc_loss_diag": 1.3172626495361328e-05, "ewc_loss_parallel": 9.5367431640625e-06, "grad_norm": 2.8388736248016357, "learning_rate": 3.0479016532428994e-06, "loss": 0.5698, "mean_token_accuracy": 0.8174530863761902, "num_tokens": 27226037.0, "step": 720 }, { "epoch": 0.0917186108637578, "ewc_loss": 0.01495361328125, "ewc_loss_diag": 1.3172626495361328e-05, "ewc_loss_parallel": 1.3589859008789062e-05, "grad_norm": 1.6770656108856201, "learning_rate": 3.0521407376006783e-06, "loss": 0.4877, "mean_token_accuracy": 0.8457393646240234, "num_tokens": 27260399.0, "step": 721 }, { "epoch": 0.09184582114234831, "ewc_loss": 0.01025390625, "ewc_loss_diag": 1.3172626495361328e-05, "ewc_loss_parallel": 8.881092071533203e-06, "grad_norm": 1.292126178741455, "learning_rate": 3.056379821958457e-06, "loss": 0.4937, "mean_token_accuracy": 0.8445159792900085, "num_tokens": 27297293.0, "step": 722 }, { "epoch": 0.09197303142093881, "ewc_loss": 0.011474609375, "ewc_loss_diag": 1.3232231140136719e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.5258134603500366, "learning_rate": 3.060618906316236e-06, "loss": 0.4909, "mean_token_accuracy": 0.8439630270004272, "num_tokens": 27333045.0, "step": 723 }, { "epoch": 0.09210024169952932, "ewc_loss": 0.01177978515625, "ewc_loss_diag": 1.3232231140136719e-05, "ewc_loss_parallel": 1.043081283569336e-05, "grad_norm": 1.4247233867645264, "learning_rate": 3.0648579906740146e-06, "loss": 0.4882, "mean_token_accuracy": 0.8435593843460083, "num_tokens": 27369044.0, "step": 724 }, { "epoch": 0.09222745197811984, "ewc_loss": 0.01141357421875, "ewc_loss_diag": 1.3232231140136719e-05, "ewc_loss_parallel": 1.0073184967041016e-05, "grad_norm": 1.4665507078170776, "learning_rate": 3.0690970750317934e-06, "loss": 0.543, "mean_token_accuracy": 0.8285766243934631, "num_tokens": 27407486.0, "step": 725 }, { "epoch": 0.09235466225671034, "ewc_loss": 0.0118408203125, "ewc_loss_diag": 1.3232231140136719e-05, "ewc_loss_parallel": 1.049041748046875e-05, "grad_norm": 1.4563437700271606, "learning_rate": 3.0733361593895723e-06, "loss": 0.5075, "mean_token_accuracy": 0.8400120735168457, "num_tokens": 27441835.0, "step": 726 }, { "epoch": 0.09248187253530085, "ewc_loss": 0.01153564453125, "ewc_loss_diag": 1.3232231140136719e-05, "ewc_loss_parallel": 1.0192394256591797e-05, "grad_norm": 1.4182555675506592, "learning_rate": 3.077575243747351e-06, "loss": 0.5205, "mean_token_accuracy": 0.834296703338623, "num_tokens": 27483105.0, "step": 727 }, { "epoch": 0.09260908281389137, "ewc_loss": 0.011474609375, "ewc_loss_diag": 1.329183578491211e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.4322564601898193, "learning_rate": 3.0818143281051292e-06, "loss": 0.5396, "mean_token_accuracy": 0.8320741057395935, "num_tokens": 27528888.0, "step": 728 }, { "epoch": 0.09273629309248187, "ewc_loss": 0.01153564453125, "ewc_loss_diag": 1.33514404296875e-05, "ewc_loss_parallel": 1.0192394256591797e-05, "grad_norm": 1.4360647201538086, "learning_rate": 3.086053412462908e-06, "loss": 0.5776, "mean_token_accuracy": 0.819386899471283, "num_tokens": 27571763.0, "step": 729 }, { "epoch": 0.09286350337107238, "ewc_loss": 0.0113525390625, "ewc_loss_diag": 1.33514404296875e-05, "ewc_loss_parallel": 1.0013580322265625e-05, "grad_norm": 1.4348163604736328, "learning_rate": 3.090292496820687e-06, "loss": 0.5085, "mean_token_accuracy": 0.8422706723213196, "num_tokens": 27606980.0, "step": 730 }, { "epoch": 0.0929907136496629, "ewc_loss": 0.0113525390625, "ewc_loss_diag": 1.33514404296875e-05, "ewc_loss_parallel": 1.0013580322265625e-05, "grad_norm": 1.4707014560699463, "learning_rate": 3.094531581178466e-06, "loss": 0.5003, "mean_token_accuracy": 0.8417327404022217, "num_tokens": 27640969.0, "step": 731 }, { "epoch": 0.0931179239282534, "ewc_loss": 0.011474609375, "ewc_loss_diag": 1.33514404296875e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.430163025856018, "learning_rate": 3.0987706655362444e-06, "loss": 0.6373, "mean_token_accuracy": 0.7993552684783936, "num_tokens": 27684360.0, "step": 732 }, { "epoch": 0.09324513420684391, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.33514404296875e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.4187672138214111, "learning_rate": 3.1030097498940232e-06, "loss": 0.4782, "mean_token_accuracy": 0.8461037874221802, "num_tokens": 27723596.0, "step": 733 }, { "epoch": 0.09337234448543442, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.33514404296875e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.4356001615524292, "learning_rate": 3.107248834251802e-06, "loss": 0.5135, "mean_token_accuracy": 0.8306320309638977, "num_tokens": 27759469.0, "step": 734 }, { "epoch": 0.09349955476402494, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3530254364013672e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.4266194105148315, "learning_rate": 3.11148791860958e-06, "loss": 0.5061, "mean_token_accuracy": 0.8404478430747986, "num_tokens": 27795783.0, "step": 735 }, { "epoch": 0.09362676504261544, "ewc_loss": 0.0111083984375, "ewc_loss_diag": 1.3530254364013672e-05, "ewc_loss_parallel": 9.775161743164062e-06, "grad_norm": 1.4556796550750732, "learning_rate": 3.115727002967359e-06, "loss": 0.5153, "mean_token_accuracy": 0.8341698050498962, "num_tokens": 27826146.0, "step": 736 }, { "epoch": 0.09375397532120595, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3530254364013672e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.4093786478042603, "learning_rate": 3.119966087325138e-06, "loss": 0.5188, "mean_token_accuracy": 0.8384107351303101, "num_tokens": 27865581.0, "step": 737 }, { "epoch": 0.09388118559979647, "ewc_loss": 0.01104736328125, "ewc_loss_diag": 1.3530254364013672e-05, "ewc_loss_parallel": 9.715557098388672e-06, "grad_norm": 1.411811351776123, "learning_rate": 3.124205171682917e-06, "loss": 0.5621, "mean_token_accuracy": 0.8243552446365356, "num_tokens": 27911209.0, "step": 738 }, { "epoch": 0.09400839587838697, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3530254364013672e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.419463872909546, "learning_rate": 3.1284442560406953e-06, "loss": 0.5249, "mean_token_accuracy": 0.833411455154419, "num_tokens": 27951639.0, "step": 739 }, { "epoch": 0.09413560615697748, "ewc_loss": 0.01116943359375, "ewc_loss_diag": 1.3530254364013672e-05, "ewc_loss_parallel": 9.834766387939453e-06, "grad_norm": 1.4106931686401367, "learning_rate": 3.132683340398474e-06, "loss": 0.4588, "mean_token_accuracy": 0.8513048887252808, "num_tokens": 27990808.0, "step": 740 }, { "epoch": 0.094262816435568, "ewc_loss": 0.01129150390625, "ewc_loss_diag": 1.3589859008789062e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.4537123441696167, "learning_rate": 3.136922424756253e-06, "loss": 0.4819, "mean_token_accuracy": 0.8494218587875366, "num_tokens": 28025421.0, "step": 741 }, { "epoch": 0.0943900267141585, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3589859008789062e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.3921864032745361, "learning_rate": 3.141161509114032e-06, "loss": 0.4529, "mean_token_accuracy": 0.8578771948814392, "num_tokens": 28068652.0, "step": 742 }, { "epoch": 0.09451723699274901, "ewc_loss": 0.0111083984375, "ewc_loss_diag": 1.3589859008789062e-05, "ewc_loss_parallel": 9.775161743164062e-06, "grad_norm": 1.433293104171753, "learning_rate": 3.14540059347181e-06, "loss": 0.497, "mean_token_accuracy": 0.8400373458862305, "num_tokens": 28104852.0, "step": 743 }, { "epoch": 0.09464444727133953, "ewc_loss": 0.01129150390625, "ewc_loss_diag": 1.3649463653564453e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.4280503988265991, "learning_rate": 3.149639677829589e-06, "loss": 0.4688, "mean_token_accuracy": 0.8501697778701782, "num_tokens": 28141646.0, "step": 744 }, { "epoch": 0.09477165754993004, "ewc_loss": 0.0111083984375, "ewc_loss_diag": 1.3649463653564453e-05, "ewc_loss_parallel": 9.775161743164062e-06, "grad_norm": 1.451610803604126, "learning_rate": 3.1538787621873677e-06, "loss": 0.5668, "mean_token_accuracy": 0.826645016670227, "num_tokens": 28178951.0, "step": 745 }, { "epoch": 0.09489886782852054, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3649463653564453e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.4332776069641113, "learning_rate": 3.1581178465451466e-06, "loss": 0.582, "mean_token_accuracy": 0.8216593265533447, "num_tokens": 28221911.0, "step": 746 }, { "epoch": 0.09502607810711106, "ewc_loss": 0.01116943359375, "ewc_loss_diag": 1.3649463653564453e-05, "ewc_loss_parallel": 9.834766387939453e-06, "grad_norm": 1.4668464660644531, "learning_rate": 3.162356930902925e-06, "loss": 0.5012, "mean_token_accuracy": 0.841727614402771, "num_tokens": 28252828.0, "step": 747 }, { "epoch": 0.09515328838570157, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3649463653564453e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.415949821472168, "learning_rate": 3.166596015260704e-06, "loss": 0.4762, "mean_token_accuracy": 0.8475759029388428, "num_tokens": 28293459.0, "step": 748 }, { "epoch": 0.09528049866429207, "ewc_loss": 0.0111083984375, "ewc_loss_diag": 1.3649463653564453e-05, "ewc_loss_parallel": 9.775161743164062e-06, "grad_norm": 1.4330050945281982, "learning_rate": 3.170835099618483e-06, "loss": 0.5114, "mean_token_accuracy": 0.8380779027938843, "num_tokens": 28332775.0, "step": 749 }, { "epoch": 0.09540770894288259, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3709068298339844e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.4215418100357056, "learning_rate": 3.1750741839762617e-06, "loss": 0.4889, "mean_token_accuracy": 0.8451105952262878, "num_tokens": 28370367.0, "step": 750 }, { "epoch": 0.0955349192214731, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3709068298339844e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.4534531831741333, "learning_rate": 3.1793132683340398e-06, "loss": 0.5212, "mean_token_accuracy": 0.8369899988174438, "num_tokens": 28404913.0, "step": 751 }, { "epoch": 0.0956621295000636, "ewc_loss": 0.0113525390625, "ewc_loss_diag": 1.3768672943115234e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.4353609085083008, "learning_rate": 3.1835523526918187e-06, "loss": 0.4631, "mean_token_accuracy": 0.8512115478515625, "num_tokens": 28441057.0, "step": 752 }, { "epoch": 0.09578933977865411, "ewc_loss": 0.01129150390625, "ewc_loss_diag": 1.3828277587890625e-05, "ewc_loss_parallel": 9.894371032714844e-06, "grad_norm": 1.450265645980835, "learning_rate": 3.1877914370495975e-06, "loss": 0.5531, "mean_token_accuracy": 0.8239123225212097, "num_tokens": 28479691.0, "step": 753 }, { "epoch": 0.09591655005724463, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3887882232666016e-05, "ewc_loss_parallel": 9.834766387939453e-06, "grad_norm": 1.4406301975250244, "learning_rate": 3.192030521407376e-06, "loss": 0.547, "mean_token_accuracy": 0.8299261331558228, "num_tokens": 28520469.0, "step": 754 }, { "epoch": 0.09604376033583513, "ewc_loss": 0.0113525390625, "ewc_loss_diag": 1.3887882232666016e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.4297893047332764, "learning_rate": 3.196269605765155e-06, "loss": 0.4877, "mean_token_accuracy": 0.8435915112495422, "num_tokens": 28560622.0, "step": 755 }, { "epoch": 0.09617097061442564, "ewc_loss": 0.01123046875, "ewc_loss_diag": 1.3887882232666016e-05, "ewc_loss_parallel": 9.834766387939453e-06, "grad_norm": 1.4590586423873901, "learning_rate": 3.2005086901229338e-06, "loss": 0.5269, "mean_token_accuracy": 0.8318090438842773, "num_tokens": 28594890.0, "step": 756 }, { "epoch": 0.09629818089301616, "ewc_loss": 0.01141357421875, "ewc_loss_diag": 1.4007091522216797e-05, "ewc_loss_parallel": 1.0013580322265625e-05, "grad_norm": 1.437485694885254, "learning_rate": 3.2047477744807127e-06, "loss": 0.52, "mean_token_accuracy": 0.8339744806289673, "num_tokens": 28633112.0, "step": 757 }, { "epoch": 0.09642539117160667, "ewc_loss": 0.0113525390625, "ewc_loss_diag": 1.4007091522216797e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.4428526163101196, "learning_rate": 3.208986858838491e-06, "loss": 0.5474, "mean_token_accuracy": 0.8250092267990112, "num_tokens": 28672195.0, "step": 758 }, { "epoch": 0.09655260145019717, "ewc_loss": 0.01141357421875, "ewc_loss_diag": 1.4007091522216797e-05, "ewc_loss_parallel": 1.0013580322265625e-05, "grad_norm": 1.437804102897644, "learning_rate": 3.2132259431962696e-06, "loss": 0.4868, "mean_token_accuracy": 0.8477769494056702, "num_tokens": 28710315.0, "step": 759 }, { "epoch": 0.09667981172878769, "ewc_loss": 0.0113525390625, "ewc_loss_diag": 1.4007091522216797e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.444374680519104, "learning_rate": 3.2174650275540485e-06, "loss": 0.5257, "mean_token_accuracy": 0.8388097286224365, "num_tokens": 28748713.0, "step": 760 }, { "epoch": 0.0968070220073782, "ewc_loss": 0.01141357421875, "ewc_loss_diag": 1.4066696166992188e-05, "ewc_loss_parallel": 1.0013580322265625e-05, "grad_norm": 1.4245213270187378, "learning_rate": 3.2217041119118274e-06, "loss": 0.4925, "mean_token_accuracy": 0.8425873517990112, "num_tokens": 28789832.0, "step": 761 }, { "epoch": 0.0969342322859687, "ewc_loss": 0.0113525390625, "ewc_loss_diag": 1.4126300811767578e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.468658685684204, "learning_rate": 3.225943196269606e-06, "loss": 0.5162, "mean_token_accuracy": 0.8385549783706665, "num_tokens": 28822781.0, "step": 762 }, { "epoch": 0.09706144256455922, "ewc_loss": 0.01153564453125, "ewc_loss_diag": 1.4185905456542969e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.450386643409729, "learning_rate": 3.2301822806273847e-06, "loss": 0.5016, "mean_token_accuracy": 0.837061882019043, "num_tokens": 28857734.0, "step": 763 }, { "epoch": 0.09718865284314973, "ewc_loss": 0.01153564453125, "ewc_loss_diag": 1.424551010131836e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.4365557432174683, "learning_rate": 3.2344213649851636e-06, "loss": 0.5335, "mean_token_accuracy": 0.8314723372459412, "num_tokens": 28901296.0, "step": 764 }, { "epoch": 0.09731586312174023, "ewc_loss": 0.01141357421875, "ewc_loss_diag": 1.424551010131836e-05, "ewc_loss_parallel": 1.0013580322265625e-05, "grad_norm": 1.4357740879058838, "learning_rate": 3.2386604493429425e-06, "loss": 0.5001, "mean_token_accuracy": 0.8387876152992249, "num_tokens": 28942053.0, "step": 765 }, { "epoch": 0.09744307340033075, "ewc_loss": 0.0115966796875, "ewc_loss_diag": 1.424551010131836e-05, "ewc_loss_parallel": 1.0192394256591797e-05, "grad_norm": 1.4746242761611938, "learning_rate": 3.242899533700721e-06, "loss": 0.4716, "mean_token_accuracy": 0.8452705144882202, "num_tokens": 28975320.0, "step": 766 }, { "epoch": 0.09757028367892126, "ewc_loss": 0.01171875, "ewc_loss_diag": 1.430511474609375e-05, "ewc_loss_parallel": 1.0251998901367188e-05, "grad_norm": 1.4222126007080078, "learning_rate": 3.2471386180585e-06, "loss": 0.5182, "mean_token_accuracy": 0.8301477432250977, "num_tokens": 29019941.0, "step": 767 }, { "epoch": 0.09769749395751176, "ewc_loss": 0.011474609375, "ewc_loss_diag": 1.430511474609375e-05, "ewc_loss_parallel": 1.0073184967041016e-05, "grad_norm": 1.4491294622421265, "learning_rate": 3.2513777024162783e-06, "loss": 0.5481, "mean_token_accuracy": 0.8299956321716309, "num_tokens": 29062608.0, "step": 768 }, { "epoch": 0.09782470423610228, "ewc_loss": 0.01171875, "ewc_loss_diag": 1.436471939086914e-05, "ewc_loss_parallel": 1.0251998901367188e-05, "grad_norm": 1.4423185586929321, "learning_rate": 3.255616786774057e-06, "loss": 0.4938, "mean_token_accuracy": 0.8450236320495605, "num_tokens": 29106423.0, "step": 769 }, { "epoch": 0.09795191451469279, "ewc_loss": 0.0115966796875, "ewc_loss_diag": 1.436471939086914e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.4326715469360352, "learning_rate": 3.2598558711318356e-06, "loss": 0.4406, "mean_token_accuracy": 0.8579147458076477, "num_tokens": 29148717.0, "step": 770 }, { "epoch": 0.0980791247932833, "ewc_loss": 0.0115966796875, "ewc_loss_diag": 1.436471939086914e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.441759705543518, "learning_rate": 3.2640949554896145e-06, "loss": 0.4872, "mean_token_accuracy": 0.8450389504432678, "num_tokens": 29189464.0, "step": 771 }, { "epoch": 0.0982063350718738, "ewc_loss": 0.0115966796875, "ewc_loss_diag": 1.436471939086914e-05, "ewc_loss_parallel": 1.0192394256591797e-05, "grad_norm": 1.4517320394515991, "learning_rate": 3.2683340398473934e-06, "loss": 0.5168, "mean_token_accuracy": 0.8364777565002441, "num_tokens": 29228384.0, "step": 772 }, { "epoch": 0.09833354535046432, "ewc_loss": 0.0115966796875, "ewc_loss_diag": 1.436471939086914e-05, "ewc_loss_parallel": 1.0192394256591797e-05, "grad_norm": 1.4484953880310059, "learning_rate": 3.272573124205172e-06, "loss": 0.5065, "mean_token_accuracy": 0.8436287045478821, "num_tokens": 29270942.0, "step": 773 }, { "epoch": 0.09846075562905483, "ewc_loss": 0.0115966796875, "ewc_loss_diag": 1.4424324035644531e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.4388833045959473, "learning_rate": 3.2768122085629507e-06, "loss": 0.4822, "mean_token_accuracy": 0.846979558467865, "num_tokens": 29313770.0, "step": 774 }, { "epoch": 0.09858796590764533, "ewc_loss": 0.01153564453125, "ewc_loss_diag": 1.4483928680419922e-05, "ewc_loss_parallel": 1.0073184967041016e-05, "grad_norm": 1.4719198942184448, "learning_rate": 3.2810512929207296e-06, "loss": 0.4794, "mean_token_accuracy": 0.8453312516212463, "num_tokens": 29347924.0, "step": 775 }, { "epoch": 0.09871517618623585, "ewc_loss": 0.0115966796875, "ewc_loss_diag": 1.4483928680419922e-05, "ewc_loss_parallel": 1.0132789611816406e-05, "grad_norm": 1.7322336435317993, "learning_rate": 3.2852903772785085e-06, "loss": 0.5292, "mean_token_accuracy": 0.8315891027450562, "num_tokens": 29386899.0, "step": 776 }, { "epoch": 0.09884238646482636, "ewc_loss": 0.01239013671875, "ewc_loss_diag": 1.4483928680419922e-05, "ewc_loss_parallel": 1.0907649993896484e-05, "grad_norm": 1.4585801362991333, "learning_rate": 3.2895294616362865e-06, "loss": 0.4751, "mean_token_accuracy": 0.8525298833847046, "num_tokens": 29423349.0, "step": 777 }, { "epoch": 0.09896959674341686, "ewc_loss": 0.01141357421875, "ewc_loss_diag": 1.4483928680419922e-05, "ewc_loss_parallel": 9.953975677490234e-06, "grad_norm": 1.4431042671203613, "learning_rate": 3.2937685459940654e-06, "loss": 0.4829, "mean_token_accuracy": 0.8436968326568604, "num_tokens": 29461127.0, "step": 778 }, { "epoch": 0.09909680702200738, "ewc_loss": 0.01171875, "ewc_loss_diag": 1.4483928680419922e-05, "ewc_loss_parallel": 1.0251998901367188e-05, "grad_norm": 1.490880012512207, "learning_rate": 3.2980076303518443e-06, "loss": 0.5664, "mean_token_accuracy": 0.8257383108139038, "num_tokens": 29498335.0, "step": 779 }, { "epoch": 0.09922401730059789, "ewc_loss": 0.01171875, "ewc_loss_diag": 1.4483928680419922e-05, "ewc_loss_parallel": 1.0251998901367188e-05, "grad_norm": 1.437883734703064, "learning_rate": 3.302246714709623e-06, "loss": 0.4936, "mean_token_accuracy": 0.8421800136566162, "num_tokens": 29541174.0, "step": 780 }, { "epoch": 0.09935122757918839, "ewc_loss": 0.01165771484375, "ewc_loss_diag": 1.4483928680419922e-05, "ewc_loss_parallel": 1.0192394256591797e-05, "grad_norm": 1.4656039476394653, "learning_rate": 3.3064857990674017e-06, "loss": 0.519, "mean_token_accuracy": 0.8375277519226074, "num_tokens": 29578684.0, "step": 781 }, { "epoch": 0.0994784378577789, "ewc_loss": 0.0118408203125, "ewc_loss_diag": 1.4543533325195312e-05, "ewc_loss_parallel": 1.0371208190917969e-05, "grad_norm": 1.4743051528930664, "learning_rate": 3.3107248834251805e-06, "loss": 0.4674, "mean_token_accuracy": 0.8510336875915527, "num_tokens": 29611236.0, "step": 782 }, { "epoch": 0.09960564813636942, "ewc_loss": 0.01190185546875, "ewc_loss_diag": 1.4603137969970703e-05, "ewc_loss_parallel": 1.043081283569336e-05, "grad_norm": 1.4541163444519043, "learning_rate": 3.3149639677829594e-06, "loss": 0.4847, "mean_token_accuracy": 0.8464756011962891, "num_tokens": 29653769.0, "step": 783 }, { "epoch": 0.09973285841495994, "ewc_loss": 0.01177978515625, "ewc_loss_diag": 1.4603137969970703e-05, "ewc_loss_parallel": 1.0311603546142578e-05, "grad_norm": 1.4628971815109253, "learning_rate": 3.3192030521407383e-06, "loss": 0.4973, "mean_token_accuracy": 0.8441412448883057, "num_tokens": 29695231.0, "step": 784 }, { "epoch": 0.09986006869355044, "ewc_loss": 0.011962890625, "ewc_loss_diag": 1.4662742614746094e-05, "ewc_loss_parallel": 1.049041748046875e-05, "grad_norm": 1.5021049976348877, "learning_rate": 3.3234421364985163e-06, "loss": 0.4789, "mean_token_accuracy": 0.8488883972167969, "num_tokens": 29725755.0, "step": 785 }, { "epoch": 0.09998727897214095, "ewc_loss": 0.011962890625, "ewc_loss_diag": 1.4662742614746094e-05, "ewc_loss_parallel": 1.049041748046875e-05, "grad_norm": 1.4863569736480713, "learning_rate": 3.3276812208562952e-06, "loss": 0.4981, "mean_token_accuracy": 0.8404551148414612, "num_tokens": 29760270.0, "step": 786 }, { "epoch": 0.10011448925073146, "ewc_loss": 0.0118408203125, "ewc_loss_diag": 1.4722347259521484e-05, "ewc_loss_parallel": 1.0371208190917969e-05, "grad_norm": 1.4797323942184448, "learning_rate": 3.331920305214074e-06, "loss": 0.5493, "mean_token_accuracy": 0.832482635974884, "num_tokens": 29797292.0, "step": 787 }, { "epoch": 0.10024169952932196, "ewc_loss": 0.01190185546875, "ewc_loss_diag": 1.4722347259521484e-05, "ewc_loss_parallel": 1.043081283569336e-05, "grad_norm": 1.4639031887054443, "learning_rate": 3.3361593895718526e-06, "loss": 0.4724, "mean_token_accuracy": 0.8532034158706665, "num_tokens": 29833265.0, "step": 788 }, { "epoch": 0.10036890980791248, "ewc_loss": 0.0118408203125, "ewc_loss_diag": 1.4722347259521484e-05, "ewc_loss_parallel": 1.0371208190917969e-05, "grad_norm": 1.4621645212173462, "learning_rate": 3.3403984739296315e-06, "loss": 0.5508, "mean_token_accuracy": 0.826330304145813, "num_tokens": 29876399.0, "step": 789 }, { "epoch": 0.100496120086503, "ewc_loss": 0.01190185546875, "ewc_loss_diag": 1.4722347259521484e-05, "ewc_loss_parallel": 1.043081283569336e-05, "grad_norm": 1.493332028388977, "learning_rate": 3.3446375582874103e-06, "loss": 0.5119, "mean_token_accuracy": 0.8392136096954346, "num_tokens": 29913883.0, "step": 790 }, { "epoch": 0.1006233303650935, "ewc_loss": 0.011962890625, "ewc_loss_diag": 1.4722347259521484e-05, "ewc_loss_parallel": 1.049041748046875e-05, "grad_norm": 1.5065488815307617, "learning_rate": 3.3488766426451892e-06, "loss": 0.5667, "mean_token_accuracy": 0.8276665210723877, "num_tokens": 29951691.0, "step": 791 }, { "epoch": 0.10075054064368401, "ewc_loss": 0.01190185546875, "ewc_loss_diag": 1.4781951904296875e-05, "ewc_loss_parallel": 1.043081283569336e-05, "grad_norm": 1.4562877416610718, "learning_rate": 3.3531157270029673e-06, "loss": 0.5298, "mean_token_accuracy": 0.8334733247756958, "num_tokens": 29992978.0, "step": 792 }, { "epoch": 0.10087775092227452, "ewc_loss": 0.011962890625, "ewc_loss_diag": 1.4781951904296875e-05, "ewc_loss_parallel": 1.049041748046875e-05, "grad_norm": 1.4609394073486328, "learning_rate": 3.357354811360746e-06, "loss": 0.5191, "mean_token_accuracy": 0.8323261737823486, "num_tokens": 30035602.0, "step": 793 }, { "epoch": 0.10100496120086502, "ewc_loss": 0.01202392578125, "ewc_loss_diag": 1.4781951904296875e-05, "ewc_loss_parallel": 1.055002212524414e-05, "grad_norm": 1.4940929412841797, "learning_rate": 3.361593895718525e-06, "loss": 0.5174, "mean_token_accuracy": 0.8320809006690979, "num_tokens": 30070063.0, "step": 794 }, { "epoch": 0.10113217147945554, "ewc_loss": 0.01202392578125, "ewc_loss_diag": 1.4781951904296875e-05, "ewc_loss_parallel": 1.055002212524414e-05, "grad_norm": 1.509856939315796, "learning_rate": 3.365832980076304e-06, "loss": 0.5566, "mean_token_accuracy": 0.8226264119148254, "num_tokens": 30101636.0, "step": 795 }, { "epoch": 0.10125938175804605, "ewc_loss": 0.0120849609375, "ewc_loss_diag": 1.4781951904296875e-05, "ewc_loss_parallel": 1.0609626770019531e-05, "grad_norm": 1.4793953895568848, "learning_rate": 3.3700720644340824e-06, "loss": 0.5149, "mean_token_accuracy": 0.8296260833740234, "num_tokens": 30137429.0, "step": 796 }, { "epoch": 0.10138659203663657, "ewc_loss": 0.01214599609375, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.0669231414794922e-05, "grad_norm": 1.6963638067245483, "learning_rate": 3.3743111487918613e-06, "loss": 0.4926, "mean_token_accuracy": 0.8446500301361084, "num_tokens": 30176173.0, "step": 797 }, { "epoch": 0.10151380231522707, "ewc_loss": 0.0128173828125, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.1324882507324219e-05, "grad_norm": 1.5060373544692993, "learning_rate": 3.37855023314964e-06, "loss": 0.4885, "mean_token_accuracy": 0.8444944024085999, "num_tokens": 30206588.0, "step": 798 }, { "epoch": 0.10164101259381758, "ewc_loss": 0.0118408203125, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.0371208190917969e-05, "grad_norm": 1.5146065950393677, "learning_rate": 3.382789317507419e-06, "loss": 0.5174, "mean_token_accuracy": 0.8359276056289673, "num_tokens": 30237345.0, "step": 799 }, { "epoch": 0.1017682228724081, "ewc_loss": 0.01220703125, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.0728836059570312e-05, "grad_norm": 1.4920575618743896, "learning_rate": 3.387028401865197e-06, "loss": 0.5298, "mean_token_accuracy": 0.8363627791404724, "num_tokens": 30274543.0, "step": 800 }, { "epoch": 0.1018954331509986, "ewc_loss": 0.01214599609375, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.0669231414794922e-05, "grad_norm": 1.4825398921966553, "learning_rate": 3.391267486222976e-06, "loss": 0.5041, "mean_token_accuracy": 0.8362894058227539, "num_tokens": 30312687.0, "step": 801 }, { "epoch": 0.10202264342958911, "ewc_loss": 0.01220703125, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.0728836059570312e-05, "grad_norm": 1.5117520093917847, "learning_rate": 3.395506570580755e-06, "loss": 0.5414, "mean_token_accuracy": 0.82625412940979, "num_tokens": 30343993.0, "step": 802 }, { "epoch": 0.10214985370817962, "ewc_loss": 0.0123291015625, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.0848045349121094e-05, "grad_norm": 1.500624656677246, "learning_rate": 3.3997456549385337e-06, "loss": 0.4884, "mean_token_accuracy": 0.8452135324478149, "num_tokens": 30378998.0, "step": 803 }, { "epoch": 0.10227706398677013, "ewc_loss": 0.01226806640625, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.0788440704345703e-05, "grad_norm": 1.5231897830963135, "learning_rate": 3.403984739296312e-06, "loss": 0.5568, "mean_token_accuracy": 0.8240017294883728, "num_tokens": 30412205.0, "step": 804 }, { "epoch": 0.10240427426536064, "ewc_loss": 0.01251220703125, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.1026859283447266e-05, "grad_norm": 1.4788322448730469, "learning_rate": 3.408223823654091e-06, "loss": 0.4768, "mean_token_accuracy": 0.8471612334251404, "num_tokens": 30448489.0, "step": 805 }, { "epoch": 0.10253148454395115, "ewc_loss": 0.01220703125, "ewc_loss_diag": 1.4901161193847656e-05, "ewc_loss_parallel": 1.0728836059570312e-05, "grad_norm": 1.4902608394622803, "learning_rate": 3.41246290801187e-06, "loss": 0.5302, "mean_token_accuracy": 0.8388813138008118, "num_tokens": 30488741.0, "step": 806 }, { "epoch": 0.10265869482254165, "ewc_loss": 0.012451171875, "ewc_loss_diag": 1.4960765838623047e-05, "ewc_loss_parallel": 1.0907649993896484e-05, "grad_norm": 1.485802412033081, "learning_rate": 3.416701992369648e-06, "loss": 0.5281, "mean_token_accuracy": 0.8401480913162231, "num_tokens": 30535715.0, "step": 807 }, { "epoch": 0.10278590510113217, "ewc_loss": 0.0123291015625, "ewc_loss_diag": 1.4960765838623047e-05, "ewc_loss_parallel": 1.0848045349121094e-05, "grad_norm": 1.4909614324569702, "learning_rate": 3.420941076727427e-06, "loss": 0.4846, "mean_token_accuracy": 0.8486360311508179, "num_tokens": 30574992.0, "step": 808 }, { "epoch": 0.10291311537972268, "ewc_loss": 0.012451171875, "ewc_loss_diag": 1.4960765838623047e-05, "ewc_loss_parallel": 1.0967254638671875e-05, "grad_norm": 1.4950963258743286, "learning_rate": 3.4251801610852058e-06, "loss": 0.555, "mean_token_accuracy": 0.8209670782089233, "num_tokens": 30615384.0, "step": 809 }, { "epoch": 0.1030403256583132, "ewc_loss": 0.0125732421875, "ewc_loss_diag": 1.5079975128173828e-05, "ewc_loss_parallel": 1.1026859283447266e-05, "grad_norm": 1.5101815462112427, "learning_rate": 3.4294192454429847e-06, "loss": 0.5045, "mean_token_accuracy": 0.8400872349739075, "num_tokens": 30651011.0, "step": 810 }, { "epoch": 0.1031675359369037, "ewc_loss": 0.0123291015625, "ewc_loss_diag": 1.5079975128173828e-05, "ewc_loss_parallel": 1.0788440704345703e-05, "grad_norm": 1.4964544773101807, "learning_rate": 3.433658329800763e-06, "loss": 0.4888, "mean_token_accuracy": 0.8456732034683228, "num_tokens": 30687213.0, "step": 811 }, { "epoch": 0.10329474621549421, "ewc_loss": 0.012451171875, "ewc_loss_diag": 1.5079975128173828e-05, "ewc_loss_parallel": 1.0907649993896484e-05, "grad_norm": 1.4940353631973267, "learning_rate": 3.437897414158542e-06, "loss": 0.5272, "mean_token_accuracy": 0.8345645070075989, "num_tokens": 30726950.0, "step": 812 }, { "epoch": 0.10342195649408473, "ewc_loss": 0.01251220703125, "ewc_loss_diag": 1.5139579772949219e-05, "ewc_loss_parallel": 1.0967254638671875e-05, "grad_norm": 1.4756091833114624, "learning_rate": 3.442136498516321e-06, "loss": 0.4633, "mean_token_accuracy": 0.851425051689148, "num_tokens": 30769359.0, "step": 813 }, { "epoch": 0.10354916677267523, "ewc_loss": 0.01239013671875, "ewc_loss_diag": 1.5139579772949219e-05, "ewc_loss_parallel": 1.0848045349121094e-05, "grad_norm": 1.646714448928833, "learning_rate": 3.4463755828740998e-06, "loss": 0.5293, "mean_token_accuracy": 0.8338611125946045, "num_tokens": 30812480.0, "step": 814 }, { "epoch": 0.10367637705126574, "ewc_loss": 0.01300048828125, "ewc_loss_diag": 1.5139579772949219e-05, "ewc_loss_parallel": 1.150369644165039e-05, "grad_norm": 1.5218276977539062, "learning_rate": 3.450614667231878e-06, "loss": 0.4707, "mean_token_accuracy": 0.847623348236084, "num_tokens": 30847287.0, "step": 815 }, { "epoch": 0.10380358732985626, "ewc_loss": 0.0123291015625, "ewc_loss_diag": 1.5139579772949219e-05, "ewc_loss_parallel": 1.0788440704345703e-05, "grad_norm": 1.4928752183914185, "learning_rate": 3.4548537515896567e-06, "loss": 0.5175, "mean_token_accuracy": 0.8371406197547913, "num_tokens": 30887050.0, "step": 816 }, { "epoch": 0.10393079760844676, "ewc_loss": 0.01251220703125, "ewc_loss_diag": 1.52587890625e-05, "ewc_loss_parallel": 1.0967254638671875e-05, "grad_norm": 1.485201358795166, "learning_rate": 3.4590928359474356e-06, "loss": 0.5173, "mean_token_accuracy": 0.8369319438934326, "num_tokens": 30926543.0, "step": 817 }, { "epoch": 0.10405800788703727, "ewc_loss": 0.012451171875, "ewc_loss_diag": 1.52587890625e-05, "ewc_loss_parallel": 1.0907649993896484e-05, "grad_norm": 1.498431921005249, "learning_rate": 3.4633319203052145e-06, "loss": 0.5371, "mean_token_accuracy": 0.8301776051521301, "num_tokens": 30966439.0, "step": 818 }, { "epoch": 0.10418521816562779, "ewc_loss": 0.01239013671875, "ewc_loss_diag": 1.52587890625e-05, "ewc_loss_parallel": 1.0848045349121094e-05, "grad_norm": 1.5145946741104126, "learning_rate": 3.467571004662993e-06, "loss": 0.4774, "mean_token_accuracy": 0.8476837277412415, "num_tokens": 30999860.0, "step": 819 }, { "epoch": 0.1043124284442183, "ewc_loss": 0.01251220703125, "ewc_loss_diag": 1.52587890625e-05, "ewc_loss_parallel": 1.0967254638671875e-05, "grad_norm": 1.5029324293136597, "learning_rate": 3.471810089020772e-06, "loss": 0.5229, "mean_token_accuracy": 0.8368313312530518, "num_tokens": 31037040.0, "step": 820 }, { "epoch": 0.1044396387228088, "ewc_loss": 0.012451171875, "ewc_loss_diag": 1.52587890625e-05, "ewc_loss_parallel": 1.0907649993896484e-05, "grad_norm": 1.5335825681686401, "learning_rate": 3.4760491733785507e-06, "loss": 0.5215, "mean_token_accuracy": 0.8301194310188293, "num_tokens": 31071508.0, "step": 821 }, { "epoch": 0.10456684900139931, "ewc_loss": 0.01263427734375, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1086463928222656e-05, "grad_norm": 1.5036590099334717, "learning_rate": 3.4802882577363296e-06, "loss": 0.5036, "mean_token_accuracy": 0.8354641199111938, "num_tokens": 31107358.0, "step": 822 }, { "epoch": 0.10469405927998983, "ewc_loss": 0.01251220703125, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.0967254638671875e-05, "grad_norm": 1.516900897026062, "learning_rate": 3.4845273420941076e-06, "loss": 0.5149, "mean_token_accuracy": 0.836266040802002, "num_tokens": 31144986.0, "step": 823 }, { "epoch": 0.10482126955858033, "ewc_loss": 0.0126953125, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1146068572998047e-05, "grad_norm": 1.5235031843185425, "learning_rate": 3.4887664264518865e-06, "loss": 0.5019, "mean_token_accuracy": 0.8351633548736572, "num_tokens": 31179709.0, "step": 824 }, { "epoch": 0.10494847983717084, "ewc_loss": 0.01263427734375, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1086463928222656e-05, "grad_norm": 1.4983726739883423, "learning_rate": 3.4930055108096654e-06, "loss": 0.5168, "mean_token_accuracy": 0.8383697271347046, "num_tokens": 31222242.0, "step": 825 }, { "epoch": 0.10507569011576136, "ewc_loss": 0.0125732421875, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1026859283447266e-05, "grad_norm": 1.4932456016540527, "learning_rate": 3.497244595167444e-06, "loss": 0.4915, "mean_token_accuracy": 0.8426017761230469, "num_tokens": 31264793.0, "step": 826 }, { "epoch": 0.10520290039435186, "ewc_loss": 0.0125732421875, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1026859283447266e-05, "grad_norm": 1.5349408388137817, "learning_rate": 3.5014836795252227e-06, "loss": 0.5517, "mean_token_accuracy": 0.8293867111206055, "num_tokens": 31300254.0, "step": 827 }, { "epoch": 0.10533011067294237, "ewc_loss": 0.01275634765625, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1205673217773438e-05, "grad_norm": 1.503951072692871, "learning_rate": 3.5057227638830016e-06, "loss": 0.4908, "mean_token_accuracy": 0.8433825373649597, "num_tokens": 31339786.0, "step": 828 }, { "epoch": 0.10545732095153289, "ewc_loss": 0.0125732421875, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1026859283447266e-05, "grad_norm": 1.5238840579986572, "learning_rate": 3.5099618482407805e-06, "loss": 0.5285, "mean_token_accuracy": 0.830538272857666, "num_tokens": 31375100.0, "step": 829 }, { "epoch": 0.10558453123012339, "ewc_loss": 0.0126953125, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1146068572998047e-05, "grad_norm": 1.509986162185669, "learning_rate": 3.514200932598559e-06, "loss": 0.5814, "mean_token_accuracy": 0.8193477392196655, "num_tokens": 31419000.0, "step": 830 }, { "epoch": 0.1057117415087139, "ewc_loss": 0.0126953125, "ewc_loss_diag": 1.537799835205078e-05, "ewc_loss_parallel": 1.1146068572998047e-05, "grad_norm": 1.5154385566711426, "learning_rate": 3.518440016956338e-06, "loss": 0.5228, "mean_token_accuracy": 0.8362046480178833, "num_tokens": 31459287.0, "step": 831 }, { "epoch": 0.10583895178730442, "ewc_loss": 0.01275634765625, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1205673217773438e-05, "grad_norm": 1.5018744468688965, "learning_rate": 3.5226791013141163e-06, "loss": 0.5306, "mean_token_accuracy": 0.8378860950469971, "num_tokens": 31502617.0, "step": 832 }, { "epoch": 0.10596616206589493, "ewc_loss": 0.01263427734375, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1086463928222656e-05, "grad_norm": 1.50816810131073, "learning_rate": 3.526918185671895e-06, "loss": 0.5451, "mean_token_accuracy": 0.8275267481803894, "num_tokens": 31545159.0, "step": 833 }, { "epoch": 0.10609337234448543, "ewc_loss": 0.01275634765625, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1205673217773438e-05, "grad_norm": 1.5318961143493652, "learning_rate": 3.5311572700296736e-06, "loss": 0.4708, "mean_token_accuracy": 0.8519821166992188, "num_tokens": 31579193.0, "step": 834 }, { "epoch": 0.10622058262307595, "ewc_loss": 0.01287841796875, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1324882507324219e-05, "grad_norm": 1.5133010149002075, "learning_rate": 3.5353963543874525e-06, "loss": 0.4892, "mean_token_accuracy": 0.8461023569107056, "num_tokens": 31618995.0, "step": 835 }, { "epoch": 0.10634779290166646, "ewc_loss": 0.01275634765625, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1205673217773438e-05, "grad_norm": 1.5273475646972656, "learning_rate": 3.5396354387452314e-06, "loss": 0.5159, "mean_token_accuracy": 0.8369368314743042, "num_tokens": 31657024.0, "step": 836 }, { "epoch": 0.10647500318025696, "ewc_loss": 0.0128173828125, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1265277862548828e-05, "grad_norm": 1.492746353149414, "learning_rate": 3.5438745231030103e-06, "loss": 0.5066, "mean_token_accuracy": 0.8401061296463013, "num_tokens": 31700521.0, "step": 837 }, { "epoch": 0.10660221345884748, "ewc_loss": 0.0128173828125, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1265277862548828e-05, "grad_norm": 1.5290549993515015, "learning_rate": 3.5481136074607888e-06, "loss": 0.4751, "mean_token_accuracy": 0.8522331714630127, "num_tokens": 31735942.0, "step": 838 }, { "epoch": 0.10672942373743799, "ewc_loss": 0.01287841796875, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1324882507324219e-05, "grad_norm": 1.5132513046264648, "learning_rate": 3.5523526918185676e-06, "loss": 0.5036, "mean_token_accuracy": 0.8423700332641602, "num_tokens": 31774092.0, "step": 839 }, { "epoch": 0.10685663401602849, "ewc_loss": 0.01287841796875, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1324882507324219e-05, "grad_norm": 1.5178508758544922, "learning_rate": 3.5565917761763465e-06, "loss": 0.4828, "mean_token_accuracy": 0.8473994135856628, "num_tokens": 31816292.0, "step": 840 }, { "epoch": 0.106983844294619, "ewc_loss": 0.01275634765625, "ewc_loss_diag": 1.5497207641601562e-05, "ewc_loss_parallel": 1.1205673217773438e-05, "grad_norm": 1.5513273477554321, "learning_rate": 3.560830860534125e-06, "loss": 0.5037, "mean_token_accuracy": 0.841741681098938, "num_tokens": 31848641.0, "step": 841 }, { "epoch": 0.10711105457320952, "ewc_loss": 0.0130615234375, "ewc_loss_diag": 1.5616416931152344e-05, "ewc_loss_parallel": 1.150369644165039e-05, "grad_norm": 1.5600987672805786, "learning_rate": 3.5650699448919035e-06, "loss": 0.5539, "mean_token_accuracy": 0.8263020515441895, "num_tokens": 31881292.0, "step": 842 }, { "epoch": 0.10723826485180002, "ewc_loss": 0.012939453125, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.1324882507324219e-05, "grad_norm": 1.5346685647964478, "learning_rate": 3.5693090292496823e-06, "loss": 0.512, "mean_token_accuracy": 0.8417354822158813, "num_tokens": 31915810.0, "step": 843 }, { "epoch": 0.10736547513039053, "ewc_loss": 0.01300048828125, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.138448715209961e-05, "grad_norm": 1.5163993835449219, "learning_rate": 3.5735481136074612e-06, "loss": 0.4953, "mean_token_accuracy": 0.8434575796127319, "num_tokens": 31953621.0, "step": 844 }, { "epoch": 0.10749268540898105, "ewc_loss": 0.012939453125, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.1324882507324219e-05, "grad_norm": 1.5413709878921509, "learning_rate": 3.5777871979652397e-06, "loss": 0.5366, "mean_token_accuracy": 0.8306158781051636, "num_tokens": 31988738.0, "step": 845 }, { "epoch": 0.10761989568757156, "ewc_loss": 0.0130615234375, "ewc_loss_diag": 1.5735626220703125e-05, "ewc_loss_parallel": 1.1444091796875e-05, "grad_norm": 1.5084586143493652, "learning_rate": 3.5820262823230186e-06, "loss": 0.4805, "mean_token_accuracy": 0.852428674697876, "num_tokens": 32029184.0, "step": 846 }, { "epoch": 0.10774710596616206, "ewc_loss": 0.0130615234375, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1444091796875e-05, "grad_norm": 1.5573564767837524, "learning_rate": 3.5862653666807975e-06, "loss": 0.5612, "mean_token_accuracy": 0.8221147060394287, "num_tokens": 32064413.0, "step": 847 }, { "epoch": 0.10787431624475258, "ewc_loss": 0.01324462890625, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1682510375976562e-05, "grad_norm": 1.532731056213379, "learning_rate": 3.5905044510385763e-06, "loss": 0.5262, "mean_token_accuracy": 0.8372516632080078, "num_tokens": 32104043.0, "step": 848 }, { "epoch": 0.10800152652334309, "ewc_loss": 0.0130615234375, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1444091796875e-05, "grad_norm": 1.5429879426956177, "learning_rate": 3.5947435353963544e-06, "loss": 0.5126, "mean_token_accuracy": 0.8403123021125793, "num_tokens": 32140956.0, "step": 849 }, { "epoch": 0.10812873680193359, "ewc_loss": 0.01312255859375, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1563301086425781e-05, "grad_norm": 1.5506856441497803, "learning_rate": 3.5989826197541333e-06, "loss": 0.561, "mean_token_accuracy": 0.8247206807136536, "num_tokens": 32181368.0, "step": 850 }, { "epoch": 0.1082559470805241, "ewc_loss": 0.0133056640625, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1742115020751953e-05, "grad_norm": 1.5238442420959473, "learning_rate": 3.603221704111912e-06, "loss": 0.5105, "mean_token_accuracy": 0.8353736996650696, "num_tokens": 32222714.0, "step": 851 }, { "epoch": 0.10838315735911462, "ewc_loss": 0.01312255859375, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1563301086425781e-05, "grad_norm": 1.5099232196807861, "learning_rate": 3.607460788469691e-06, "loss": 0.4762, "mean_token_accuracy": 0.8527116775512695, "num_tokens": 32267870.0, "step": 852 }, { "epoch": 0.10851036763770512, "ewc_loss": 0.01324462890625, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1682510375976562e-05, "grad_norm": 1.5192972421646118, "learning_rate": 3.6116998728274695e-06, "loss": 0.451, "mean_token_accuracy": 0.8558235168457031, "num_tokens": 32307408.0, "step": 853 }, { "epoch": 0.10863757791629564, "ewc_loss": 0.01324462890625, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1682510375976562e-05, "grad_norm": 1.5654652118682861, "learning_rate": 3.6159389571852484e-06, "loss": 0.5855, "mean_token_accuracy": 0.822453498840332, "num_tokens": 32343503.0, "step": 854 }, { "epoch": 0.10876478819488615, "ewc_loss": 0.01348876953125, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1920928955078125e-05, "grad_norm": 1.5247725248336792, "learning_rate": 3.6201780415430273e-06, "loss": 0.528, "mean_token_accuracy": 0.8349663019180298, "num_tokens": 32386375.0, "step": 855 }, { "epoch": 0.10889199847347665, "ewc_loss": 0.01324462890625, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1682510375976562e-05, "grad_norm": 1.5526105165481567, "learning_rate": 3.624417125900806e-06, "loss": 0.5409, "mean_token_accuracy": 0.830160915851593, "num_tokens": 32424370.0, "step": 856 }, { "epoch": 0.10901920875206716, "ewc_loss": 0.013427734375, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1861324310302734e-05, "grad_norm": 1.5486699342727661, "learning_rate": 3.628656210258584e-06, "loss": 0.5485, "mean_token_accuracy": 0.8262367248535156, "num_tokens": 32461998.0, "step": 857 }, { "epoch": 0.10914641903065768, "ewc_loss": 0.01336669921875, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1801719665527344e-05, "grad_norm": 1.5339152812957764, "learning_rate": 3.632895294616363e-06, "loss": 0.4854, "mean_token_accuracy": 0.8476614952087402, "num_tokens": 32498124.0, "step": 858 }, { "epoch": 0.1092736293092482, "ewc_loss": 0.01336669921875, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1801719665527344e-05, "grad_norm": 1.53246009349823, "learning_rate": 3.637134378974142e-06, "loss": 0.4795, "mean_token_accuracy": 0.8468127846717834, "num_tokens": 32537735.0, "step": 859 }, { "epoch": 0.1094008395878387, "ewc_loss": 0.01336669921875, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1801719665527344e-05, "grad_norm": 1.5450129508972168, "learning_rate": 3.641373463331921e-06, "loss": 0.5241, "mean_token_accuracy": 0.8365118503570557, "num_tokens": 32576219.0, "step": 860 }, { "epoch": 0.10952804986642921, "ewc_loss": 0.01348876953125, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1920928955078125e-05, "grad_norm": 1.5495831966400146, "learning_rate": 3.6456125476896993e-06, "loss": 0.4857, "mean_token_accuracy": 0.8395135402679443, "num_tokens": 32612868.0, "step": 861 }, { "epoch": 0.10965526014501972, "ewc_loss": 0.013427734375, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1861324310302734e-05, "grad_norm": 1.5254250764846802, "learning_rate": 3.649851632047478e-06, "loss": 0.4805, "mean_token_accuracy": 0.8508632779121399, "num_tokens": 32655412.0, "step": 862 }, { "epoch": 0.10978247042361022, "ewc_loss": 0.01324462890625, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1682510375976562e-05, "grad_norm": 1.5360113382339478, "learning_rate": 3.654090716405257e-06, "loss": 0.4477, "mean_token_accuracy": 0.8590680360794067, "num_tokens": 32692579.0, "step": 863 }, { "epoch": 0.10990968070220074, "ewc_loss": 0.01348876953125, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1920928955078125e-05, "grad_norm": 1.5527582168579102, "learning_rate": 3.658329800763035e-06, "loss": 0.5446, "mean_token_accuracy": 0.8250581622123718, "num_tokens": 32731642.0, "step": 864 }, { "epoch": 0.11003689098079125, "ewc_loss": 0.013427734375, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1861324310302734e-05, "grad_norm": 1.5601564645767212, "learning_rate": 3.662568885120814e-06, "loss": 0.5112, "mean_token_accuracy": 0.8367549777030945, "num_tokens": 32765707.0, "step": 865 }, { "epoch": 0.11016410125938175, "ewc_loss": 0.0135498046875, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1980533599853516e-05, "grad_norm": 1.557778000831604, "learning_rate": 3.666807969478593e-06, "loss": 0.5424, "mean_token_accuracy": 0.8263812065124512, "num_tokens": 32804613.0, "step": 866 }, { "epoch": 0.11029131153797227, "ewc_loss": 0.0135498046875, "ewc_loss_diag": 1.5854835510253906e-05, "ewc_loss_parallel": 1.1980533599853516e-05, "grad_norm": 1.5545529127120972, "learning_rate": 3.6710470538363718e-06, "loss": 0.5274, "mean_token_accuracy": 0.8322532176971436, "num_tokens": 32841198.0, "step": 867 }, { "epoch": 0.11041852181656278, "ewc_loss": 0.01361083984375, "ewc_loss_diag": 1.609325408935547e-05, "ewc_loss_parallel": 1.2040138244628906e-05, "grad_norm": 1.5482566356658936, "learning_rate": 3.6752861381941502e-06, "loss": 0.4922, "mean_token_accuracy": 0.8446155786514282, "num_tokens": 32878241.0, "step": 868 }, { "epoch": 0.11054573209515328, "ewc_loss": 0.013671875, "ewc_loss_diag": 1.609325408935547e-05, "ewc_loss_parallel": 1.2099742889404297e-05, "grad_norm": 1.568982720375061, "learning_rate": 3.679525222551929e-06, "loss": 0.5533, "mean_token_accuracy": 0.8277521729469299, "num_tokens": 32915510.0, "step": 869 }, { "epoch": 0.1106729423737438, "ewc_loss": 0.01373291015625, "ewc_loss_diag": 1.609325408935547e-05, "ewc_loss_parallel": 1.2159347534179688e-05, "grad_norm": 1.5357184410095215, "learning_rate": 3.683764306909708e-06, "loss": 0.4525, "mean_token_accuracy": 0.8552119731903076, "num_tokens": 32953371.0, "step": 870 }, { "epoch": 0.11080015265233431, "ewc_loss": 0.013671875, "ewc_loss_diag": 1.609325408935547e-05, "ewc_loss_parallel": 1.2099742889404297e-05, "grad_norm": 1.5402237176895142, "learning_rate": 3.688003391267487e-06, "loss": 0.5134, "mean_token_accuracy": 0.841343104839325, "num_tokens": 32998167.0, "step": 871 }, { "epoch": 0.11092736293092482, "ewc_loss": 0.0137939453125, "ewc_loss_diag": 1.609325408935547e-05, "ewc_loss_parallel": 1.2218952178955078e-05, "grad_norm": 1.5237914323806763, "learning_rate": 3.692242475625265e-06, "loss": 0.4835, "mean_token_accuracy": 0.8455475568771362, "num_tokens": 33046660.0, "step": 872 }, { "epoch": 0.11105457320951533, "ewc_loss": 0.0137939453125, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.2159347534179688e-05, "grad_norm": 1.586122751235962, "learning_rate": 3.696481559983044e-06, "loss": 0.5263, "mean_token_accuracy": 0.830772876739502, "num_tokens": 33080672.0, "step": 873 }, { "epoch": 0.11118178348810584, "ewc_loss": 0.0137939453125, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.2218952178955078e-05, "grad_norm": 1.5659325122833252, "learning_rate": 3.7007206443408227e-06, "loss": 0.4646, "mean_token_accuracy": 0.8509774208068848, "num_tokens": 33117023.0, "step": 874 }, { "epoch": 0.11130899376669635, "ewc_loss": 0.0137939453125, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.2218952178955078e-05, "grad_norm": 1.5632134675979614, "learning_rate": 3.7049597286986016e-06, "loss": 0.582, "mean_token_accuracy": 0.8186184167861938, "num_tokens": 33154843.0, "step": 875 }, { "epoch": 0.11143620404528685, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.2278556823730469e-05, "grad_norm": 1.5652928352355957, "learning_rate": 3.70919881305638e-06, "loss": 0.5358, "mean_token_accuracy": 0.8334031105041504, "num_tokens": 33193500.0, "step": 876 }, { "epoch": 0.11156341432387737, "ewc_loss": 0.0137939453125, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.2218952178955078e-05, "grad_norm": 1.5370196104049683, "learning_rate": 3.713437897414159e-06, "loss": 0.5265, "mean_token_accuracy": 0.8350634574890137, "num_tokens": 33240528.0, "step": 877 }, { "epoch": 0.11169062460246788, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.2278556823730469e-05, "grad_norm": 1.5719265937805176, "learning_rate": 3.717676981771938e-06, "loss": 0.5333, "mean_token_accuracy": 0.8317015171051025, "num_tokens": 33276959.0, "step": 878 }, { "epoch": 0.11181783488105838, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.2278556823730469e-05, "grad_norm": 1.5565472841262817, "learning_rate": 3.721916066129716e-06, "loss": 0.4461, "mean_token_accuracy": 0.8578537702560425, "num_tokens": 33313573.0, "step": 879 }, { "epoch": 0.1119450451596489, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.233816146850586e-05, "grad_norm": 1.588427186012268, "learning_rate": 3.7261551504874947e-06, "loss": 0.519, "mean_token_accuracy": 0.8395584225654602, "num_tokens": 33347408.0, "step": 880 }, { "epoch": 0.11207225543823941, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.621246337890625e-05, "ewc_loss_parallel": 1.2278556823730469e-05, "grad_norm": 1.5590336322784424, "learning_rate": 3.7303942348452736e-06, "loss": 0.4516, "mean_token_accuracy": 0.8544071912765503, "num_tokens": 33381894.0, "step": 881 }, { "epoch": 0.11219946571682991, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.633167266845703e-05, "ewc_loss_parallel": 1.2278556823730469e-05, "grad_norm": 1.561496376991272, "learning_rate": 3.7346333192030525e-06, "loss": 0.4664, "mean_token_accuracy": 0.8539113998413086, "num_tokens": 33421746.0, "step": 882 }, { "epoch": 0.11232667599542043, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.633167266845703e-05, "ewc_loss_parallel": 1.2278556823730469e-05, "grad_norm": 1.5425827503204346, "learning_rate": 3.738872403560831e-06, "loss": 0.4874, "mean_token_accuracy": 0.8500776886940002, "num_tokens": 33465203.0, "step": 883 }, { "epoch": 0.11245388627401094, "ewc_loss": 0.01397705078125, "ewc_loss_diag": 1.633167266845703e-05, "ewc_loss_parallel": 1.233816146850586e-05, "grad_norm": 1.5852162837982178, "learning_rate": 3.74311148791861e-06, "loss": 0.5276, "mean_token_accuracy": 0.8337001204490662, "num_tokens": 33500333.0, "step": 884 }, { "epoch": 0.11258109655260146, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.6450881958007812e-05, "ewc_loss_parallel": 1.2278556823730469e-05, "grad_norm": 1.5792994499206543, "learning_rate": 3.7473505722763887e-06, "loss": 0.5459, "mean_token_accuracy": 0.8271442651748657, "num_tokens": 33535180.0, "step": 885 }, { "epoch": 0.11270830683119196, "ewc_loss": 0.01397705078125, "ewc_loss_diag": 1.6450881958007812e-05, "ewc_loss_parallel": 1.233816146850586e-05, "grad_norm": 1.5737719535827637, "learning_rate": 3.7515896566341676e-06, "loss": 0.4975, "mean_token_accuracy": 0.8412398099899292, "num_tokens": 33571298.0, "step": 886 }, { "epoch": 0.11283551710978247, "ewc_loss": 0.013916015625, "ewc_loss_diag": 1.6450881958007812e-05, "ewc_loss_parallel": 1.2278556823730469e-05, "grad_norm": 1.5541729927062988, "learning_rate": 3.7558287409919456e-06, "loss": 0.5226, "mean_token_accuracy": 0.8331565856933594, "num_tokens": 33613693.0, "step": 887 }, { "epoch": 0.11296272738837299, "ewc_loss": 0.01397705078125, "ewc_loss_diag": 1.6450881958007812e-05, "ewc_loss_parallel": 1.233816146850586e-05, "grad_norm": 1.5726789236068726, "learning_rate": 3.7600678253497245e-06, "loss": 0.4874, "mean_token_accuracy": 0.8481377363204956, "num_tokens": 33649760.0, "step": 888 }, { "epoch": 0.11308993766696349, "ewc_loss": 0.01397705078125, "ewc_loss_diag": 1.6450881958007812e-05, "ewc_loss_parallel": 1.233816146850586e-05, "grad_norm": 1.5372627973556519, "learning_rate": 3.7643069097075034e-06, "loss": 0.4496, "mean_token_accuracy": 0.8580436110496521, "num_tokens": 33692441.0, "step": 889 }, { "epoch": 0.113217147945554, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.6450881958007812e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.608588695526123, "learning_rate": 3.7685459940652823e-06, "loss": 0.5016, "mean_token_accuracy": 0.8404654264450073, "num_tokens": 33723034.0, "step": 890 }, { "epoch": 0.11334435822414451, "ewc_loss": 0.0140380859375, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.239776611328125e-05, "grad_norm": 1.5586676597595215, "learning_rate": 3.7727850784230608e-06, "loss": 0.5167, "mean_token_accuracy": 0.8402525186538696, "num_tokens": 33764102.0, "step": 891 }, { "epoch": 0.11347156850273502, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.5686908960342407, "learning_rate": 3.7770241627808396e-06, "loss": 0.4764, "mean_token_accuracy": 0.8488349914550781, "num_tokens": 33803293.0, "step": 892 }, { "epoch": 0.11359877878132553, "ewc_loss": 0.01397705078125, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.233816146850586e-05, "grad_norm": 1.566957712173462, "learning_rate": 3.7812632471386185e-06, "loss": 0.4771, "mean_token_accuracy": 0.8489139676094055, "num_tokens": 33841681.0, "step": 893 }, { "epoch": 0.11372598905991604, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.5857133865356445, "learning_rate": 3.7855023314963974e-06, "loss": 0.5009, "mean_token_accuracy": 0.8360157608985901, "num_tokens": 33877207.0, "step": 894 }, { "epoch": 0.11385319933850654, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.5778712034225464, "learning_rate": 3.7897414158541754e-06, "loss": 0.5152, "mean_token_accuracy": 0.8345842361450195, "num_tokens": 33917703.0, "step": 895 }, { "epoch": 0.11398040961709706, "ewc_loss": 0.01416015625, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.2516975402832031e-05, "grad_norm": 1.5788938999176025, "learning_rate": 3.7939805002119543e-06, "loss": 0.5028, "mean_token_accuracy": 0.8430025577545166, "num_tokens": 33957007.0, "step": 896 }, { "epoch": 0.11410761989568757, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.5623666048049927, "learning_rate": 3.7982195845697332e-06, "loss": 0.4371, "mean_token_accuracy": 0.8595082759857178, "num_tokens": 33992615.0, "step": 897 }, { "epoch": 0.11423483017427809, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.6047145128250122, "learning_rate": 3.8024586689275117e-06, "loss": 0.5563, "mean_token_accuracy": 0.8297964334487915, "num_tokens": 34032049.0, "step": 898 }, { "epoch": 0.11436204045286859, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.5701311826705933, "learning_rate": 3.8066977532852906e-06, "loss": 0.486, "mean_token_accuracy": 0.8440842032432556, "num_tokens": 34065978.0, "step": 899 }, { "epoch": 0.1144892507314591, "ewc_loss": 0.01416015625, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.2516975402832031e-05, "grad_norm": 1.577157974243164, "learning_rate": 3.8109368376430694e-06, "loss": 0.4652, "mean_token_accuracy": 0.8536657094955444, "num_tokens": 34103454.0, "step": 900 }, { "epoch": 0.11461646101004962, "ewc_loss": 0.01409912109375, "ewc_loss_diag": 1.6689300537109375e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.5968270301818848, "learning_rate": 3.815175922000848e-06, "loss": 0.4426, "mean_token_accuracy": 0.8548862934112549, "num_tokens": 34137970.0, "step": 901 }, { "epoch": 0.11474367128864012, "ewc_loss": 0.0142822265625, "ewc_loss_diag": 1.6808509826660156e-05, "ewc_loss_parallel": 1.2576580047607422e-05, "grad_norm": 1.57504141330719, "learning_rate": 3.819415006358627e-06, "loss": 0.5373, "mean_token_accuracy": 0.8331773281097412, "num_tokens": 34176997.0, "step": 902 }, { "epoch": 0.11487088156723063, "ewc_loss": 0.01416015625, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.245737075805664e-05, "grad_norm": 1.581525444984436, "learning_rate": 3.823654090716406e-06, "loss": 0.5082, "mean_token_accuracy": 0.837155818939209, "num_tokens": 34215790.0, "step": 903 }, { "epoch": 0.11499809184582115, "ewc_loss": 0.01422119140625, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.2516975402832031e-05, "grad_norm": 1.5983225107192993, "learning_rate": 3.8278931750741846e-06, "loss": 0.5187, "mean_token_accuracy": 0.8433468341827393, "num_tokens": 34255212.0, "step": 904 }, { "epoch": 0.11512530212441165, "ewc_loss": 0.01434326171875, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.2636184692382812e-05, "grad_norm": 1.6061731576919556, "learning_rate": 3.8321322594319634e-06, "loss": 0.4535, "mean_token_accuracy": 0.856247067451477, "num_tokens": 34288432.0, "step": 905 }, { "epoch": 0.11525251240300216, "ewc_loss": 0.01422119140625, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.2516975402832031e-05, "grad_norm": 1.5885913372039795, "learning_rate": 3.8363713437897415e-06, "loss": 0.5618, "mean_token_accuracy": 0.821763813495636, "num_tokens": 34328991.0, "step": 906 }, { "epoch": 0.11537972268159268, "ewc_loss": 0.01434326171875, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.2636184692382812e-05, "grad_norm": 1.6281800270080566, "learning_rate": 3.84061042814752e-06, "loss": 0.5507, "mean_token_accuracy": 0.8309437036514282, "num_tokens": 34366712.0, "step": 907 }, { "epoch": 0.11550693296018319, "ewc_loss": 0.01434326171875, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.2636184692382812e-05, "grad_norm": 1.5775693655014038, "learning_rate": 3.844849512505299e-06, "loss": 0.537, "mean_token_accuracy": 0.8289456963539124, "num_tokens": 34405757.0, "step": 908 }, { "epoch": 0.11563414323877369, "ewc_loss": 0.01434326171875, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.2636184692382812e-05, "grad_norm": 1.5843477249145508, "learning_rate": 3.849088596863078e-06, "loss": 0.492, "mean_token_accuracy": 0.8458471894264221, "num_tokens": 34448019.0, "step": 909 }, { "epoch": 0.1157613535173642, "ewc_loss": 0.01446533203125, "ewc_loss_diag": 1.6927719116210938e-05, "ewc_loss_parallel": 1.2755393981933594e-05, "grad_norm": 1.5806926488876343, "learning_rate": 3.853327681220856e-06, "loss": 0.5095, "mean_token_accuracy": 0.8402620553970337, "num_tokens": 34491237.0, "step": 910 }, { "epoch": 0.11588856379595472, "ewc_loss": 0.01434326171875, "ewc_loss_diag": 1.704692840576172e-05, "ewc_loss_parallel": 1.2636184692382812e-05, "grad_norm": 1.6174523830413818, "learning_rate": 3.857566765578635e-06, "loss": 0.4778, "mean_token_accuracy": 0.8479752540588379, "num_tokens": 34522842.0, "step": 911 }, { "epoch": 0.11601577407454522, "ewc_loss": 0.01446533203125, "ewc_loss_diag": 1.704692840576172e-05, "ewc_loss_parallel": 1.2755393981933594e-05, "grad_norm": 1.5776636600494385, "learning_rate": 3.861805849936414e-06, "loss": 0.5016, "mean_token_accuracy": 0.8447983264923096, "num_tokens": 34560551.0, "step": 912 }, { "epoch": 0.11614298435313573, "ewc_loss": 0.0145263671875, "ewc_loss_diag": 1.704692840576172e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.5987004041671753, "learning_rate": 3.866044934294193e-06, "loss": 0.5495, "mean_token_accuracy": 0.8329727053642273, "num_tokens": 34603064.0, "step": 913 }, { "epoch": 0.11627019463172625, "ewc_loss": 0.01434326171875, "ewc_loss_diag": 1.704692840576172e-05, "ewc_loss_parallel": 1.2636184692382812e-05, "grad_norm": 1.5979403257369995, "learning_rate": 3.870284018651972e-06, "loss": 0.526, "mean_token_accuracy": 0.8344576358795166, "num_tokens": 34640283.0, "step": 914 }, { "epoch": 0.11639740491031675, "ewc_loss": 0.01446533203125, "ewc_loss_diag": 1.704692840576172e-05, "ewc_loss_parallel": 1.2755393981933594e-05, "grad_norm": 1.5916011333465576, "learning_rate": 3.874523103009751e-06, "loss": 0.4754, "mean_token_accuracy": 0.8511337637901306, "num_tokens": 34679840.0, "step": 915 }, { "epoch": 0.11652461518890726, "ewc_loss": 0.0145263671875, "ewc_loss_diag": 1.71661376953125e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.5883740186691284, "learning_rate": 3.8787621873675295e-06, "loss": 0.4633, "mean_token_accuracy": 0.8529553413391113, "num_tokens": 34717045.0, "step": 916 }, { "epoch": 0.11665182546749778, "ewc_loss": 0.0145263671875, "ewc_loss_diag": 1.71661376953125e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.5844762325286865, "learning_rate": 3.8830012717253075e-06, "loss": 0.4662, "mean_token_accuracy": 0.8531386852264404, "num_tokens": 34755849.0, "step": 917 }, { "epoch": 0.11677903574608828, "ewc_loss": 0.01458740234375, "ewc_loss_diag": 1.71661376953125e-05, "ewc_loss_parallel": 1.2874603271484375e-05, "grad_norm": 1.5937950611114502, "learning_rate": 3.887240356083086e-06, "loss": 0.441, "mean_token_accuracy": 0.8579339385032654, "num_tokens": 34795202.0, "step": 918 }, { "epoch": 0.11690624602467879, "ewc_loss": 0.0145263671875, "ewc_loss_diag": 1.71661376953125e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.6406278610229492, "learning_rate": 3.891479440440865e-06, "loss": 0.4791, "mean_token_accuracy": 0.8504418730735779, "num_tokens": 34824279.0, "step": 919 }, { "epoch": 0.1170334563032693, "ewc_loss": 0.01446533203125, "ewc_loss_diag": 1.71661376953125e-05, "ewc_loss_parallel": 1.2755393981933594e-05, "grad_norm": 1.6256918907165527, "learning_rate": 3.895718524798644e-06, "loss": 0.5202, "mean_token_accuracy": 0.8356650471687317, "num_tokens": 34857631.0, "step": 920 }, { "epoch": 0.11716066658185982, "ewc_loss": 0.0145263671875, "ewc_loss_diag": 1.71661376953125e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.6016876697540283, "learning_rate": 3.899957609156422e-06, "loss": 0.4908, "mean_token_accuracy": 0.8455914258956909, "num_tokens": 34896969.0, "step": 921 }, { "epoch": 0.11728787686045032, "ewc_loss": 0.0145263671875, "ewc_loss_diag": 1.7404556274414062e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.5799280405044556, "learning_rate": 3.904196693514201e-06, "loss": 0.5087, "mean_token_accuracy": 0.8375570178031921, "num_tokens": 34942177.0, "step": 922 }, { "epoch": 0.11741508713904084, "ewc_loss": 0.0145263671875, "ewc_loss_diag": 1.7404556274414062e-05, "ewc_loss_parallel": 1.2755393981933594e-05, "grad_norm": 1.625863790512085, "learning_rate": 3.90843577787198e-06, "loss": 0.5778, "mean_token_accuracy": 0.8197633028030396, "num_tokens": 34980440.0, "step": 923 }, { "epoch": 0.11754229741763135, "ewc_loss": 0.0145263671875, "ewc_loss_diag": 1.7404556274414062e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.6198844909667969, "learning_rate": 3.912674862229759e-06, "loss": 0.5286, "mean_token_accuracy": 0.8331900835037231, "num_tokens": 35015343.0, "step": 924 }, { "epoch": 0.11766950769622185, "ewc_loss": 0.01458740234375, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.6275262832641602, "learning_rate": 3.916913946587537e-06, "loss": 0.5201, "mean_token_accuracy": 0.836418628692627, "num_tokens": 35050265.0, "step": 925 }, { "epoch": 0.11779671797481236, "ewc_loss": 0.0146484375, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.2874603271484375e-05, "grad_norm": 1.603304147720337, "learning_rate": 3.921153030945316e-06, "loss": 0.5436, "mean_token_accuracy": 0.8285021781921387, "num_tokens": 35089235.0, "step": 926 }, { "epoch": 0.11792392825340288, "ewc_loss": 0.01458740234375, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.2814998626708984e-05, "grad_norm": 1.6293314695358276, "learning_rate": 3.925392115303095e-06, "loss": 0.4915, "mean_token_accuracy": 0.844569206237793, "num_tokens": 35122887.0, "step": 927 }, { "epoch": 0.11805113853199338, "ewc_loss": 0.0146484375, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.2874603271484375e-05, "grad_norm": 1.6147364377975464, "learning_rate": 3.9296311996608736e-06, "loss": 0.4955, "mean_token_accuracy": 0.8415961265563965, "num_tokens": 35158686.0, "step": 928 }, { "epoch": 0.1181783488105839, "ewc_loss": 0.0146484375, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.2874603271484375e-05, "grad_norm": 1.6086328029632568, "learning_rate": 3.9338702840186524e-06, "loss": 0.4925, "mean_token_accuracy": 0.8419528007507324, "num_tokens": 35195215.0, "step": 929 }, { "epoch": 0.11830555908917441, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.7523765563964844e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.5940766334533691, "learning_rate": 3.938109368376431e-06, "loss": 0.4972, "mean_token_accuracy": 0.8441861867904663, "num_tokens": 35236499.0, "step": 930 }, { "epoch": 0.11843276936776491, "ewc_loss": 0.01470947265625, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2934207916259766e-05, "grad_norm": 1.6117632389068604, "learning_rate": 3.94234845273421e-06, "loss": 0.5499, "mean_token_accuracy": 0.8274291753768921, "num_tokens": 35278809.0, "step": 931 }, { "epoch": 0.11855997964635542, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.612455129623413, "learning_rate": 3.946587537091989e-06, "loss": 0.4964, "mean_token_accuracy": 0.841107189655304, "num_tokens": 35318342.0, "step": 932 }, { "epoch": 0.11868718992494594, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.6025351285934448, "learning_rate": 3.950826621449767e-06, "loss": 0.4948, "mean_token_accuracy": 0.8429083824157715, "num_tokens": 35361802.0, "step": 933 }, { "epoch": 0.11881440020353645, "ewc_loss": 0.01470947265625, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2934207916259766e-05, "grad_norm": 1.6090428829193115, "learning_rate": 3.955065705807546e-06, "loss": 0.4672, "mean_token_accuracy": 0.8521807193756104, "num_tokens": 35401244.0, "step": 934 }, { "epoch": 0.11894161048212695, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.6349480152130127, "learning_rate": 3.959304790165325e-06, "loss": 0.5059, "mean_token_accuracy": 0.8424311876296997, "num_tokens": 35438557.0, "step": 935 }, { "epoch": 0.11906882076071747, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.633614420890808, "learning_rate": 3.963543874523103e-06, "loss": 0.4948, "mean_token_accuracy": 0.8450181484222412, "num_tokens": 35471686.0, "step": 936 }, { "epoch": 0.11919603103930798, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.5939033031463623, "learning_rate": 3.967782958880882e-06, "loss": 0.5206, "mean_token_accuracy": 0.8356812596321106, "num_tokens": 35514288.0, "step": 937 }, { "epoch": 0.11932324131789848, "ewc_loss": 0.01470947265625, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2934207916259766e-05, "grad_norm": 1.6416579484939575, "learning_rate": 3.972022043238661e-06, "loss": 0.5285, "mean_token_accuracy": 0.8335316777229309, "num_tokens": 35547588.0, "step": 938 }, { "epoch": 0.119450451596489, "ewc_loss": 0.01495361328125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3172626495361328e-05, "grad_norm": 1.6160123348236084, "learning_rate": 3.97626112759644e-06, "loss": 0.4667, "mean_token_accuracy": 0.8535376191139221, "num_tokens": 35582099.0, "step": 939 }, { "epoch": 0.11957766187507951, "ewc_loss": 0.01483154296875, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3053417205810547e-05, "grad_norm": 1.62013840675354, "learning_rate": 3.980500211954218e-06, "loss": 0.4457, "mean_token_accuracy": 0.8564291596412659, "num_tokens": 35616173.0, "step": 940 }, { "epoch": 0.11970487215367001, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.6245427131652832, "learning_rate": 3.9847392963119965e-06, "loss": 0.5438, "mean_token_accuracy": 0.8302146196365356, "num_tokens": 35654801.0, "step": 941 }, { "epoch": 0.11983208243226053, "ewc_loss": 0.0147705078125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.2993812561035156e-05, "grad_norm": 1.6186542510986328, "learning_rate": 3.988978380669775e-06, "loss": 0.4897, "mean_token_accuracy": 0.8479042053222656, "num_tokens": 35692219.0, "step": 942 }, { "epoch": 0.11995929271085104, "ewc_loss": 0.01483154296875, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3053417205810547e-05, "grad_norm": 1.6246893405914307, "learning_rate": 3.993217465027554e-06, "loss": 0.5514, "mean_token_accuracy": 0.8294855356216431, "num_tokens": 35732886.0, "step": 943 }, { "epoch": 0.12008650298944154, "ewc_loss": 0.01483154296875, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3053417205810547e-05, "grad_norm": 1.6325310468673706, "learning_rate": 3.997456549385333e-06, "loss": 0.4391, "mean_token_accuracy": 0.8604484796524048, "num_tokens": 35766796.0, "step": 944 }, { "epoch": 0.12021371326803205, "ewc_loss": 0.014892578125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3113021850585938e-05, "grad_norm": 1.6299068927764893, "learning_rate": 4.001695633743112e-06, "loss": 0.5067, "mean_token_accuracy": 0.8404204249382019, "num_tokens": 35803065.0, "step": 945 }, { "epoch": 0.12034092354662257, "ewc_loss": 0.014892578125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3113021850585938e-05, "grad_norm": 1.6435792446136475, "learning_rate": 4.005934718100891e-06, "loss": 0.4959, "mean_token_accuracy": 0.8426705598831177, "num_tokens": 35836490.0, "step": 946 }, { "epoch": 0.12046813382521308, "ewc_loss": 0.014892578125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3113021850585938e-05, "grad_norm": 1.6040679216384888, "learning_rate": 4.01017380245867e-06, "loss": 0.4543, "mean_token_accuracy": 0.8573326468467712, "num_tokens": 35877239.0, "step": 947 }, { "epoch": 0.12059534410380358, "ewc_loss": 0.014892578125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3113021850585938e-05, "grad_norm": 1.6384646892547607, "learning_rate": 4.014412886816448e-06, "loss": 0.5617, "mean_token_accuracy": 0.8252594470977783, "num_tokens": 35917779.0, "step": 948 }, { "epoch": 0.1207225543823941, "ewc_loss": 0.0150146484375, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3232231140136719e-05, "grad_norm": 1.6091585159301758, "learning_rate": 4.018651971174227e-06, "loss": 0.4884, "mean_token_accuracy": 0.8455044627189636, "num_tokens": 35955083.0, "step": 949 }, { "epoch": 0.12084976466098461, "ewc_loss": 0.01495361328125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3172626495361328e-05, "grad_norm": 1.6393054723739624, "learning_rate": 4.022891055532006e-06, "loss": 0.5183, "mean_token_accuracy": 0.8361811637878418, "num_tokens": 35999581.0, "step": 950 }, { "epoch": 0.12097697493957511, "ewc_loss": 0.01507568359375, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.329183578491211e-05, "grad_norm": 1.6217833757400513, "learning_rate": 4.0271301398897845e-06, "loss": 0.5602, "mean_token_accuracy": 0.8292936682701111, "num_tokens": 36040891.0, "step": 951 }, { "epoch": 0.12110418521816563, "ewc_loss": 0.01495361328125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3172626495361328e-05, "grad_norm": 1.6557447910308838, "learning_rate": 4.0313692242475625e-06, "loss": 0.4846, "mean_token_accuracy": 0.843153715133667, "num_tokens": 36074958.0, "step": 952 }, { "epoch": 0.12123139549675614, "ewc_loss": 0.0150146484375, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3232231140136719e-05, "grad_norm": 1.6126960515975952, "learning_rate": 4.0356083086053414e-06, "loss": 0.5156, "mean_token_accuracy": 0.8352594971656799, "num_tokens": 36115211.0, "step": 953 }, { "epoch": 0.12135860577534664, "ewc_loss": 0.014892578125, "ewc_loss_diag": 1.7642974853515625e-05, "ewc_loss_parallel": 1.3113021850585938e-05, "grad_norm": 1.636744737625122, "learning_rate": 4.03984739296312e-06, "loss": 0.5159, "mean_token_accuracy": 0.8398962616920471, "num_tokens": 36156876.0, "step": 954 }, { "epoch": 0.12148581605393716, "ewc_loss": 0.0150146484375, "ewc_loss_diag": 1.7762184143066406e-05, "ewc_loss_parallel": 1.3232231140136719e-05, "grad_norm": 1.6114165782928467, "learning_rate": 4.044086477320898e-06, "loss": 0.4941, "mean_token_accuracy": 0.8414020538330078, "num_tokens": 36197277.0, "step": 955 }, { "epoch": 0.12161302633252767, "ewc_loss": 0.014892578125, "ewc_loss_diag": 1.7762184143066406e-05, "ewc_loss_parallel": 1.3113021850585938e-05, "grad_norm": 1.6654231548309326, "learning_rate": 4.048325561678677e-06, "loss": 0.5265, "mean_token_accuracy": 0.834458589553833, "num_tokens": 36232839.0, "step": 956 }, { "epoch": 0.12174023661111817, "ewc_loss": 0.0150146484375, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.3232231140136719e-05, "grad_norm": 1.6237070560455322, "learning_rate": 4.052564646036456e-06, "loss": 0.5172, "mean_token_accuracy": 0.8363804817199707, "num_tokens": 36273531.0, "step": 957 }, { "epoch": 0.12186744688970869, "ewc_loss": 0.01495361328125, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.3172626495361328e-05, "grad_norm": 1.6834148168563843, "learning_rate": 4.056803730394235e-06, "loss": 0.5301, "mean_token_accuracy": 0.8314868807792664, "num_tokens": 36309325.0, "step": 958 }, { "epoch": 0.1219946571682992, "ewc_loss": 0.0150146484375, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.3232231140136719e-05, "grad_norm": 1.626957893371582, "learning_rate": 4.061042814752014e-06, "loss": 0.5073, "mean_token_accuracy": 0.8413708209991455, "num_tokens": 36344457.0, "step": 959 }, { "epoch": 0.12212186744688971, "ewc_loss": 0.0150146484375, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.3232231140136719e-05, "grad_norm": 1.6889758110046387, "learning_rate": 4.065281899109793e-06, "loss": 0.5112, "mean_token_accuracy": 0.8363319635391235, "num_tokens": 36376731.0, "step": 960 }, { "epoch": 0.12224907772548022, "ewc_loss": 0.01513671875, "ewc_loss_diag": 1.7881393432617188e-05, "ewc_loss_parallel": 1.33514404296875e-05, "grad_norm": 1.6120373010635376, "learning_rate": 4.069520983467572e-06, "loss": 0.4642, "mean_token_accuracy": 0.8537977933883667, "num_tokens": 36415529.0, "step": 961 }, { "epoch": 0.12237628800407073, "ewc_loss": 0.0152587890625, "ewc_loss_diag": 1.800060272216797e-05, "ewc_loss_parallel": 1.341104507446289e-05, "grad_norm": 1.6671490669250488, "learning_rate": 4.0737600678253505e-06, "loss": 0.4922, "mean_token_accuracy": 0.8425018787384033, "num_tokens": 36449522.0, "step": 962 }, { "epoch": 0.12250349828266124, "ewc_loss": 0.01513671875, "ewc_loss_diag": 1.800060272216797e-05, "ewc_loss_parallel": 1.33514404296875e-05, "grad_norm": 1.6285415887832642, "learning_rate": 4.077999152183129e-06, "loss": 0.4429, "mean_token_accuracy": 0.8589215278625488, "num_tokens": 36483010.0, "step": 963 }, { "epoch": 0.12263070856125174, "ewc_loss": 0.01513671875, "ewc_loss_diag": 1.800060272216797e-05, "ewc_loss_parallel": 1.33514404296875e-05, "grad_norm": 1.6472853422164917, "learning_rate": 4.0822382365409075e-06, "loss": 0.5129, "mean_token_accuracy": 0.8407294154167175, "num_tokens": 36527993.0, "step": 964 }, { "epoch": 0.12275791883984226, "ewc_loss": 0.0152587890625, "ewc_loss_diag": 1.800060272216797e-05, "ewc_loss_parallel": 1.3470649719238281e-05, "grad_norm": 1.6296868324279785, "learning_rate": 4.086477320898686e-06, "loss": 0.4846, "mean_token_accuracy": 0.8451645970344543, "num_tokens": 36568479.0, "step": 965 }, { "epoch": 0.12288512911843277, "ewc_loss": 0.01519775390625, "ewc_loss_diag": 1.811981201171875e-05, "ewc_loss_parallel": 1.33514404296875e-05, "grad_norm": 1.6634117364883423, "learning_rate": 4.090716405256465e-06, "loss": 0.5234, "mean_token_accuracy": 0.8360323905944824, "num_tokens": 36607196.0, "step": 966 }, { "epoch": 0.12301233939702327, "ewc_loss": 0.015380859375, "ewc_loss_diag": 1.811981201171875e-05, "ewc_loss_parallel": 1.3530254364013672e-05, "grad_norm": 1.6572448015213013, "learning_rate": 4.094955489614243e-06, "loss": 0.4898, "mean_token_accuracy": 0.8434737920761108, "num_tokens": 36640845.0, "step": 967 }, { "epoch": 0.12313954967561379, "ewc_loss": 0.01513671875, "ewc_loss_diag": 1.811981201171875e-05, "ewc_loss_parallel": 1.329183578491211e-05, "grad_norm": 1.6418750286102295, "learning_rate": 4.099194573972022e-06, "loss": 0.4887, "mean_token_accuracy": 0.8460650444030762, "num_tokens": 36682182.0, "step": 968 }, { "epoch": 0.1232667599542043, "ewc_loss": 0.01531982421875, "ewc_loss_diag": 1.811981201171875e-05, "ewc_loss_parallel": 1.3470649719238281e-05, "grad_norm": 1.644265055656433, "learning_rate": 4.103433658329801e-06, "loss": 0.532, "mean_token_accuracy": 0.8330702185630798, "num_tokens": 36723629.0, "step": 969 }, { "epoch": 0.1233939702327948, "ewc_loss": 0.0152587890625, "ewc_loss_diag": 1.823902130126953e-05, "ewc_loss_parallel": 1.341104507446289e-05, "grad_norm": 1.6417902708053589, "learning_rate": 4.10767274268758e-06, "loss": 0.5154, "mean_token_accuracy": 0.8368839025497437, "num_tokens": 36763560.0, "step": 970 }, { "epoch": 0.12352118051138532, "ewc_loss": 0.015380859375, "ewc_loss_diag": 1.823902130126953e-05, "ewc_loss_parallel": 1.3530254364013672e-05, "grad_norm": 1.6654326915740967, "learning_rate": 4.111911827045359e-06, "loss": 0.5807, "mean_token_accuracy": 0.8149648904800415, "num_tokens": 36802267.0, "step": 971 }, { "epoch": 0.12364839078997583, "ewc_loss": 0.0152587890625, "ewc_loss_diag": 1.823902130126953e-05, "ewc_loss_parallel": 1.341104507446289e-05, "grad_norm": 1.643585205078125, "learning_rate": 4.116150911403137e-06, "loss": 0.4699, "mean_token_accuracy": 0.853858232498169, "num_tokens": 36840047.0, "step": 972 }, { "epoch": 0.12377560106856635, "ewc_loss": 0.01531982421875, "ewc_loss_diag": 1.823902130126953e-05, "ewc_loss_parallel": 1.3470649719238281e-05, "grad_norm": 1.6682062149047852, "learning_rate": 4.120389995760916e-06, "loss": 0.533, "mean_token_accuracy": 0.8344979286193848, "num_tokens": 36876788.0, "step": 973 }, { "epoch": 0.12390281134715685, "ewc_loss": 0.0152587890625, "ewc_loss_diag": 1.823902130126953e-05, "ewc_loss_parallel": 1.341104507446289e-05, "grad_norm": 1.6653732061386108, "learning_rate": 4.124629080118695e-06, "loss": 0.4816, "mean_token_accuracy": 0.8449991345405579, "num_tokens": 36909954.0, "step": 974 }, { "epoch": 0.12403002162574736, "ewc_loss": 0.01544189453125, "ewc_loss_diag": 1.823902130126953e-05, "ewc_loss_parallel": 1.3589859008789062e-05, "grad_norm": 1.6525412797927856, "learning_rate": 4.1288681644764735e-06, "loss": 0.5209, "mean_token_accuracy": 0.8345348834991455, "num_tokens": 36948371.0, "step": 975 }, { "epoch": 0.12415723190433788, "ewc_loss": 0.01531982421875, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3470649719238281e-05, "grad_norm": 1.6946966648101807, "learning_rate": 4.133107248834252e-06, "loss": 0.5305, "mean_token_accuracy": 0.8330633640289307, "num_tokens": 36983304.0, "step": 976 }, { "epoch": 0.12428444218292838, "ewc_loss": 0.01544189453125, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3589859008789062e-05, "grad_norm": 1.6406736373901367, "learning_rate": 4.137346333192031e-06, "loss": 0.451, "mean_token_accuracy": 0.8554241061210632, "num_tokens": 37019644.0, "step": 977 }, { "epoch": 0.12441165246151889, "ewc_loss": 0.01531982421875, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3470649719238281e-05, "grad_norm": 1.6623512506484985, "learning_rate": 4.141585417549809e-06, "loss": 0.5214, "mean_token_accuracy": 0.8337526321411133, "num_tokens": 37056906.0, "step": 978 }, { "epoch": 0.1245388627401094, "ewc_loss": 0.01544189453125, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3589859008789062e-05, "grad_norm": 1.6621381044387817, "learning_rate": 4.145824501907588e-06, "loss": 0.4774, "mean_token_accuracy": 0.8467569351196289, "num_tokens": 37090396.0, "step": 979 }, { "epoch": 0.1246660730186999, "ewc_loss": 0.01556396484375, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3709068298339844e-05, "grad_norm": 1.6597535610198975, "learning_rate": 4.150063586265367e-06, "loss": 0.5152, "mean_token_accuracy": 0.8383090496063232, "num_tokens": 37129246.0, "step": 980 }, { "epoch": 0.12479328329729042, "ewc_loss": 0.01544189453125, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3589859008789062e-05, "grad_norm": 1.6531827449798584, "learning_rate": 4.154302670623146e-06, "loss": 0.4626, "mean_token_accuracy": 0.8524682521820068, "num_tokens": 37166344.0, "step": 981 }, { "epoch": 0.12492049357588093, "ewc_loss": 0.01556396484375, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3709068298339844e-05, "grad_norm": 1.6515259742736816, "learning_rate": 4.158541754980924e-06, "loss": 0.4468, "mean_token_accuracy": 0.8594360947608948, "num_tokens": 37202827.0, "step": 982 }, { "epoch": 0.12504770385447145, "ewc_loss": 0.0155029296875, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3649463653564453e-05, "grad_norm": 1.670959234237671, "learning_rate": 4.162780839338703e-06, "loss": 0.5123, "mean_token_accuracy": 0.8346361517906189, "num_tokens": 37239606.0, "step": 983 }, { "epoch": 0.12517491413306195, "ewc_loss": 0.01556396484375, "ewc_loss_diag": 1.8477439880371094e-05, "ewc_loss_parallel": 1.3709068298339844e-05, "grad_norm": 1.6677957773208618, "learning_rate": 4.167019923696482e-06, "loss": 0.5076, "mean_token_accuracy": 0.836860179901123, "num_tokens": 37275154.0, "step": 984 }, { "epoch": 0.12530212441165245, "ewc_loss": 0.015625, "ewc_loss_diag": 1.8596649169921875e-05, "ewc_loss_parallel": 1.3709068298339844e-05, "grad_norm": 1.657885193824768, "learning_rate": 4.171259008054261e-06, "loss": 0.4633, "mean_token_accuracy": 0.8546632528305054, "num_tokens": 37311849.0, "step": 985 }, { "epoch": 0.12542933469024298, "ewc_loss": 0.015625, "ewc_loss_diag": 1.8596649169921875e-05, "ewc_loss_parallel": 1.3709068298339844e-05, "grad_norm": 1.6626564264297485, "learning_rate": 4.1754980924120395e-06, "loss": 0.5164, "mean_token_accuracy": 0.8390350341796875, "num_tokens": 37353174.0, "step": 986 }, { "epoch": 0.12555654496883348, "ewc_loss": 0.015625, "ewc_loss_diag": 1.8596649169921875e-05, "ewc_loss_parallel": 1.3768672943115234e-05, "grad_norm": 1.6647350788116455, "learning_rate": 4.1797371767698184e-06, "loss": 0.4719, "mean_token_accuracy": 0.85194993019104, "num_tokens": 37392081.0, "step": 987 }, { "epoch": 0.12568375524742398, "ewc_loss": 0.015625, "ewc_loss_diag": 1.8596649169921875e-05, "ewc_loss_parallel": 1.3768672943115234e-05, "grad_norm": 1.6582039594650269, "learning_rate": 4.183976261127597e-06, "loss": 0.4697, "mean_token_accuracy": 0.8512144088745117, "num_tokens": 37430639.0, "step": 988 }, { "epoch": 0.1258109655260145, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.8596649169921875e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.6878271102905273, "learning_rate": 4.188215345485375e-06, "loss": 0.5555, "mean_token_accuracy": 0.8229565620422363, "num_tokens": 37467126.0, "step": 989 }, { "epoch": 0.125938175804605, "ewc_loss": 0.015625, "ewc_loss_diag": 1.8715858459472656e-05, "ewc_loss_parallel": 1.3709068298339844e-05, "grad_norm": 1.6743884086608887, "learning_rate": 4.192454429843154e-06, "loss": 0.4869, "mean_token_accuracy": 0.847218930721283, "num_tokens": 37505657.0, "step": 990 }, { "epoch": 0.12606538608319554, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.8715858459472656e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.6579785346984863, "learning_rate": 4.196693514200933e-06, "loss": 0.4922, "mean_token_accuracy": 0.8455968499183655, "num_tokens": 37549262.0, "step": 991 }, { "epoch": 0.12619259636178604, "ewc_loss": 0.015625, "ewc_loss_diag": 1.8715858459472656e-05, "ewc_loss_parallel": 1.3768672943115234e-05, "grad_norm": 1.6591535806655884, "learning_rate": 4.200932598558712e-06, "loss": 0.5058, "mean_token_accuracy": 0.8386380672454834, "num_tokens": 37589967.0, "step": 992 }, { "epoch": 0.12631980664037654, "ewc_loss": 0.015625, "ewc_loss_diag": 1.8715858459472656e-05, "ewc_loss_parallel": 1.3709068298339844e-05, "grad_norm": 1.6572426557540894, "learning_rate": 4.20517168291649e-06, "loss": 0.5084, "mean_token_accuracy": 0.8421539664268494, "num_tokens": 37635686.0, "step": 993 }, { "epoch": 0.12644701691896706, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.8835067749023438e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.6818639039993286, "learning_rate": 4.209410767274269e-06, "loss": 0.5489, "mean_token_accuracy": 0.832669734954834, "num_tokens": 37673753.0, "step": 994 }, { "epoch": 0.12657422719755757, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.8835067749023438e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.6844968795776367, "learning_rate": 4.213649851632048e-06, "loss": 0.4583, "mean_token_accuracy": 0.8555031418800354, "num_tokens": 37706345.0, "step": 995 }, { "epoch": 0.12670143747614807, "ewc_loss": 0.015625, "ewc_loss_diag": 1.8835067749023438e-05, "ewc_loss_parallel": 1.3768672943115234e-05, "grad_norm": 1.6582657098770142, "learning_rate": 4.217888935989827e-06, "loss": 0.529, "mean_token_accuracy": 0.8315626978874207, "num_tokens": 37750356.0, "step": 996 }, { "epoch": 0.1268286477547386, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.8835067749023438e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.686865210533142, "learning_rate": 4.222128020347605e-06, "loss": 0.5222, "mean_token_accuracy": 0.8359296321868896, "num_tokens": 37791466.0, "step": 997 }, { "epoch": 0.1269558580333291, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.9073486328125e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.6612874269485474, "learning_rate": 4.226367104705384e-06, "loss": 0.528, "mean_token_accuracy": 0.8325000405311584, "num_tokens": 37837257.0, "step": 998 }, { "epoch": 0.1270830683119196, "ewc_loss": 0.015869140625, "ewc_loss_diag": 1.9073486328125e-05, "ewc_loss_parallel": 1.3887882232666016e-05, "grad_norm": 1.6762009859085083, "learning_rate": 4.2306061890631625e-06, "loss": 0.5163, "mean_token_accuracy": 0.8382238149642944, "num_tokens": 37880223.0, "step": 999 }, { "epoch": 0.12721027859051012, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.9073486328125e-05, "ewc_loss_parallel": 1.3768672943115234e-05, "grad_norm": 1.7302049398422241, "learning_rate": 4.234845273420941e-06, "loss": 0.5957, "mean_token_accuracy": 0.8204203844070435, "num_tokens": 37917662.0, "step": 1000 }, { "epoch": 0.12733748886910062, "ewc_loss": 0.015869140625, "ewc_loss_diag": 1.9073486328125e-05, "ewc_loss_parallel": 1.3887882232666016e-05, "grad_norm": 1.6550266742706299, "learning_rate": 4.23908435777872e-06, "loss": 0.5209, "mean_token_accuracy": 0.8336340188980103, "num_tokens": 37961985.0, "step": 1001 }, { "epoch": 0.12746469914769112, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.9073486328125e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.6978713274002075, "learning_rate": 4.243323442136499e-06, "loss": 0.5097, "mean_token_accuracy": 0.8432714939117432, "num_tokens": 37998960.0, "step": 1002 }, { "epoch": 0.12759190942628165, "ewc_loss": 0.015625, "ewc_loss_diag": 1.9073486328125e-05, "ewc_loss_parallel": 1.3709068298339844e-05, "grad_norm": 1.7405376434326172, "learning_rate": 4.247562526494278e-06, "loss": 0.4824, "mean_token_accuracy": 0.8492013812065125, "num_tokens": 38040920.0, "step": 1003 }, { "epoch": 0.12771911970487215, "ewc_loss": 0.015869140625, "ewc_loss_diag": 1.919269561767578e-05, "ewc_loss_parallel": 1.3947486877441406e-05, "grad_norm": 1.7184807062149048, "learning_rate": 4.251801610852057e-06, "loss": 0.5294, "mean_token_accuracy": 0.8339322209358215, "num_tokens": 38071814.0, "step": 1004 }, { "epoch": 0.12784632998346265, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.919269561767578e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.6648051738739014, "learning_rate": 4.256040695209835e-06, "loss": 0.5469, "mean_token_accuracy": 0.8308582305908203, "num_tokens": 38112784.0, "step": 1005 }, { "epoch": 0.12797354026205318, "ewc_loss": 0.0157470703125, "ewc_loss_diag": 1.919269561767578e-05, "ewc_loss_parallel": 1.3828277587890625e-05, "grad_norm": 1.6905666589736938, "learning_rate": 4.260279779567614e-06, "loss": 0.5628, "mean_token_accuracy": 0.8276876211166382, "num_tokens": 38152140.0, "step": 1006 }, { "epoch": 0.12810075054064368, "ewc_loss": 0.015869140625, "ewc_loss_diag": 1.919269561767578e-05, "ewc_loss_parallel": 1.3947486877441406e-05, "grad_norm": 1.6679359674453735, "learning_rate": 4.264518863925393e-06, "loss": 0.4668, "mean_token_accuracy": 0.852512001991272, "num_tokens": 38188293.0, "step": 1007 }, { "epoch": 0.12822796081923418, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.919269561767578e-05, "ewc_loss_parallel": 1.4066696166992188e-05, "grad_norm": 1.7047501802444458, "learning_rate": 4.268757948283171e-06, "loss": 0.5205, "mean_token_accuracy": 0.836297333240509, "num_tokens": 38229136.0, "step": 1008 }, { "epoch": 0.1283551710978247, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.9311904907226562e-05, "ewc_loss_parallel": 1.4066696166992188e-05, "grad_norm": 1.665967345237732, "learning_rate": 4.27299703264095e-06, "loss": 0.538, "mean_token_accuracy": 0.8318650722503662, "num_tokens": 38273330.0, "step": 1009 }, { "epoch": 0.1284823813764152, "ewc_loss": 0.015869140625, "ewc_loss_diag": 1.9311904907226562e-05, "ewc_loss_parallel": 1.4007091522216797e-05, "grad_norm": 1.6997628211975098, "learning_rate": 4.2772361169987285e-06, "loss": 0.5011, "mean_token_accuracy": 0.8426017761230469, "num_tokens": 38315066.0, "step": 1010 }, { "epoch": 0.1286095916550057, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.9311904907226562e-05, "ewc_loss_parallel": 1.4126300811767578e-05, "grad_norm": 1.6805907487869263, "learning_rate": 4.281475201356507e-06, "loss": 0.507, "mean_token_accuracy": 0.8408564925193787, "num_tokens": 38353430.0, "step": 1011 }, { "epoch": 0.12873680193359624, "ewc_loss": 0.0162353515625, "ewc_loss_diag": 1.9311904907226562e-05, "ewc_loss_parallel": 1.430511474609375e-05, "grad_norm": 1.7145520448684692, "learning_rate": 4.2857142857142855e-06, "loss": 0.5391, "mean_token_accuracy": 0.8322800397872925, "num_tokens": 38391919.0, "step": 1012 }, { "epoch": 0.12886401221218674, "ewc_loss": 0.01611328125, "ewc_loss_diag": 1.9550323486328125e-05, "ewc_loss_parallel": 1.4126300811767578e-05, "grad_norm": 1.6715209484100342, "learning_rate": 4.289953370072064e-06, "loss": 0.4813, "mean_token_accuracy": 0.8495872616767883, "num_tokens": 38430104.0, "step": 1013 }, { "epoch": 0.12899122249077727, "ewc_loss": 0.01611328125, "ewc_loss_diag": 1.9550323486328125e-05, "ewc_loss_parallel": 1.4185905456542969e-05, "grad_norm": 1.714624047279358, "learning_rate": 4.294192454429843e-06, "loss": 0.5281, "mean_token_accuracy": 0.8359402418136597, "num_tokens": 38468775.0, "step": 1014 }, { "epoch": 0.12911843276936777, "ewc_loss": 0.0159912109375, "ewc_loss_diag": 1.9550323486328125e-05, "ewc_loss_parallel": 1.4066696166992188e-05, "grad_norm": 1.6540919542312622, "learning_rate": 4.298431538787622e-06, "loss": 0.4154, "mean_token_accuracy": 0.8698602318763733, "num_tokens": 38510330.0, "step": 1015 }, { "epoch": 0.12924564304795827, "ewc_loss": 0.01611328125, "ewc_loss_diag": 1.9550323486328125e-05, "ewc_loss_parallel": 1.424551010131836e-05, "grad_norm": 1.7519357204437256, "learning_rate": 4.302670623145401e-06, "loss": 0.5565, "mean_token_accuracy": 0.826653778553009, "num_tokens": 38543840.0, "step": 1016 }, { "epoch": 0.1293728533265488, "ewc_loss": 0.0162353515625, "ewc_loss_diag": 1.9550323486328125e-05, "ewc_loss_parallel": 1.430511474609375e-05, "grad_norm": 1.6892809867858887, "learning_rate": 4.30690970750318e-06, "loss": 0.5096, "mean_token_accuracy": 0.8422849178314209, "num_tokens": 38581670.0, "step": 1017 }, { "epoch": 0.1295000636051393, "ewc_loss": 0.0162353515625, "ewc_loss_diag": 1.9550323486328125e-05, "ewc_loss_parallel": 1.430511474609375e-05, "grad_norm": 1.7582980394363403, "learning_rate": 4.311148791860959e-06, "loss": 0.497, "mean_token_accuracy": 0.8398500084877014, "num_tokens": 38612642.0, "step": 1018 }, { "epoch": 0.1296272738837298, "ewc_loss": 0.01611328125, "ewc_loss_diag": 1.9669532775878906e-05, "ewc_loss_parallel": 1.4185905456542969e-05, "grad_norm": 1.6837217807769775, "learning_rate": 4.315387876218738e-06, "loss": 0.5348, "mean_token_accuracy": 0.8370119333267212, "num_tokens": 38647579.0, "step": 1019 }, { "epoch": 0.12975448416232033, "ewc_loss": 0.0162353515625, "ewc_loss_diag": 1.9669532775878906e-05, "ewc_loss_parallel": 1.430511474609375e-05, "grad_norm": 1.7425990104675293, "learning_rate": 4.319626960576516e-06, "loss": 0.5348, "mean_token_accuracy": 0.8290468454360962, "num_tokens": 38683114.0, "step": 1020 }, { "epoch": 0.12988169444091083, "ewc_loss": 0.01611328125, "ewc_loss_diag": 1.9788742065429688e-05, "ewc_loss_parallel": 1.4185905456542969e-05, "grad_norm": 1.6756726503372192, "learning_rate": 4.3238660449342946e-06, "loss": 0.4782, "mean_token_accuracy": 0.8467249870300293, "num_tokens": 38722990.0, "step": 1021 }, { "epoch": 0.13000890471950133, "ewc_loss": 0.016357421875, "ewc_loss_diag": 1.9788742065429688e-05, "ewc_loss_parallel": 1.4424324035644531e-05, "grad_norm": 1.729636311531067, "learning_rate": 4.3281051292920735e-06, "loss": 0.5046, "mean_token_accuracy": 0.840968132019043, "num_tokens": 38765041.0, "step": 1022 }, { "epoch": 0.13013611499809186, "ewc_loss": 0.0162353515625, "ewc_loss_diag": 1.9788742065429688e-05, "ewc_loss_parallel": 1.430511474609375e-05, "grad_norm": 1.6771174669265747, "learning_rate": 4.332344213649852e-06, "loss": 0.4696, "mean_token_accuracy": 0.8498029708862305, "num_tokens": 38802241.0, "step": 1023 }, { "epoch": 0.13026332527668236, "ewc_loss": 0.0164794921875, "ewc_loss_diag": 1.990795135498047e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.744627833366394, "learning_rate": 4.33658329800763e-06, "loss": 0.4976, "mean_token_accuracy": 0.8446699380874634, "num_tokens": 38839444.0, "step": 1024 }, { "epoch": 0.13039053555527286, "ewc_loss": 0.016357421875, "ewc_loss_diag": 2.014636993408203e-05, "ewc_loss_parallel": 1.430511474609375e-05, "grad_norm": 1.6722105741500854, "learning_rate": 4.340822382365409e-06, "loss": 0.4795, "mean_token_accuracy": 0.8479989767074585, "num_tokens": 38882102.0, "step": 1025 }, { "epoch": 0.13051774583386339, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.014636993408203e-05, "ewc_loss_parallel": 1.4603137969970703e-05, "grad_norm": 1.7217772006988525, "learning_rate": 4.345061466723188e-06, "loss": 0.4692, "mean_token_accuracy": 0.8538708686828613, "num_tokens": 38925855.0, "step": 1026 }, { "epoch": 0.13064495611245389, "ewc_loss": 0.016357421875, "ewc_loss_diag": 2.014636993408203e-05, "ewc_loss_parallel": 1.436471939086914e-05, "grad_norm": 1.7051799297332764, "learning_rate": 4.349300551080966e-06, "loss": 0.526, "mean_token_accuracy": 0.8357696533203125, "num_tokens": 38962293.0, "step": 1027 }, { "epoch": 0.1307721663910444, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.014636993408203e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.7302393913269043, "learning_rate": 4.353539635438745e-06, "loss": 0.4935, "mean_token_accuracy": 0.8425018191337585, "num_tokens": 38999000.0, "step": 1028 }, { "epoch": 0.13089937666963491, "ewc_loss": 0.016357421875, "ewc_loss_diag": 2.014636993408203e-05, "ewc_loss_parallel": 1.436471939086914e-05, "grad_norm": 1.7257508039474487, "learning_rate": 4.357778719796524e-06, "loss": 0.5549, "mean_token_accuracy": 0.8232777118682861, "num_tokens": 39032819.0, "step": 1029 }, { "epoch": 0.13102658694822542, "ewc_loss": 0.0164794921875, "ewc_loss_diag": 2.0384788513183594e-05, "ewc_loss_parallel": 1.4483928680419922e-05, "grad_norm": 1.717147946357727, "learning_rate": 4.362017804154303e-06, "loss": 0.4765, "mean_token_accuracy": 0.85148024559021, "num_tokens": 39070247.0, "step": 1030 }, { "epoch": 0.13115379722681592, "ewc_loss": 0.0164794921875, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4424324035644531e-05, "grad_norm": 1.711970567703247, "learning_rate": 4.366256888512082e-06, "loss": 0.4742, "mean_token_accuracy": 0.8519484996795654, "num_tokens": 39105310.0, "step": 1031 }, { "epoch": 0.13128100750540644, "ewc_loss": 0.0164794921875, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4424324035644531e-05, "grad_norm": 1.7125356197357178, "learning_rate": 4.370495972869861e-06, "loss": 0.5047, "mean_token_accuracy": 0.8406953811645508, "num_tokens": 39144218.0, "step": 1032 }, { "epoch": 0.13140821778399694, "ewc_loss": 0.016357421875, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.436471939086914e-05, "grad_norm": 1.7330750226974487, "learning_rate": 4.3747350572276395e-06, "loss": 0.4951, "mean_token_accuracy": 0.8424201607704163, "num_tokens": 39177248.0, "step": 1033 }, { "epoch": 0.13153542806258745, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.7069387435913086, "learning_rate": 4.378974141585418e-06, "loss": 0.5049, "mean_token_accuracy": 0.8411657810211182, "num_tokens": 39214230.0, "step": 1034 }, { "epoch": 0.13166263834117797, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4603137969970703e-05, "grad_norm": 1.71219003200531, "learning_rate": 4.383213225943196e-06, "loss": 0.497, "mean_token_accuracy": 0.8414089679718018, "num_tokens": 39252189.0, "step": 1035 }, { "epoch": 0.13178984861976847, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.7099692821502686, "learning_rate": 4.387452310300975e-06, "loss": 0.4464, "mean_token_accuracy": 0.8592270016670227, "num_tokens": 39288058.0, "step": 1036 }, { "epoch": 0.13191705889835897, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4603137969970703e-05, "grad_norm": 1.7210153341293335, "learning_rate": 4.391691394658754e-06, "loss": 0.4551, "mean_token_accuracy": 0.8535659313201904, "num_tokens": 39322157.0, "step": 1037 }, { "epoch": 0.1320442691769495, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.7151657342910767, "learning_rate": 4.395930479016533e-06, "loss": 0.4714, "mean_token_accuracy": 0.8489189743995667, "num_tokens": 39357678.0, "step": 1038 }, { "epoch": 0.13217147945554, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4722347259521484e-05, "grad_norm": 1.7085767984390259, "learning_rate": 4.400169563374311e-06, "loss": 0.4922, "mean_token_accuracy": 0.8447028398513794, "num_tokens": 39398909.0, "step": 1039 }, { "epoch": 0.13229868973413053, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4603137969970703e-05, "grad_norm": 1.7373442649841309, "learning_rate": 4.40440864773209e-06, "loss": 0.5683, "mean_token_accuracy": 0.8235609531402588, "num_tokens": 39437934.0, "step": 1040 }, { "epoch": 0.13242590001272103, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4543533325195312e-05, "grad_norm": 1.7188150882720947, "learning_rate": 4.408647732089869e-06, "loss": 0.4889, "mean_token_accuracy": 0.8424385190010071, "num_tokens": 39471772.0, "step": 1041 }, { "epoch": 0.13255311029131153, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4662742614746094e-05, "grad_norm": 1.715482473373413, "learning_rate": 4.412886816447648e-06, "loss": 0.5038, "mean_token_accuracy": 0.8458299040794373, "num_tokens": 39512398.0, "step": 1042 }, { "epoch": 0.13268032056990206, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 2.0503997802734375e-05, "ewc_loss_parallel": 1.4662742614746094e-05, "grad_norm": 1.719929814338684, "learning_rate": 4.417125900805427e-06, "loss": 0.4974, "mean_token_accuracy": 0.8442096710205078, "num_tokens": 39554735.0, "step": 1043 }, { "epoch": 0.13280753084849256, "ewc_loss": 0.0166015625, "ewc_loss_diag": 2.0623207092285156e-05, "ewc_loss_parallel": 1.4603137969970703e-05, "grad_norm": 1.7238669395446777, "learning_rate": 4.4213649851632055e-06, "loss": 0.5469, "mean_token_accuracy": 0.8324205279350281, "num_tokens": 39595397.0, "step": 1044 }, { "epoch": 0.13293474112708306, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 2.0623207092285156e-05, "ewc_loss_parallel": 1.4662742614746094e-05, "grad_norm": 1.8235459327697754, "learning_rate": 4.4256040695209836e-06, "loss": 0.5306, "mean_token_accuracy": 0.834571123123169, "num_tokens": 39632474.0, "step": 1045 }, { "epoch": 0.1330619514056736, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 2.0623207092285156e-05, "ewc_loss_parallel": 1.4722347259521484e-05, "grad_norm": 1.7170275449752808, "learning_rate": 4.4298431538787625e-06, "loss": 0.5169, "mean_token_accuracy": 0.832883894443512, "num_tokens": 39666408.0, "step": 1046 }, { "epoch": 0.1331891616842641, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 2.0623207092285156e-05, "ewc_loss_parallel": 1.4662742614746094e-05, "grad_norm": 1.7374153137207031, "learning_rate": 4.434082238236541e-06, "loss": 0.5526, "mean_token_accuracy": 0.825762152671814, "num_tokens": 39709033.0, "step": 1047 }, { "epoch": 0.1333163719628546, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 2.0623207092285156e-05, "ewc_loss_parallel": 1.4722347259521484e-05, "grad_norm": 1.702697992324829, "learning_rate": 4.43832132259432e-06, "loss": 0.5244, "mean_token_accuracy": 0.8372011780738831, "num_tokens": 39749732.0, "step": 1048 }, { "epoch": 0.13344358224144512, "ewc_loss": 0.0167236328125, "ewc_loss_diag": 2.0623207092285156e-05, "ewc_loss_parallel": 1.4722347259521484e-05, "grad_norm": 1.7555382251739502, "learning_rate": 4.442560406952099e-06, "loss": 0.5329, "mean_token_accuracy": 0.830142617225647, "num_tokens": 39788125.0, "step": 1049 }, { "epoch": 0.13357079252003562, "ewc_loss": 0.016845703125, "ewc_loss_diag": 2.0742416381835938e-05, "ewc_loss_parallel": 1.4841556549072266e-05, "grad_norm": 1.6905015707015991, "learning_rate": 4.446799491309877e-06, "loss": 0.4751, "mean_token_accuracy": 0.8518537282943726, "num_tokens": 39830125.0, "step": 1050 }, { "epoch": 0.13369800279862612, "ewc_loss": 0.0169677734375, "ewc_loss_diag": 2.0742416381835938e-05, "ewc_loss_parallel": 1.4901161193847656e-05, "grad_norm": 1.7538068294525146, "learning_rate": 4.451038575667656e-06, "loss": 0.5314, "mean_token_accuracy": 0.8310684561729431, "num_tokens": 39868236.0, "step": 1051 }, { "epoch": 0.13382521307721665, "ewc_loss": 0.0169677734375, "ewc_loss_diag": 2.086162567138672e-05, "ewc_loss_parallel": 1.4841556549072266e-05, "grad_norm": 1.693219542503357, "learning_rate": 4.455277660025435e-06, "loss": 0.4987, "mean_token_accuracy": 0.841537356376648, "num_tokens": 39913254.0, "step": 1052 }, { "epoch": 0.13395242335580715, "ewc_loss": 0.0169677734375, "ewc_loss_diag": 2.086162567138672e-05, "ewc_loss_parallel": 1.4901161193847656e-05, "grad_norm": 1.7780139446258545, "learning_rate": 4.459516744383214e-06, "loss": 0.509, "mean_token_accuracy": 0.837923526763916, "num_tokens": 39948242.0, "step": 1053 }, { "epoch": 0.13407963363439765, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.110004425048828e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7201197147369385, "learning_rate": 4.463755828740992e-06, "loss": 0.5243, "mean_token_accuracy": 0.8365963697433472, "num_tokens": 39989196.0, "step": 1054 }, { "epoch": 0.13420684391298818, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.110004425048828e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7659472227096558, "learning_rate": 4.467994913098771e-06, "loss": 0.5237, "mean_token_accuracy": 0.8433293700218201, "num_tokens": 40028662.0, "step": 1055 }, { "epoch": 0.13433405419157868, "ewc_loss": 0.0169677734375, "ewc_loss_diag": 2.110004425048828e-05, "ewc_loss_parallel": 1.4901161193847656e-05, "grad_norm": 1.7258250713348389, "learning_rate": 4.47223399745655e-06, "loss": 0.5241, "mean_token_accuracy": 0.8357131481170654, "num_tokens": 40066178.0, "step": 1056 }, { "epoch": 0.13446126447016918, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.110004425048828e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7439097166061401, "learning_rate": 4.4764730818143285e-06, "loss": 0.5154, "mean_token_accuracy": 0.8370441794395447, "num_tokens": 40106908.0, "step": 1057 }, { "epoch": 0.1345884747487597, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.110004425048828e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7227355241775513, "learning_rate": 4.480712166172107e-06, "loss": 0.4783, "mean_token_accuracy": 0.8497384190559387, "num_tokens": 40146872.0, "step": 1058 }, { "epoch": 0.1347156850273502, "ewc_loss": 0.016845703125, "ewc_loss_diag": 2.110004425048828e-05, "ewc_loss_parallel": 1.4781951904296875e-05, "grad_norm": 1.7505052089691162, "learning_rate": 4.484951250529886e-06, "loss": 0.508, "mean_token_accuracy": 0.8395065665245056, "num_tokens": 40186717.0, "step": 1059 }, { "epoch": 0.1348428953059407, "ewc_loss": 0.0169677734375, "ewc_loss_diag": 2.110004425048828e-05, "ewc_loss_parallel": 1.4901161193847656e-05, "grad_norm": 1.7344682216644287, "learning_rate": 4.489190334887665e-06, "loss": 0.5471, "mean_token_accuracy": 0.8328132629394531, "num_tokens": 40227119.0, "step": 1060 }, { "epoch": 0.13497010558453124, "ewc_loss": 0.0169677734375, "ewc_loss_diag": 2.110004425048828e-05, "ewc_loss_parallel": 1.4841556549072266e-05, "grad_norm": 1.76557195186615, "learning_rate": 4.493429419245444e-06, "loss": 0.5245, "mean_token_accuracy": 0.8364803791046143, "num_tokens": 40262302.0, "step": 1061 }, { "epoch": 0.13509731586312174, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7329598665237427, "learning_rate": 4.497668503603222e-06, "loss": 0.4713, "mean_token_accuracy": 0.8521323204040527, "num_tokens": 40296783.0, "step": 1062 }, { "epoch": 0.13522452614171224, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7720569372177124, "learning_rate": 4.501907587961001e-06, "loss": 0.5339, "mean_token_accuracy": 0.8295125365257263, "num_tokens": 40331323.0, "step": 1063 }, { "epoch": 0.13535173642030277, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7288678884506226, "learning_rate": 4.50614667231878e-06, "loss": 0.5054, "mean_token_accuracy": 0.840994119644165, "num_tokens": 40369965.0, "step": 1064 }, { "epoch": 0.13547894669889327, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7441436052322388, "learning_rate": 4.510385756676558e-06, "loss": 0.5549, "mean_token_accuracy": 0.8298575282096863, "num_tokens": 40410820.0, "step": 1065 }, { "epoch": 0.1356061569774838, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.5020370483398438e-05, "grad_norm": 1.7369384765625, "learning_rate": 4.514624841034337e-06, "loss": 0.4597, "mean_token_accuracy": 0.8549644351005554, "num_tokens": 40446942.0, "step": 1066 }, { "epoch": 0.1357333672560743, "ewc_loss": 0.0172119140625, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.5079975128173828e-05, "grad_norm": 1.74681556224823, "learning_rate": 4.518863925392116e-06, "loss": 0.4764, "mean_token_accuracy": 0.8494673371315002, "num_tokens": 40484684.0, "step": 1067 }, { "epoch": 0.1358605775346648, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7358149290084839, "learning_rate": 4.5231030097498945e-06, "loss": 0.4557, "mean_token_accuracy": 0.8530499339103699, "num_tokens": 40521164.0, "step": 1068 }, { "epoch": 0.13598778781325532, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.7275445461273193, "learning_rate": 4.5273420941076726e-06, "loss": 0.4922, "mean_token_accuracy": 0.8428850769996643, "num_tokens": 40566192.0, "step": 1069 }, { "epoch": 0.13611499809184582, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.5020370483398438e-05, "grad_norm": 1.7492640018463135, "learning_rate": 4.5315811784654514e-06, "loss": 0.5268, "mean_token_accuracy": 0.8349916338920593, "num_tokens": 40605534.0, "step": 1070 }, { "epoch": 0.13624220837043632, "ewc_loss": 0.0172119140625, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.5079975128173828e-05, "grad_norm": 1.7391754388809204, "learning_rate": 4.53582026282323e-06, "loss": 0.48, "mean_token_accuracy": 0.8463140726089478, "num_tokens": 40645478.0, "step": 1071 }, { "epoch": 0.13636941864902685, "ewc_loss": 0.017333984375, "ewc_loss_diag": 2.1219253540039062e-05, "ewc_loss_parallel": 1.519918441772461e-05, "grad_norm": 1.7722737789154053, "learning_rate": 4.540059347181009e-06, "loss": 0.5651, "mean_token_accuracy": 0.8222339153289795, "num_tokens": 40684591.0, "step": 1072 }, { "epoch": 0.13649662892761735, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1338462829589844e-05, "ewc_loss_parallel": 1.4960765838623047e-05, "grad_norm": 1.737848162651062, "learning_rate": 4.544298431538788e-06, "loss": 0.4992, "mean_token_accuracy": 0.8419824242591858, "num_tokens": 40724437.0, "step": 1073 }, { "epoch": 0.13662383920620785, "ewc_loss": 0.01708984375, "ewc_loss_diag": 2.1338462829589844e-05, "ewc_loss_parallel": 1.5020370483398438e-05, "grad_norm": 1.780148983001709, "learning_rate": 4.548537515896567e-06, "loss": 0.5441, "mean_token_accuracy": 0.8290265202522278, "num_tokens": 40758324.0, "step": 1074 }, { "epoch": 0.13675104948479838, "ewc_loss": 0.0172119140625, "ewc_loss_diag": 2.1338462829589844e-05, "ewc_loss_parallel": 1.5079975128173828e-05, "grad_norm": 1.7542765140533447, "learning_rate": 4.552776600254346e-06, "loss": 0.5276, "mean_token_accuracy": 0.8348628282546997, "num_tokens": 40794182.0, "step": 1075 }, { "epoch": 0.13687825976338888, "ewc_loss": 0.017333984375, "ewc_loss_diag": 2.1338462829589844e-05, "ewc_loss_parallel": 1.5139579772949219e-05, "grad_norm": 1.7535659074783325, "learning_rate": 4.557015684612125e-06, "loss": 0.4776, "mean_token_accuracy": 0.8504137396812439, "num_tokens": 40832293.0, "step": 1076 }, { "epoch": 0.13700547004197938, "ewc_loss": 0.017333984375, "ewc_loss_diag": 2.1457672119140625e-05, "ewc_loss_parallel": 1.5139579772949219e-05, "grad_norm": 1.7370812892913818, "learning_rate": 4.561254768969903e-06, "loss": 0.4729, "mean_token_accuracy": 0.8490417003631592, "num_tokens": 40871078.0, "step": 1077 }, { "epoch": 0.1371326803205699, "ewc_loss": 0.0172119140625, "ewc_loss_diag": 2.1457672119140625e-05, "ewc_loss_parallel": 1.5079975128173828e-05, "grad_norm": 1.7673771381378174, "learning_rate": 4.565493853327682e-06, "loss": 0.4838, "mean_token_accuracy": 0.8490264415740967, "num_tokens": 40909061.0, "step": 1078 }, { "epoch": 0.1372598905991604, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 2.1457672119140625e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.7607613801956177, "learning_rate": 4.5697329376854606e-06, "loss": 0.522, "mean_token_accuracy": 0.841090738773346, "num_tokens": 40945623.0, "step": 1079 }, { "epoch": 0.1373871008777509, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 2.1576881408691406e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.7683663368225098, "learning_rate": 4.573972022043239e-06, "loss": 0.5866, "mean_token_accuracy": 0.81612628698349, "num_tokens": 40985009.0, "step": 1080 }, { "epoch": 0.13751431115634144, "ewc_loss": 0.017333984375, "ewc_loss_diag": 2.1576881408691406e-05, "ewc_loss_parallel": 1.519918441772461e-05, "grad_norm": 1.7456676959991455, "learning_rate": 4.5782111064010175e-06, "loss": 0.536, "mean_token_accuracy": 0.833492636680603, "num_tokens": 41027377.0, "step": 1081 }, { "epoch": 0.13764152143493194, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 2.1696090698242188e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.764891266822815, "learning_rate": 4.582450190758796e-06, "loss": 0.5397, "mean_token_accuracy": 0.8306216597557068, "num_tokens": 41068816.0, "step": 1082 }, { "epoch": 0.13776873171352244, "ewc_loss": 0.017578125, "ewc_loss_diag": 2.1696090698242188e-05, "ewc_loss_parallel": 1.537799835205078e-05, "grad_norm": 1.7467786073684692, "learning_rate": 4.586689275116575e-06, "loss": 0.4603, "mean_token_accuracy": 0.8547773957252502, "num_tokens": 41106603.0, "step": 1083 }, { "epoch": 0.13789594199211297, "ewc_loss": 0.017578125, "ewc_loss_diag": 2.1696090698242188e-05, "ewc_loss_parallel": 1.537799835205078e-05, "grad_norm": 1.7749298810958862, "learning_rate": 4.590928359474353e-06, "loss": 0.5485, "mean_token_accuracy": 0.8278417587280273, "num_tokens": 41142862.0, "step": 1084 }, { "epoch": 0.13802315227070347, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 2.1696090698242188e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.7396621704101562, "learning_rate": 4.595167443832132e-06, "loss": 0.4989, "mean_token_accuracy": 0.8444094657897949, "num_tokens": 41182761.0, "step": 1085 }, { "epoch": 0.13815036254929397, "ewc_loss": 0.017578125, "ewc_loss_diag": 2.1696090698242188e-05, "ewc_loss_parallel": 1.537799835205078e-05, "grad_norm": 1.7648005485534668, "learning_rate": 4.599406528189911e-06, "loss": 0.45, "mean_token_accuracy": 0.8584628105163574, "num_tokens": 41221318.0, "step": 1086 }, { "epoch": 0.1382775728278845, "ewc_loss": 0.0177001953125, "ewc_loss_diag": 2.181529998779297e-05, "ewc_loss_parallel": 1.5497207641601562e-05, "grad_norm": 1.7534637451171875, "learning_rate": 4.60364561254769e-06, "loss": 0.4895, "mean_token_accuracy": 0.8417688608169556, "num_tokens": 41257748.0, "step": 1087 }, { "epoch": 0.138404783106475, "ewc_loss": 0.017578125, "ewc_loss_diag": 2.181529998779297e-05, "ewc_loss_parallel": 1.537799835205078e-05, "grad_norm": 1.7748159170150757, "learning_rate": 4.607884696905469e-06, "loss": 0.4932, "mean_token_accuracy": 0.8437632918357849, "num_tokens": 41299287.0, "step": 1088 }, { "epoch": 0.1385319933850655, "ewc_loss": 0.0174560546875, "ewc_loss_diag": 2.181529998779297e-05, "ewc_loss_parallel": 1.52587890625e-05, "grad_norm": 1.7513946294784546, "learning_rate": 4.612123781263248e-06, "loss": 0.5343, "mean_token_accuracy": 0.8311601877212524, "num_tokens": 41344600.0, "step": 1089 }, { "epoch": 0.13865920366365603, "ewc_loss": 0.0177001953125, "ewc_loss_diag": 2.181529998779297e-05, "ewc_loss_parallel": 1.5497207641601562e-05, "grad_norm": 1.7727222442626953, "learning_rate": 4.616362865621027e-06, "loss": 0.4272, "mean_token_accuracy": 0.8621019721031189, "num_tokens": 41380681.0, "step": 1090 }, { "epoch": 0.13878641394224653, "ewc_loss": 0.0177001953125, "ewc_loss_diag": 2.181529998779297e-05, "ewc_loss_parallel": 1.5497207641601562e-05, "grad_norm": 1.7615635395050049, "learning_rate": 4.6206019499788055e-06, "loss": 0.5226, "mean_token_accuracy": 0.8390158414840698, "num_tokens": 41418751.0, "step": 1091 }, { "epoch": 0.13891362422083706, "ewc_loss": 0.017822265625, "ewc_loss_diag": 2.181529998779297e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.7929939031600952, "learning_rate": 4.6248410343365835e-06, "loss": 0.5695, "mean_token_accuracy": 0.8192853927612305, "num_tokens": 41455617.0, "step": 1092 }, { "epoch": 0.13904083449942756, "ewc_loss": 0.0177001953125, "ewc_loss_diag": 2.193450927734375e-05, "ewc_loss_parallel": 1.5497207641601562e-05, "grad_norm": 1.7467013597488403, "learning_rate": 4.629080118694362e-06, "loss": 0.4802, "mean_token_accuracy": 0.849450945854187, "num_tokens": 41494480.0, "step": 1093 }, { "epoch": 0.13916804477801806, "ewc_loss": 0.017578125, "ewc_loss_diag": 2.193450927734375e-05, "ewc_loss_parallel": 1.537799835205078e-05, "grad_norm": 1.7924635410308838, "learning_rate": 4.633319203052141e-06, "loss": 0.4955, "mean_token_accuracy": 0.845756471157074, "num_tokens": 41529294.0, "step": 1094 }, { "epoch": 0.13929525505660859, "ewc_loss": 0.017578125, "ewc_loss_diag": 2.2172927856445312e-05, "ewc_loss_parallel": 1.537799835205078e-05, "grad_norm": 1.7447930574417114, "learning_rate": 4.63755828740992e-06, "loss": 0.4714, "mean_token_accuracy": 0.8505572080612183, "num_tokens": 41570787.0, "step": 1095 }, { "epoch": 0.1394224653351991, "ewc_loss": 0.0177001953125, "ewc_loss_diag": 2.2172927856445312e-05, "ewc_loss_parallel": 1.5497207641601562e-05, "grad_norm": 1.797861099243164, "learning_rate": 4.641797371767698e-06, "loss": 0.4853, "mean_token_accuracy": 0.8506263494491577, "num_tokens": 41606596.0, "step": 1096 }, { "epoch": 0.1395496756137896, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2172927856445312e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7489248514175415, "learning_rate": 4.646036456125477e-06, "loss": 0.5069, "mean_token_accuracy": 0.8375741243362427, "num_tokens": 41645317.0, "step": 1097 }, { "epoch": 0.13967688589238011, "ewc_loss": 0.017822265625, "ewc_loss_diag": 2.2172927856445312e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.8092098236083984, "learning_rate": 4.650275540483256e-06, "loss": 0.473, "mean_token_accuracy": 0.8469456434249878, "num_tokens": 41680424.0, "step": 1098 }, { "epoch": 0.13980409617097062, "ewc_loss": 0.017822265625, "ewc_loss_diag": 2.2172927856445312e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.7753980159759521, "learning_rate": 4.654514624841035e-06, "loss": 0.5001, "mean_token_accuracy": 0.8430453538894653, "num_tokens": 41712492.0, "step": 1099 }, { "epoch": 0.13993130644956112, "ewc_loss": 0.017822265625, "ewc_loss_diag": 2.2292137145996094e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.8029241561889648, "learning_rate": 4.658753709198813e-06, "loss": 0.5023, "mean_token_accuracy": 0.8434878587722778, "num_tokens": 41747666.0, "step": 1100 }, { "epoch": 0.14005851672815164, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2292137145996094e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7617276906967163, "learning_rate": 4.662992793556592e-06, "loss": 0.4821, "mean_token_accuracy": 0.8483748435974121, "num_tokens": 41784638.0, "step": 1101 }, { "epoch": 0.14018572700674214, "ewc_loss": 0.017822265625, "ewc_loss_diag": 2.2292137145996094e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.7920255661010742, "learning_rate": 4.667231877914371e-06, "loss": 0.5686, "mean_token_accuracy": 0.8228157162666321, "num_tokens": 41826547.0, "step": 1102 }, { "epoch": 0.14031293728533265, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2292137145996094e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7836648225784302, "learning_rate": 4.6714709622721496e-06, "loss": 0.5158, "mean_token_accuracy": 0.8371590375900269, "num_tokens": 41859526.0, "step": 1103 }, { "epoch": 0.14044014756392317, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2292137145996094e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7834618091583252, "learning_rate": 4.6757100466299284e-06, "loss": 0.5174, "mean_token_accuracy": 0.8382497429847717, "num_tokens": 41901177.0, "step": 1104 }, { "epoch": 0.14056735784251367, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2292137145996094e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.769469976425171, "learning_rate": 4.679949130987707e-06, "loss": 0.5286, "mean_token_accuracy": 0.8322110176086426, "num_tokens": 41941579.0, "step": 1105 }, { "epoch": 0.14069456812110417, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2411346435546875e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.8030084371566772, "learning_rate": 4.684188215345486e-06, "loss": 0.4587, "mean_token_accuracy": 0.8548136949539185, "num_tokens": 41975315.0, "step": 1106 }, { "epoch": 0.1408217783996947, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2411346435546875e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.767500877380371, "learning_rate": 4.688427299703264e-06, "loss": 0.4493, "mean_token_accuracy": 0.8525882959365845, "num_tokens": 42011950.0, "step": 1107 }, { "epoch": 0.1409489886782852, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2411346435546875e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7889981269836426, "learning_rate": 4.692666384061043e-06, "loss": 0.5114, "mean_token_accuracy": 0.8394050002098083, "num_tokens": 42051764.0, "step": 1108 }, { "epoch": 0.1410761989568757, "ewc_loss": 0.017822265625, "ewc_loss_diag": 2.2411346435546875e-05, "ewc_loss_parallel": 1.5616416931152344e-05, "grad_norm": 1.7666709423065186, "learning_rate": 4.696905468418822e-06, "loss": 0.5258, "mean_token_accuracy": 0.8322681188583374, "num_tokens": 42094218.0, "step": 1109 }, { "epoch": 0.14120340923546623, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2411346435546875e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.813161015510559, "learning_rate": 4.701144552776601e-06, "loss": 0.5012, "mean_token_accuracy": 0.8425562381744385, "num_tokens": 42129120.0, "step": 1110 }, { "epoch": 0.14133061951405673, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2411346435546875e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7737467288970947, "learning_rate": 4.705383637134379e-06, "loss": 0.5174, "mean_token_accuracy": 0.8372156620025635, "num_tokens": 42167437.0, "step": 1111 }, { "epoch": 0.14145782979264723, "ewc_loss": 0.0179443359375, "ewc_loss_diag": 2.2411346435546875e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.807271957397461, "learning_rate": 4.709622721492158e-06, "loss": 0.4897, "mean_token_accuracy": 0.8477798104286194, "num_tokens": 42203143.0, "step": 1112 }, { "epoch": 0.14158504007123776, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7817755937576294, "learning_rate": 4.713861805849937e-06, "loss": 0.4864, "mean_token_accuracy": 0.8464757204055786, "num_tokens": 42237024.0, "step": 1113 }, { "epoch": 0.14171225034982826, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.8251479864120483, "learning_rate": 4.718100890207716e-06, "loss": 0.5627, "mean_token_accuracy": 0.824500322341919, "num_tokens": 42272847.0, "step": 1114 }, { "epoch": 0.1418394606284188, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7947250604629517, "learning_rate": 4.7223399745654945e-06, "loss": 0.5167, "mean_token_accuracy": 0.8365163803100586, "num_tokens": 42309335.0, "step": 1115 }, { "epoch": 0.1419666709070093, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7993720769882202, "learning_rate": 4.726579058923273e-06, "loss": 0.5124, "mean_token_accuracy": 0.8418656587600708, "num_tokens": 42346349.0, "step": 1116 }, { "epoch": 0.1420938811855998, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.795867681503296, "learning_rate": 4.730818143281051e-06, "loss": 0.54, "mean_token_accuracy": 0.8325494527816772, "num_tokens": 42389310.0, "step": 1117 }, { "epoch": 0.14222109146419032, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.797289252281189, "learning_rate": 4.73505722763883e-06, "loss": 0.4796, "mean_token_accuracy": 0.8486104011535645, "num_tokens": 42427808.0, "step": 1118 }, { "epoch": 0.14234830174278082, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.7936551570892334, "learning_rate": 4.739296311996609e-06, "loss": 0.5034, "mean_token_accuracy": 0.8404754996299744, "num_tokens": 42465531.0, "step": 1119 }, { "epoch": 0.14247551202137132, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5735626220703125e-05, "grad_norm": 1.7821009159088135, "learning_rate": 4.743535396354388e-06, "loss": 0.4981, "mean_token_accuracy": 0.8467788100242615, "num_tokens": 42507613.0, "step": 1120 }, { "epoch": 0.14260272229996185, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.8172075748443604, "learning_rate": 4.747774480712167e-06, "loss": 0.522, "mean_token_accuracy": 0.8386576771736145, "num_tokens": 42542876.0, "step": 1121 }, { "epoch": 0.14272993257855235, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 1.805763840675354, "learning_rate": 4.752013565069945e-06, "loss": 0.5386, "mean_token_accuracy": 0.8315227031707764, "num_tokens": 42577587.0, "step": 1122 }, { "epoch": 0.14285714285714285, "ewc_loss": 0.01806640625, "ewc_loss_diag": 2.2649765014648438e-05, "ewc_loss_parallel": 1.5854835510253906e-05, "grad_norm": 2.02294659614563, "learning_rate": 4.756252649427724e-06, "loss": 0.5067, "mean_token_accuracy": 0.8405817747116089, "num_tokens": 42612733.0, "step": 1123 }, { "epoch": 0.14298435313573338, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.276897430419922e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.7672969102859497, "learning_rate": 4.760491733785503e-06, "loss": 0.452, "mean_token_accuracy": 0.8571346402168274, "num_tokens": 42646831.0, "step": 1124 }, { "epoch": 0.14311156341432388, "ewc_loss": 0.018310546875, "ewc_loss_diag": 2.276897430419922e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.8699734210968018, "learning_rate": 4.764730818143282e-06, "loss": 0.5014, "mean_token_accuracy": 0.8362252712249756, "num_tokens": 42681351.0, "step": 1125 }, { "epoch": 0.14323877369291438, "ewc_loss": 0.018310546875, "ewc_loss_diag": 2.276897430419922e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.7524625062942505, "learning_rate": 4.76896990250106e-06, "loss": 0.5523, "mean_token_accuracy": 0.8260117769241333, "num_tokens": 42723491.0, "step": 1126 }, { "epoch": 0.1433659839715049, "ewc_loss": 0.0185546875, "ewc_loss_diag": 2.276897430419922e-05, "ewc_loss_parallel": 1.621246337890625e-05, "grad_norm": 1.869284749031067, "learning_rate": 4.7732089868588386e-06, "loss": 0.5119, "mean_token_accuracy": 0.8380838632583618, "num_tokens": 42763349.0, "step": 1127 }, { "epoch": 0.1434931942500954, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.276897430419922e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.76705002784729, "learning_rate": 4.7774480712166174e-06, "loss": 0.5248, "mean_token_accuracy": 0.8347504138946533, "num_tokens": 42803977.0, "step": 1128 }, { "epoch": 0.1436204045286859, "ewc_loss": 0.0186767578125, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.8592650890350342, "learning_rate": 4.781687155574396e-06, "loss": 0.5278, "mean_token_accuracy": 0.833337128162384, "num_tokens": 42847313.0, "step": 1129 }, { "epoch": 0.14374761480727644, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.7672975063323975, "learning_rate": 4.785926239932175e-06, "loss": 0.4745, "mean_token_accuracy": 0.848570704460144, "num_tokens": 42887468.0, "step": 1130 }, { "epoch": 0.14387482508586694, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.8712373971939087, "learning_rate": 4.790165324289954e-06, "loss": 0.5036, "mean_token_accuracy": 0.8427897095680237, "num_tokens": 42925502.0, "step": 1131 }, { "epoch": 0.14400203536445744, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.797998070716858, "learning_rate": 4.794404408647733e-06, "loss": 0.5056, "mean_token_accuracy": 0.8433880805969238, "num_tokens": 42961930.0, "step": 1132 }, { "epoch": 0.14412924564304797, "ewc_loss": 0.0186767578125, "ewc_loss_diag": 2.288818359375e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.8406440019607544, "learning_rate": 4.798643493005512e-06, "loss": 0.4512, "mean_token_accuracy": 0.8570562601089478, "num_tokens": 43002121.0, "step": 1133 }, { "epoch": 0.14425645592163847, "ewc_loss": 0.018310546875, "ewc_loss_diag": 2.300739288330078e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.8126921653747559, "learning_rate": 4.80288257736329e-06, "loss": 0.5486, "mean_token_accuracy": 0.8298819065093994, "num_tokens": 43039343.0, "step": 1134 }, { "epoch": 0.14438366620022897, "ewc_loss": 0.018310546875, "ewc_loss_diag": 2.300739288330078e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.840822696685791, "learning_rate": 4.807121661721069e-06, "loss": 0.451, "mean_token_accuracy": 0.8592267036437988, "num_tokens": 43075035.0, "step": 1135 }, { "epoch": 0.1445108764788195, "ewc_loss": 0.018310546875, "ewc_loss_diag": 2.300739288330078e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.794526219367981, "learning_rate": 4.811360746078848e-06, "loss": 0.5049, "mean_token_accuracy": 0.8432796597480774, "num_tokens": 43114350.0, "step": 1136 }, { "epoch": 0.14463808675741, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.3126602172851562e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.8705662488937378, "learning_rate": 4.815599830436626e-06, "loss": 0.5033, "mean_token_accuracy": 0.8414038419723511, "num_tokens": 43146445.0, "step": 1137 }, { "epoch": 0.1447652970360005, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.3126602172851562e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.8095426559448242, "learning_rate": 4.819838914794405e-06, "loss": 0.4882, "mean_token_accuracy": 0.8465647101402283, "num_tokens": 43182216.0, "step": 1138 }, { "epoch": 0.14489250731459102, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.3126602172851562e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.8483784198760986, "learning_rate": 4.8240779991521835e-06, "loss": 0.4964, "mean_token_accuracy": 0.8419538736343384, "num_tokens": 43216014.0, "step": 1139 }, { "epoch": 0.14501971759318152, "ewc_loss": 0.018310546875, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.7956478595733643, "learning_rate": 4.828317083509962e-06, "loss": 0.5034, "mean_token_accuracy": 0.8414785861968994, "num_tokens": 43252909.0, "step": 1140 }, { "epoch": 0.14514692787177205, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.8502650260925293, "learning_rate": 4.83255616786774e-06, "loss": 0.4804, "mean_token_accuracy": 0.8518824577331543, "num_tokens": 43285653.0, "step": 1141 }, { "epoch": 0.14527413815036255, "ewc_loss": 0.018310546875, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 1.8076348304748535, "learning_rate": 4.836795252225519e-06, "loss": 0.4768, "mean_token_accuracy": 0.8522994518280029, "num_tokens": 43322087.0, "step": 1142 }, { "epoch": 0.14540134842895305, "ewc_loss": 0.0185546875, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 1.621246337890625e-05, "grad_norm": 1.8486849069595337, "learning_rate": 4.841034336583298e-06, "loss": 0.4879, "mean_token_accuracy": 0.8487658500671387, "num_tokens": 43360526.0, "step": 1143 }, { "epoch": 0.14552855870754358, "ewc_loss": 0.018310546875, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 1.5974044799804688e-05, "grad_norm": 3.113248825073242, "learning_rate": 4.845273420941077e-06, "loss": 0.4671, "mean_token_accuracy": 0.8548877239227295, "num_tokens": 43400267.0, "step": 1144 }, { "epoch": 0.14565576898613408, "ewc_loss": 0.0234375, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 2.110004425048828e-05, "grad_norm": 2.0646026134490967, "learning_rate": 4.849512505298856e-06, "loss": 0.5118, "mean_token_accuracy": 0.8368158936500549, "num_tokens": 43430411.0, "step": 1145 }, { "epoch": 0.14578297926472458, "ewc_loss": 0.0184326171875, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 1.609325408935547e-05, "grad_norm": 1.7792729139328003, "learning_rate": 4.853751589656635e-06, "loss": 0.4923, "mean_token_accuracy": 0.8472906351089478, "num_tokens": 43472199.0, "step": 1146 }, { "epoch": 0.1459101895433151, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.3365020751953125e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.863063097000122, "learning_rate": 4.857990674014414e-06, "loss": 0.5158, "mean_token_accuracy": 0.8383859395980835, "num_tokens": 43515896.0, "step": 1147 }, { "epoch": 0.1460373998219056, "ewc_loss": 0.0194091796875, "ewc_loss_diag": 2.3484230041503906e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.8546714782714844, "learning_rate": 4.862229758372193e-06, "loss": 0.5382, "mean_token_accuracy": 0.8282394409179688, "num_tokens": 43556088.0, "step": 1148 }, { "epoch": 0.1461646101004961, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.3484230041503906e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.8592678308486938, "learning_rate": 4.866468842729971e-06, "loss": 0.5425, "mean_token_accuracy": 0.8282506465911865, "num_tokens": 43596623.0, "step": 1149 }, { "epoch": 0.14629182037908664, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.3484230041503906e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.8547985553741455, "learning_rate": 4.8707079270877495e-06, "loss": 0.5155, "mean_token_accuracy": 0.8390441536903381, "num_tokens": 43634757.0, "step": 1150 }, { "epoch": 0.14641903065767714, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.8539408445358276, "learning_rate": 4.874947011445528e-06, "loss": 0.5206, "mean_token_accuracy": 0.8388400077819824, "num_tokens": 43671979.0, "step": 1151 }, { "epoch": 0.14654624093626764, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.824684500694275, "learning_rate": 4.879186095803307e-06, "loss": 0.4446, "mean_token_accuracy": 0.8602399230003357, "num_tokens": 43711898.0, "step": 1152 }, { "epoch": 0.14667345121485817, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8608360290527344, "learning_rate": 4.883425180161085e-06, "loss": 0.476, "mean_token_accuracy": 0.8512613773345947, "num_tokens": 43748394.0, "step": 1153 }, { "epoch": 0.14680066149344867, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.848061442375183, "learning_rate": 4.887664264518864e-06, "loss": 0.5271, "mean_token_accuracy": 0.8417179584503174, "num_tokens": 43784730.0, "step": 1154 }, { "epoch": 0.14692787177203917, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.854364037513733, "learning_rate": 4.891903348876643e-06, "loss": 0.5161, "mean_token_accuracy": 0.8421686887741089, "num_tokens": 43823993.0, "step": 1155 }, { "epoch": 0.1470550820506297, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8508274555206299, "learning_rate": 4.896142433234421e-06, "loss": 0.5509, "mean_token_accuracy": 0.8335475325584412, "num_tokens": 43861673.0, "step": 1156 }, { "epoch": 0.1471822923292202, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.3603439331054688e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8529516458511353, "learning_rate": 4.9003815175922e-06, "loss": 0.5464, "mean_token_accuracy": 0.8342392444610596, "num_tokens": 43900123.0, "step": 1157 }, { "epoch": 0.1473095026078107, "ewc_loss": 0.0186767578125, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.8312656879425049, "learning_rate": 4.904620601949979e-06, "loss": 0.5011, "mean_token_accuracy": 0.8419293165206909, "num_tokens": 43939371.0, "step": 1158 }, { "epoch": 0.14743671288640123, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.8654086589813232, "learning_rate": 4.908859686307758e-06, "loss": 0.5354, "mean_token_accuracy": 0.8345514535903931, "num_tokens": 43972960.0, "step": 1159 }, { "epoch": 0.14756392316499173, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8430378437042236, "learning_rate": 4.913098770665537e-06, "loss": 0.5386, "mean_token_accuracy": 0.835716962814331, "num_tokens": 44009919.0, "step": 1160 }, { "epoch": 0.14769113344358223, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8416880369186401, "learning_rate": 4.9173378550233155e-06, "loss": 0.4857, "mean_token_accuracy": 0.8441251516342163, "num_tokens": 44050491.0, "step": 1161 }, { "epoch": 0.14781834372217276, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8330016136169434, "learning_rate": 4.9215769393810944e-06, "loss": 0.5042, "mean_token_accuracy": 0.846686840057373, "num_tokens": 44087999.0, "step": 1162 }, { "epoch": 0.14794555400076326, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.8519741296768188, "learning_rate": 4.925816023738873e-06, "loss": 0.5541, "mean_token_accuracy": 0.828027606010437, "num_tokens": 44126014.0, "step": 1163 }, { "epoch": 0.14807276427935376, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.8614866733551025, "learning_rate": 4.930055108096651e-06, "loss": 0.4424, "mean_token_accuracy": 0.8602995276451111, "num_tokens": 44154713.0, "step": 1164 }, { "epoch": 0.1481999745579443, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.870448112487793, "learning_rate": 4.93429419245443e-06, "loss": 0.5687, "mean_token_accuracy": 0.824022114276886, "num_tokens": 44189624.0, "step": 1165 }, { "epoch": 0.1483271848365348, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.843546986579895, "learning_rate": 4.938533276812209e-06, "loss": 0.5099, "mean_token_accuracy": 0.8408939838409424, "num_tokens": 44225010.0, "step": 1166 }, { "epoch": 0.14845439511512531, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.372264862060547e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8300745487213135, "learning_rate": 4.942772361169988e-06, "loss": 0.4234, "mean_token_accuracy": 0.8617244362831116, "num_tokens": 44264987.0, "step": 1167 }, { "epoch": 0.14858160539371582, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.384185791015625e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.838464379310608, "learning_rate": 4.947011445527766e-06, "loss": 0.4997, "mean_token_accuracy": 0.8454224467277527, "num_tokens": 44306705.0, "step": 1168 }, { "epoch": 0.14870881567230632, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.396106719970703e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.8304734230041504, "learning_rate": 4.951250529885545e-06, "loss": 0.4554, "mean_token_accuracy": 0.856991171836853, "num_tokens": 44347227.0, "step": 1169 }, { "epoch": 0.14883602595089684, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.396106719970703e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.8739333152770996, "learning_rate": 4.955489614243324e-06, "loss": 0.5902, "mean_token_accuracy": 0.8129534721374512, "num_tokens": 44385273.0, "step": 1170 }, { "epoch": 0.14896323622948734, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.396106719970703e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.8273011445999146, "learning_rate": 4.959728698601103e-06, "loss": 0.509, "mean_token_accuracy": 0.8366454243659973, "num_tokens": 44426177.0, "step": 1171 }, { "epoch": 0.14909044650807785, "ewc_loss": 0.019287109375, "ewc_loss_diag": 2.396106719970703e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.8548985719680786, "learning_rate": 4.963967782958881e-06, "loss": 0.4854, "mean_token_accuracy": 0.8476084470748901, "num_tokens": 44468609.0, "step": 1172 }, { "epoch": 0.14921765678666837, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.396106719970703e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.833042860031128, "learning_rate": 4.96820686731666e-06, "loss": 0.486, "mean_token_accuracy": 0.845668375492096, "num_tokens": 44507212.0, "step": 1173 }, { "epoch": 0.14934486706525887, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.4080276489257812e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8850972652435303, "learning_rate": 4.9724459516744385e-06, "loss": 0.4693, "mean_token_accuracy": 0.85023033618927, "num_tokens": 44538551.0, "step": 1174 }, { "epoch": 0.14947207734384937, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.4199485778808594e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.8482201099395752, "learning_rate": 4.976685036032217e-06, "loss": 0.5606, "mean_token_accuracy": 0.8237705826759338, "num_tokens": 44577105.0, "step": 1175 }, { "epoch": 0.1495992876224399, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.4199485778808594e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8545392751693726, "learning_rate": 4.980924120389996e-06, "loss": 0.439, "mean_token_accuracy": 0.861636221408844, "num_tokens": 44615548.0, "step": 1176 }, { "epoch": 0.1497264979010304, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.4199485778808594e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.822832703590393, "learning_rate": 4.985163204747775e-06, "loss": 0.4462, "mean_token_accuracy": 0.85928875207901, "num_tokens": 44654447.0, "step": 1177 }, { "epoch": 0.1498537081796209, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.4199485778808594e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.866029977798462, "learning_rate": 4.989402289105554e-06, "loss": 0.5004, "mean_token_accuracy": 0.8458572030067444, "num_tokens": 44697995.0, "step": 1178 }, { "epoch": 0.14998091845821143, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.4199485778808594e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.872258186340332, "learning_rate": 4.993641373463332e-06, "loss": 0.5606, "mean_token_accuracy": 0.8242085576057434, "num_tokens": 44734376.0, "step": 1179 }, { "epoch": 0.15010812873680193, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.4199485778808594e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.859845757484436, "learning_rate": 4.997880457821111e-06, "loss": 0.5092, "mean_token_accuracy": 0.8385822772979736, "num_tokens": 44772745.0, "step": 1180 }, { "epoch": 0.15023533901539243, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.4199485778808594e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.835492730140686, "learning_rate": 5.002119542178889e-06, "loss": 0.4975, "mean_token_accuracy": 0.8451451659202576, "num_tokens": 44816530.0, "step": 1181 }, { "epoch": 0.15036254929398296, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.872054934501648, "learning_rate": 5.006358626536668e-06, "loss": 0.458, "mean_token_accuracy": 0.853905200958252, "num_tokens": 44853428.0, "step": 1182 }, { "epoch": 0.15048975957257346, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.8423470258712769, "learning_rate": 5.010597710894447e-06, "loss": 0.4794, "mean_token_accuracy": 0.848106324672699, "num_tokens": 44891118.0, "step": 1183 }, { "epoch": 0.15061696985116396, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.8754067420959473, "learning_rate": 5.014836795252226e-06, "loss": 0.4698, "mean_token_accuracy": 0.8511818647384644, "num_tokens": 44929143.0, "step": 1184 }, { "epoch": 0.1507441801297545, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.8510754108428955, "learning_rate": 5.0190758796100045e-06, "loss": 0.4758, "mean_token_accuracy": 0.852276086807251, "num_tokens": 44966361.0, "step": 1185 }, { "epoch": 0.150871390408345, "ewc_loss": 0.0194091796875, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.8739001750946045, "learning_rate": 5.0233149639677834e-06, "loss": 0.512, "mean_token_accuracy": 0.8373115062713623, "num_tokens": 45007993.0, "step": 1186 }, { "epoch": 0.1509986006869355, "ewc_loss": 0.018798828125, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 1.633167266845703e-05, "grad_norm": 1.8788827657699585, "learning_rate": 5.027554048325562e-06, "loss": 0.5489, "mean_token_accuracy": 0.8268972635269165, "num_tokens": 45040337.0, "step": 1187 }, { "epoch": 0.15112581096552602, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.4318695068359375e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.874761700630188, "learning_rate": 5.031793132683341e-06, "loss": 0.5583, "mean_token_accuracy": 0.8250395655632019, "num_tokens": 45080730.0, "step": 1188 }, { "epoch": 0.15125302124411652, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.4437904357910156e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.867081642150879, "learning_rate": 5.036032217041119e-06, "loss": 0.517, "mean_token_accuracy": 0.8354249000549316, "num_tokens": 45116748.0, "step": 1189 }, { "epoch": 0.15138023152270705, "ewc_loss": 0.0189208984375, "ewc_loss_diag": 2.4437904357910156e-05, "ewc_loss_parallel": 1.6450881958007812e-05, "grad_norm": 1.8564764261245728, "learning_rate": 5.040271301398898e-06, "loss": 0.507, "mean_token_accuracy": 0.8366499543190002, "num_tokens": 45158596.0, "step": 1190 }, { "epoch": 0.15150744180129755, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.4437904357910156e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8492100238800049, "learning_rate": 5.044510385756677e-06, "loss": 0.4818, "mean_token_accuracy": 0.845903754234314, "num_tokens": 45202433.0, "step": 1191 }, { "epoch": 0.15163465207988805, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.4437904357910156e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.8740017414093018, "learning_rate": 5.048749470114456e-06, "loss": 0.4909, "mean_token_accuracy": 0.8468124866485596, "num_tokens": 45239202.0, "step": 1192 }, { "epoch": 0.15176186235847858, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.4437904357910156e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8988053798675537, "learning_rate": 5.052988554472235e-06, "loss": 0.5073, "mean_token_accuracy": 0.8396793007850647, "num_tokens": 45270150.0, "step": 1193 }, { "epoch": 0.15188907263706908, "ewc_loss": 0.019287109375, "ewc_loss_diag": 2.4437904357910156e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.8665568828582764, "learning_rate": 5.057227638830014e-06, "loss": 0.5062, "mean_token_accuracy": 0.8425495624542236, "num_tokens": 45306394.0, "step": 1194 }, { "epoch": 0.15201628291565958, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.4437904357910156e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.882283329963684, "learning_rate": 5.0614667231877925e-06, "loss": 0.5231, "mean_token_accuracy": 0.8369388580322266, "num_tokens": 45344945.0, "step": 1195 }, { "epoch": 0.1521434931942501, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.4437904357910156e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.8801440000534058, "learning_rate": 5.06570580754557e-06, "loss": 0.4759, "mean_token_accuracy": 0.8521997332572937, "num_tokens": 45378179.0, "step": 1196 }, { "epoch": 0.1522707034728406, "ewc_loss": 0.019287109375, "ewc_loss_diag": 2.4557113647460938e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.855193853378296, "learning_rate": 5.069944891903349e-06, "loss": 0.4915, "mean_token_accuracy": 0.8436764478683472, "num_tokens": 45421194.0, "step": 1197 }, { "epoch": 0.1523979137514311, "ewc_loss": 0.019287109375, "ewc_loss_diag": 2.4557113647460938e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.8793981075286865, "learning_rate": 5.0741839762611275e-06, "loss": 0.5014, "mean_token_accuracy": 0.8407264947891235, "num_tokens": 45458226.0, "step": 1198 }, { "epoch": 0.15252512403002164, "ewc_loss": 0.019287109375, "ewc_loss_diag": 2.4557113647460938e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.8602792024612427, "learning_rate": 5.078423060618906e-06, "loss": 0.5028, "mean_token_accuracy": 0.8426219820976257, "num_tokens": 45500007.0, "step": 1199 }, { "epoch": 0.15265233430861214, "ewc_loss": 0.0194091796875, "ewc_loss_diag": 2.4557113647460938e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.8937406539916992, "learning_rate": 5.082662144976685e-06, "loss": 0.4961, "mean_token_accuracy": 0.8399636149406433, "num_tokens": 45536605.0, "step": 1200 }, { "epoch": 0.15277954458720264, "ewc_loss": 0.019287109375, "ewc_loss_diag": 2.4557113647460938e-05, "ewc_loss_parallel": 1.6808509826660156e-05, "grad_norm": 1.8696147203445435, "learning_rate": 5.086901229334464e-06, "loss": 0.4918, "mean_token_accuracy": 0.8488495349884033, "num_tokens": 45572137.0, "step": 1201 }, { "epoch": 0.15290675486579317, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 2.467632293701172e-05, "ewc_loss_parallel": 1.71661376953125e-05, "grad_norm": 1.8850516080856323, "learning_rate": 5.091140313692243e-06, "loss": 0.4975, "mean_token_accuracy": 0.842055082321167, "num_tokens": 45608659.0, "step": 1202 }, { "epoch": 0.15303396514438367, "ewc_loss": 0.01904296875, "ewc_loss_diag": 2.47955322265625e-05, "ewc_loss_parallel": 1.6570091247558594e-05, "grad_norm": 1.8646247386932373, "learning_rate": 5.095379398050022e-06, "loss": 0.4828, "mean_token_accuracy": 0.8466936945915222, "num_tokens": 45647002.0, "step": 1203 }, { "epoch": 0.15316117542297417, "ewc_loss": 0.01953125, "ewc_loss_diag": 2.47955322265625e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.912889003753662, "learning_rate": 5.0996184824078e-06, "loss": 0.5101, "mean_token_accuracy": 0.8367040157318115, "num_tokens": 45680194.0, "step": 1204 }, { "epoch": 0.1532883857015647, "ewc_loss": 0.0194091796875, "ewc_loss_diag": 2.47955322265625e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.855789065361023, "learning_rate": 5.103857566765579e-06, "loss": 0.4825, "mean_token_accuracy": 0.8472450971603394, "num_tokens": 45719629.0, "step": 1205 }, { "epoch": 0.1534155959801552, "ewc_loss": 0.019775390625, "ewc_loss_diag": 2.47955322265625e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.900123953819275, "learning_rate": 5.108096651123358e-06, "loss": 0.4521, "mean_token_accuracy": 0.857195258140564, "num_tokens": 45754630.0, "step": 1206 }, { "epoch": 0.1535428062587457, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.8633977174758911, "learning_rate": 5.112335735481137e-06, "loss": 0.4882, "mean_token_accuracy": 0.8441480994224548, "num_tokens": 45791904.0, "step": 1207 }, { "epoch": 0.15367001653733622, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.8912168741226196, "learning_rate": 5.1165748198389155e-06, "loss": 0.4079, "mean_token_accuracy": 0.8730998039245605, "num_tokens": 45830553.0, "step": 1208 }, { "epoch": 0.15379722681592672, "ewc_loss": 0.0191650390625, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.6689300537109375e-05, "grad_norm": 1.8756287097930908, "learning_rate": 5.120813904196694e-06, "loss": 0.4819, "mean_token_accuracy": 0.8504694700241089, "num_tokens": 45865957.0, "step": 1209 }, { "epoch": 0.15392443709451722, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.9244673252105713, "learning_rate": 5.125052988554473e-06, "loss": 0.5427, "mean_token_accuracy": 0.8315256834030151, "num_tokens": 45898993.0, "step": 1210 }, { "epoch": 0.15405164737310775, "ewc_loss": 0.0194091796875, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.8745579719543457, "learning_rate": 5.1292920729122505e-06, "loss": 0.4727, "mean_token_accuracy": 0.851355791091919, "num_tokens": 45932843.0, "step": 1211 }, { "epoch": 0.15417885765169825, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.8962379693984985, "learning_rate": 5.133531157270029e-06, "loss": 0.4717, "mean_token_accuracy": 0.8545844554901123, "num_tokens": 45973247.0, "step": 1212 }, { "epoch": 0.15430606793028875, "ewc_loss": 0.0194091796875, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.864598274230957, "learning_rate": 5.137770241627808e-06, "loss": 0.4321, "mean_token_accuracy": 0.8629172444343567, "num_tokens": 46008641.0, "step": 1213 }, { "epoch": 0.15443327820887928, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.903693437576294, "learning_rate": 5.142009325985587e-06, "loss": 0.4643, "mean_token_accuracy": 0.8571783304214478, "num_tokens": 46046789.0, "step": 1214 }, { "epoch": 0.15456048848746978, "ewc_loss": 0.0194091796875, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.6927719116210938e-05, "grad_norm": 1.8657423257827759, "learning_rate": 5.146248410343366e-06, "loss": 0.5026, "mean_token_accuracy": 0.843863308429718, "num_tokens": 46086904.0, "step": 1215 }, { "epoch": 0.1546876987660603, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.491474151611328e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.8950315713882446, "learning_rate": 5.150487494701145e-06, "loss": 0.4537, "mean_token_accuracy": 0.8594020009040833, "num_tokens": 46129743.0, "step": 1216 }, { "epoch": 0.1548149090446508, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 2.5272369384765625e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.885166049003601, "learning_rate": 5.154726579058924e-06, "loss": 0.4967, "mean_token_accuracy": 0.8439109921455383, "num_tokens": 46164902.0, "step": 1217 }, { "epoch": 0.1549421193232413, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.5272369384765625e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.8844833374023438, "learning_rate": 5.158965663416703e-06, "loss": 0.4936, "mean_token_accuracy": 0.847240686416626, "num_tokens": 46209689.0, "step": 1218 }, { "epoch": 0.15506932960183184, "ewc_loss": 0.019775390625, "ewc_loss_diag": 2.5272369384765625e-05, "ewc_loss_parallel": 1.71661376953125e-05, "grad_norm": 1.8707280158996582, "learning_rate": 5.163204747774481e-06, "loss": 0.4438, "mean_token_accuracy": 0.8587988615036011, "num_tokens": 46251365.0, "step": 1219 }, { "epoch": 0.15519653988042234, "ewc_loss": 0.0201416015625, "ewc_loss_diag": 2.5272369384765625e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.898465871810913, "learning_rate": 5.1674438321322596e-06, "loss": 0.4963, "mean_token_accuracy": 0.8424840569496155, "num_tokens": 46290961.0, "step": 1220 }, { "epoch": 0.15532375015901284, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.894661545753479, "learning_rate": 5.1716829164900385e-06, "loss": 0.4737, "mean_token_accuracy": 0.8514467477798462, "num_tokens": 46326317.0, "step": 1221 }, { "epoch": 0.15545096043760337, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.9015711545944214, "learning_rate": 5.175922000847817e-06, "loss": 0.4439, "mean_token_accuracy": 0.8592554330825806, "num_tokens": 46361018.0, "step": 1222 }, { "epoch": 0.15557817071619387, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.890036940574646, "learning_rate": 5.180161085205596e-06, "loss": 0.4739, "mean_token_accuracy": 0.8485256433486938, "num_tokens": 46397344.0, "step": 1223 }, { "epoch": 0.15570538099478437, "ewc_loss": 0.0201416015625, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.9038746356964111, "learning_rate": 5.184400169563375e-06, "loss": 0.5345, "mean_token_accuracy": 0.8327842950820923, "num_tokens": 46437257.0, "step": 1224 }, { "epoch": 0.1558325912733749, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.8927295207977295, "learning_rate": 5.188639253921154e-06, "loss": 0.5088, "mean_token_accuracy": 0.8402165174484253, "num_tokens": 46477844.0, "step": 1225 }, { "epoch": 0.1559598015519654, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.9014184474945068, "learning_rate": 5.192878338278933e-06, "loss": 0.4735, "mean_token_accuracy": 0.849951982498169, "num_tokens": 46514556.0, "step": 1226 }, { "epoch": 0.1560870118305559, "ewc_loss": 0.019775390625, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.71661376953125e-05, "grad_norm": 1.8824186325073242, "learning_rate": 5.19711742263671e-06, "loss": 0.5477, "mean_token_accuracy": 0.8296213746070862, "num_tokens": 46559274.0, "step": 1227 }, { "epoch": 0.15621422210914643, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.926389455795288, "learning_rate": 5.201356506994489e-06, "loss": 0.4772, "mean_token_accuracy": 0.845788300037384, "num_tokens": 46591104.0, "step": 1228 }, { "epoch": 0.15634143238773693, "ewc_loss": 0.019775390625, "ewc_loss_diag": 2.5391578674316406e-05, "ewc_loss_parallel": 1.71661376953125e-05, "grad_norm": 1.8888031244277954, "learning_rate": 5.205595591352268e-06, "loss": 0.5659, "mean_token_accuracy": 0.8222606182098389, "num_tokens": 46631296.0, "step": 1229 }, { "epoch": 0.15646864266632743, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.5510787963867188e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.9093358516693115, "learning_rate": 5.209834675710047e-06, "loss": 0.5033, "mean_token_accuracy": 0.8437908887863159, "num_tokens": 46673079.0, "step": 1230 }, { "epoch": 0.15659585294491796, "ewc_loss": 0.0196533203125, "ewc_loss_diag": 2.5510787963867188e-05, "ewc_loss_parallel": 1.704692840576172e-05, "grad_norm": 1.9054921865463257, "learning_rate": 5.214073760067826e-06, "loss": 0.566, "mean_token_accuracy": 0.8241429924964905, "num_tokens": 46706820.0, "step": 1231 }, { "epoch": 0.15672306322350846, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.5510787963867188e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.914122462272644, "learning_rate": 5.2183128444256045e-06, "loss": 0.4945, "mean_token_accuracy": 0.8470996022224426, "num_tokens": 46742466.0, "step": 1232 }, { "epoch": 0.15685027350209896, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.5510787963867188e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.8958834409713745, "learning_rate": 5.222551928783383e-06, "loss": 0.4725, "mean_token_accuracy": 0.8492810726165771, "num_tokens": 46778882.0, "step": 1233 }, { "epoch": 0.1569774837806895, "ewc_loss": 0.0201416015625, "ewc_loss_diag": 2.5510787963867188e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.9123966693878174, "learning_rate": 5.226791013141161e-06, "loss": 0.5329, "mean_token_accuracy": 0.8367117643356323, "num_tokens": 46821509.0, "step": 1234 }, { "epoch": 0.15710469405928, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.562999725341797e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.8864552974700928, "learning_rate": 5.23103009749894e-06, "loss": 0.479, "mean_token_accuracy": 0.8485929369926453, "num_tokens": 46862247.0, "step": 1235 }, { "epoch": 0.1572319043378705, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.562999725341797e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.9346551895141602, "learning_rate": 5.235269181856719e-06, "loss": 0.502, "mean_token_accuracy": 0.8433778285980225, "num_tokens": 46906645.0, "step": 1236 }, { "epoch": 0.15735911461646102, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 2.562999725341797e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.9137113094329834, "learning_rate": 5.239508266214498e-06, "loss": 0.5486, "mean_token_accuracy": 0.8250929713249207, "num_tokens": 46940492.0, "step": 1237 }, { "epoch": 0.15748632489505152, "ewc_loss": 0.020263671875, "ewc_loss_diag": 2.574920654296875e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.9338455200195312, "learning_rate": 5.243747350572277e-06, "loss": 0.5078, "mean_token_accuracy": 0.843475341796875, "num_tokens": 46976871.0, "step": 1238 }, { "epoch": 0.15761353517364202, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.574920654296875e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.9417511224746704, "learning_rate": 5.247986434930056e-06, "loss": 0.5, "mean_token_accuracy": 0.8451030850410461, "num_tokens": 47013689.0, "step": 1239 }, { "epoch": 0.15774074545223254, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 2.586841583251953e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.9214450120925903, "learning_rate": 5.252225519287835e-06, "loss": 0.5598, "mean_token_accuracy": 0.8250799775123596, "num_tokens": 47050995.0, "step": 1240 }, { "epoch": 0.15786795573082305, "ewc_loss": 0.02001953125, "ewc_loss_diag": 2.586841583251953e-05, "ewc_loss_parallel": 1.7404556274414062e-05, "grad_norm": 1.9012893438339233, "learning_rate": 5.256464603645614e-06, "loss": 0.4877, "mean_token_accuracy": 0.8499581813812256, "num_tokens": 47088571.0, "step": 1241 }, { "epoch": 0.15799516600941357, "ewc_loss": 0.020263671875, "ewc_loss_diag": 2.5987625122070312e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.9092298746109009, "learning_rate": 5.260703688003392e-06, "loss": 0.5013, "mean_token_accuracy": 0.8439478874206543, "num_tokens": 47130547.0, "step": 1242 }, { "epoch": 0.15812237628800407, "ewc_loss": 0.0201416015625, "ewc_loss_diag": 2.586841583251953e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.9278621673583984, "learning_rate": 5.2649427723611705e-06, "loss": 0.5, "mean_token_accuracy": 0.8428401947021484, "num_tokens": 47162724.0, "step": 1243 }, { "epoch": 0.15824958656659457, "ewc_loss": 0.020263671875, "ewc_loss_diag": 2.586841583251953e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.9341315031051636, "learning_rate": 5.269181856718949e-06, "loss": 0.5414, "mean_token_accuracy": 0.8301957845687866, "num_tokens": 47200081.0, "step": 1244 }, { "epoch": 0.1583767968451851, "ewc_loss": 0.020263671875, "ewc_loss_diag": 2.586841583251953e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.9267399311065674, "learning_rate": 5.2734209410767275e-06, "loss": 0.4802, "mean_token_accuracy": 0.8516523241996765, "num_tokens": 47235074.0, "step": 1245 }, { "epoch": 0.1585040071237756, "ewc_loss": 0.020263671875, "ewc_loss_diag": 2.586841583251953e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.906469702720642, "learning_rate": 5.277660025434506e-06, "loss": 0.4396, "mean_token_accuracy": 0.859241247177124, "num_tokens": 47274258.0, "step": 1246 }, { "epoch": 0.1586312174023661, "ewc_loss": 0.020263671875, "ewc_loss_diag": 2.586841583251953e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.909216046333313, "learning_rate": 5.281899109792285e-06, "loss": 0.4662, "mean_token_accuracy": 0.853022038936615, "num_tokens": 47314232.0, "step": 1247 }, { "epoch": 0.15875842768095663, "ewc_loss": 0.0203857421875, "ewc_loss_diag": 2.5987625122070312e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.9165679216384888, "learning_rate": 5.286138194150064e-06, "loss": 0.5111, "mean_token_accuracy": 0.8417950868606567, "num_tokens": 47355245.0, "step": 1248 }, { "epoch": 0.15888563795954713, "ewc_loss": 0.0203857421875, "ewc_loss_diag": 2.5987625122070312e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.9313852787017822, "learning_rate": 5.290377278507842e-06, "loss": 0.545, "mean_token_accuracy": 0.8354203104972839, "num_tokens": 47394356.0, "step": 1249 }, { "epoch": 0.15901284823813763, "ewc_loss": 0.020263671875, "ewc_loss_diag": 2.5987625122070312e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.9205142259597778, "learning_rate": 5.294616362865621e-06, "loss": 0.4835, "mean_token_accuracy": 0.849411129951477, "num_tokens": 47435220.0, "step": 1250 }, { "epoch": 0.15914005851672816, "ewc_loss": 0.0203857421875, "ewc_loss_diag": 2.5987625122070312e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.997146725654602, "learning_rate": 5.2988554472234e-06, "loss": 0.5599, "mean_token_accuracy": 0.8335703611373901, "num_tokens": 47477294.0, "step": 1251 }, { "epoch": 0.15926726879531866, "ewc_loss": 0.0201416015625, "ewc_loss_diag": 2.5987625122070312e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.9182482957839966, "learning_rate": 5.303094531581179e-06, "loss": 0.5163, "mean_token_accuracy": 0.8382837176322937, "num_tokens": 47512216.0, "step": 1252 }, { "epoch": 0.15939447907390916, "ewc_loss": 0.0203857421875, "ewc_loss_diag": 2.5987625122070312e-05, "ewc_loss_parallel": 1.7881393432617188e-05, "grad_norm": 1.9443122148513794, "learning_rate": 5.307333615938958e-06, "loss": 0.4356, "mean_token_accuracy": 0.8647780418395996, "num_tokens": 47550075.0, "step": 1253 }, { "epoch": 0.1595216893524997, "ewc_loss": 0.0198974609375, "ewc_loss_diag": 2.5987625122070312e-05, "ewc_loss_parallel": 1.728534698486328e-05, "grad_norm": 1.9343661069869995, "learning_rate": 5.3115727002967366e-06, "loss": 0.5206, "mean_token_accuracy": 0.8369123935699463, "num_tokens": 47581595.0, "step": 1254 }, { "epoch": 0.1596488996310902, "ewc_loss": 0.0205078125, "ewc_loss_diag": 2.6106834411621094e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.9379769563674927, "learning_rate": 5.3158117846545155e-06, "loss": 0.4236, "mean_token_accuracy": 0.8628835678100586, "num_tokens": 47617336.0, "step": 1255 }, { "epoch": 0.1597761099096807, "ewc_loss": 0.0205078125, "ewc_loss_diag": 2.6106834411621094e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.911041498184204, "learning_rate": 5.320050869012294e-06, "loss": 0.4984, "mean_token_accuracy": 0.8455273509025574, "num_tokens": 47662701.0, "step": 1256 }, { "epoch": 0.15990332018827122, "ewc_loss": 0.0206298828125, "ewc_loss_diag": 2.6106834411621094e-05, "ewc_loss_parallel": 1.811981201171875e-05, "grad_norm": 1.9360355138778687, "learning_rate": 5.324289953370072e-06, "loss": 0.5475, "mean_token_accuracy": 0.8316063284873962, "num_tokens": 47703335.0, "step": 1257 }, { "epoch": 0.16003053046686172, "ewc_loss": 0.020751953125, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 1.811981201171875e-05, "grad_norm": 1.9095954895019531, "learning_rate": 5.328529037727851e-06, "loss": 0.4888, "mean_token_accuracy": 0.8465735912322998, "num_tokens": 47746980.0, "step": 1258 }, { "epoch": 0.16015774074545222, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.957075834274292, "learning_rate": 5.33276812208563e-06, "loss": 0.4857, "mean_token_accuracy": 0.8465107679367065, "num_tokens": 47784606.0, "step": 1259 }, { "epoch": 0.16028495102404275, "ewc_loss": 0.0205078125, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 1.7881393432617188e-05, "grad_norm": 1.9272352457046509, "learning_rate": 5.337007206443409e-06, "loss": 0.5212, "mean_token_accuracy": 0.8378770351409912, "num_tokens": 47821298.0, "step": 1260 }, { "epoch": 0.16041216130263325, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.9432646036148071, "learning_rate": 5.341246290801188e-06, "loss": 0.4849, "mean_token_accuracy": 0.8503453731536865, "num_tokens": 47863102.0, "step": 1261 }, { "epoch": 0.16053937158122375, "ewc_loss": 0.0206298828125, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.910636305809021, "learning_rate": 5.345485375158966e-06, "loss": 0.468, "mean_token_accuracy": 0.852428674697876, "num_tokens": 47903124.0, "step": 1262 }, { "epoch": 0.16066658185981428, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.6345252990722656e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.9586845636367798, "learning_rate": 5.349724459516745e-06, "loss": 0.4654, "mean_token_accuracy": 0.8509179949760437, "num_tokens": 47939441.0, "step": 1263 }, { "epoch": 0.16079379213840478, "ewc_loss": 0.0205078125, "ewc_loss_diag": 2.658367156982422e-05, "ewc_loss_parallel": 1.7881393432617188e-05, "grad_norm": 1.9139803647994995, "learning_rate": 5.353963543874523e-06, "loss": 0.5316, "mean_token_accuracy": 0.8357313275337219, "num_tokens": 47981236.0, "step": 1264 }, { "epoch": 0.1609210024169953, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.658367156982422e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.9833743572235107, "learning_rate": 5.358202628232302e-06, "loss": 0.4988, "mean_token_accuracy": 0.8462093472480774, "num_tokens": 48011431.0, "step": 1265 }, { "epoch": 0.1610482126955858, "ewc_loss": 0.0205078125, "ewc_loss_diag": 2.658367156982422e-05, "ewc_loss_parallel": 1.7881393432617188e-05, "grad_norm": 1.901297688484192, "learning_rate": 5.362441712590081e-06, "loss": 0.4862, "mean_token_accuracy": 0.848230242729187, "num_tokens": 48051003.0, "step": 1266 }, { "epoch": 0.1611754229741763, "ewc_loss": 0.02099609375, "ewc_loss_diag": 2.658367156982422e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 1.9778319597244263, "learning_rate": 5.3666807969478595e-06, "loss": 0.5313, "mean_token_accuracy": 0.8349945545196533, "num_tokens": 48091239.0, "step": 1267 }, { "epoch": 0.16130263325276684, "ewc_loss": 0.020263671875, "ewc_loss_diag": 2.658367156982422e-05, "ewc_loss_parallel": 1.7523765563964844e-05, "grad_norm": 1.926853060722351, "learning_rate": 5.370919881305638e-06, "loss": 0.5183, "mean_token_accuracy": 0.8425315618515015, "num_tokens": 48126021.0, "step": 1268 }, { "epoch": 0.16142984353135734, "ewc_loss": 0.0211181640625, "ewc_loss_diag": 2.658367156982422e-05, "ewc_loss_parallel": 1.8477439880371094e-05, "grad_norm": 1.9887412786483765, "learning_rate": 5.375158965663417e-06, "loss": 0.5022, "mean_token_accuracy": 0.8450275659561157, "num_tokens": 48160459.0, "step": 1269 }, { "epoch": 0.16155705380994784, "ewc_loss": 0.0203857421875, "ewc_loss_diag": 2.658367156982422e-05, "ewc_loss_parallel": 1.7642974853515625e-05, "grad_norm": 1.9047837257385254, "learning_rate": 5.379398050021196e-06, "loss": 0.4897, "mean_token_accuracy": 0.8476810455322266, "num_tokens": 48200199.0, "step": 1270 }, { "epoch": 0.16168426408853837, "ewc_loss": 0.02099609375, "ewc_loss_diag": 2.6702880859375e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 1.9935683012008667, "learning_rate": 5.383637134378975e-06, "loss": 0.4735, "mean_token_accuracy": 0.8485430479049683, "num_tokens": 48233926.0, "step": 1271 }, { "epoch": 0.16181147436712887, "ewc_loss": 0.0205078125, "ewc_loss_diag": 2.6702880859375e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.9062681198120117, "learning_rate": 5.387876218736753e-06, "loss": 0.5662, "mean_token_accuracy": 0.8205777406692505, "num_tokens": 48279667.0, "step": 1272 }, { "epoch": 0.16193868464571937, "ewc_loss": 0.021240234375, "ewc_loss_diag": 2.682209014892578e-05, "ewc_loss_parallel": 1.8596649169921875e-05, "grad_norm": 1.9847683906555176, "learning_rate": 5.392115303094532e-06, "loss": 0.4555, "mean_token_accuracy": 0.8579322099685669, "num_tokens": 48319185.0, "step": 1273 }, { "epoch": 0.1620658949243099, "ewc_loss": 0.0205078125, "ewc_loss_diag": 2.6941299438476562e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.9135353565216064, "learning_rate": 5.396354387452311e-06, "loss": 0.4427, "mean_token_accuracy": 0.86287522315979, "num_tokens": 48357250.0, "step": 1274 }, { "epoch": 0.1621931052029004, "ewc_loss": 0.0211181640625, "ewc_loss_diag": 2.7060508728027344e-05, "ewc_loss_parallel": 1.8477439880371094e-05, "grad_norm": 1.998223066329956, "learning_rate": 5.40059347181009e-06, "loss": 0.5265, "mean_token_accuracy": 0.8366482853889465, "num_tokens": 48396577.0, "step": 1275 }, { "epoch": 0.1623203154814909, "ewc_loss": 0.0205078125, "ewc_loss_diag": 2.7060508728027344e-05, "ewc_loss_parallel": 1.7881393432617188e-05, "grad_norm": 1.9362105131149292, "learning_rate": 5.404832556167869e-06, "loss": 0.5235, "mean_token_accuracy": 0.8348569869995117, "num_tokens": 48431360.0, "step": 1276 }, { "epoch": 0.16244752576008142, "ewc_loss": 0.0211181640625, "ewc_loss_diag": 2.7060508728027344e-05, "ewc_loss_parallel": 1.8477439880371094e-05, "grad_norm": 1.98043692111969, "learning_rate": 5.4090716405256475e-06, "loss": 0.4893, "mean_token_accuracy": 0.8495104908943176, "num_tokens": 48471640.0, "step": 1277 }, { "epoch": 0.16257473603867192, "ewc_loss": 0.0206298828125, "ewc_loss_diag": 2.7179718017578125e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.9287700653076172, "learning_rate": 5.413310724883426e-06, "loss": 0.5454, "mean_token_accuracy": 0.8238011002540588, "num_tokens": 48509870.0, "step": 1278 }, { "epoch": 0.16270194631726242, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.7179718017578125e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.9779036045074463, "learning_rate": 5.417549809241205e-06, "loss": 0.5228, "mean_token_accuracy": 0.8402390480041504, "num_tokens": 48554174.0, "step": 1279 }, { "epoch": 0.16282915659585295, "ewc_loss": 0.020751953125, "ewc_loss_diag": 2.7179718017578125e-05, "ewc_loss_parallel": 1.811981201171875e-05, "grad_norm": 1.9522509574890137, "learning_rate": 5.4217888935989825e-06, "loss": 0.5284, "mean_token_accuracy": 0.8364766836166382, "num_tokens": 48588590.0, "step": 1280 }, { "epoch": 0.16295636687444345, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.7179718017578125e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.9643464088439941, "learning_rate": 5.426027977956761e-06, "loss": 0.5472, "mean_token_accuracy": 0.8288629055023193, "num_tokens": 48633567.0, "step": 1281 }, { "epoch": 0.16308357715303395, "ewc_loss": 0.0206298828125, "ewc_loss_diag": 2.7298927307128906e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.9568414688110352, "learning_rate": 5.43026706231454e-06, "loss": 0.567, "mean_token_accuracy": 0.8246039152145386, "num_tokens": 48673764.0, "step": 1282 }, { "epoch": 0.16321078743162448, "ewc_loss": 0.020751953125, "ewc_loss_diag": 2.7298927307128906e-05, "ewc_loss_parallel": 1.811981201171875e-05, "grad_norm": 1.946963906288147, "learning_rate": 5.434506146672319e-06, "loss": 0.5433, "mean_token_accuracy": 0.831337571144104, "num_tokens": 48722531.0, "step": 1283 }, { "epoch": 0.16333799771021498, "ewc_loss": 0.0206298828125, "ewc_loss_diag": 2.7298927307128906e-05, "ewc_loss_parallel": 1.800060272216797e-05, "grad_norm": 1.972567081451416, "learning_rate": 5.438745231030098e-06, "loss": 0.5336, "mean_token_accuracy": 0.8309614658355713, "num_tokens": 48758273.0, "step": 1284 }, { "epoch": 0.16346520798880548, "ewc_loss": 0.020751953125, "ewc_loss_diag": 2.7298927307128906e-05, "ewc_loss_parallel": 1.811981201171875e-05, "grad_norm": 1.9481306076049805, "learning_rate": 5.442984315387877e-06, "loss": 0.4279, "mean_token_accuracy": 0.8639746308326721, "num_tokens": 48793016.0, "step": 1285 }, { "epoch": 0.163592418267396, "ewc_loss": 0.02099609375, "ewc_loss_diag": 2.753734588623047e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.9652631282806396, "learning_rate": 5.447223399745656e-06, "loss": 0.5014, "mean_token_accuracy": 0.8460062742233276, "num_tokens": 48832204.0, "step": 1286 }, { "epoch": 0.1637196285459865, "ewc_loss": 0.02099609375, "ewc_loss_diag": 2.753734588623047e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 1.946706771850586, "learning_rate": 5.451462484103434e-06, "loss": 0.4739, "mean_token_accuracy": 0.8529937863349915, "num_tokens": 48867199.0, "step": 1287 }, { "epoch": 0.163846838824577, "ewc_loss": 0.02099609375, "ewc_loss_diag": 2.753734588623047e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.9775664806365967, "learning_rate": 5.455701568461213e-06, "loss": 0.4917, "mean_token_accuracy": 0.8476836681365967, "num_tokens": 48901423.0, "step": 1288 }, { "epoch": 0.16397404910316754, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.811981201171875e-05, "grad_norm": 1.956678032875061, "learning_rate": 5.459940652818992e-06, "loss": 0.5732, "mean_token_accuracy": 0.8223620653152466, "num_tokens": 48941883.0, "step": 1289 }, { "epoch": 0.16410125938175804, "ewc_loss": 0.0211181640625, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 2.0007588863372803, "learning_rate": 5.4641797371767705e-06, "loss": 0.5035, "mean_token_accuracy": 0.8418571949005127, "num_tokens": 48971764.0, "step": 1290 }, { "epoch": 0.16422846966034857, "ewc_loss": 0.02099609375, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.94469153881073, "learning_rate": 5.468418821534549e-06, "loss": 0.4451, "mean_token_accuracy": 0.8582192063331604, "num_tokens": 49003687.0, "step": 1291 }, { "epoch": 0.16435567993893907, "ewc_loss": 0.0213623046875, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.8596649169921875e-05, "grad_norm": 1.9869972467422485, "learning_rate": 5.472657905892328e-06, "loss": 0.453, "mean_token_accuracy": 0.8581277132034302, "num_tokens": 49036896.0, "step": 1292 }, { "epoch": 0.16448289021752957, "ewc_loss": 0.02099609375, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.945733666419983, "learning_rate": 5.476896990250107e-06, "loss": 0.5688, "mean_token_accuracy": 0.8236019611358643, "num_tokens": 49076549.0, "step": 1293 }, { "epoch": 0.1646101004961201, "ewc_loss": 0.0213623046875, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.8596649169921875e-05, "grad_norm": 1.9814242124557495, "learning_rate": 5.481136074607886e-06, "loss": 0.4546, "mean_token_accuracy": 0.8529297113418579, "num_tokens": 49114507.0, "step": 1294 }, { "epoch": 0.1647373107747106, "ewc_loss": 0.0211181640625, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 1.9299554824829102, "learning_rate": 5.485375158965663e-06, "loss": 0.465, "mean_token_accuracy": 0.8560978770256042, "num_tokens": 49159794.0, "step": 1295 }, { "epoch": 0.1648645210533011, "ewc_loss": 0.0213623046875, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.8596649169921875e-05, "grad_norm": 2.0111169815063477, "learning_rate": 5.489614243323442e-06, "loss": 0.5051, "mean_token_accuracy": 0.8423638343811035, "num_tokens": 49198766.0, "step": 1296 }, { "epoch": 0.16499173133189163, "ewc_loss": 0.021240234375, "ewc_loss_diag": 2.765655517578125e-05, "ewc_loss_parallel": 1.8477439880371094e-05, "grad_norm": 1.9287506341934204, "learning_rate": 5.493853327681221e-06, "loss": 0.4489, "mean_token_accuracy": 0.8591877222061157, "num_tokens": 49236481.0, "step": 1297 }, { "epoch": 0.16511894161048213, "ewc_loss": 0.021484375, "ewc_loss_diag": 2.777576446533203e-05, "ewc_loss_parallel": 1.8715858459472656e-05, "grad_norm": 2.0413708686828613, "learning_rate": 5.498092412039e-06, "loss": 0.5725, "mean_token_accuracy": 0.8270360231399536, "num_tokens": 49271730.0, "step": 1298 }, { "epoch": 0.16524615188907263, "ewc_loss": 0.0211181640625, "ewc_loss_diag": 2.7894973754882812e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 1.8994747400283813, "learning_rate": 5.502331496396779e-06, "loss": 0.4628, "mean_token_accuracy": 0.8537296056747437, "num_tokens": 49316208.0, "step": 1299 }, { "epoch": 0.16537336216766316, "ewc_loss": 0.021484375, "ewc_loss_diag": 2.7894973754882812e-05, "ewc_loss_parallel": 1.8715858459472656e-05, "grad_norm": 2.061596155166626, "learning_rate": 5.506570580754558e-06, "loss": 0.5011, "mean_token_accuracy": 0.8434716463088989, "num_tokens": 49355343.0, "step": 1300 }, { "epoch": 0.16550057244625366, "ewc_loss": 0.0208740234375, "ewc_loss_diag": 2.7894973754882812e-05, "ewc_loss_parallel": 1.811981201171875e-05, "grad_norm": 1.9180433750152588, "learning_rate": 5.5108096651123365e-06, "loss": 0.4763, "mean_token_accuracy": 0.8518010377883911, "num_tokens": 49394055.0, "step": 1301 }, { "epoch": 0.16562778272484416, "ewc_loss": 0.0218505859375, "ewc_loss_diag": 2.7894973754882812e-05, "ewc_loss_parallel": 1.9073486328125e-05, "grad_norm": 2.0799784660339355, "learning_rate": 5.5150487494701146e-06, "loss": 0.5048, "mean_token_accuracy": 0.8448261022567749, "num_tokens": 49432466.0, "step": 1302 }, { "epoch": 0.1657549930034347, "ewc_loss": 0.0211181640625, "ewc_loss_diag": 2.7894973754882812e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 1.9098598957061768, "learning_rate": 5.5192878338278934e-06, "loss": 0.4866, "mean_token_accuracy": 0.8458842039108276, "num_tokens": 49470746.0, "step": 1303 }, { "epoch": 0.1658822032820252, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.8014183044433594e-05, "ewc_loss_parallel": 1.9431114196777344e-05, "grad_norm": 2.1213369369506836, "learning_rate": 5.523526918185672e-06, "loss": 0.5329, "mean_token_accuracy": 0.8318308591842651, "num_tokens": 49502879.0, "step": 1304 }, { "epoch": 0.1660094135606157, "ewc_loss": 0.02099609375, "ewc_loss_diag": 2.8014183044433594e-05, "ewc_loss_parallel": 1.823902130126953e-05, "grad_norm": 1.921502709388733, "learning_rate": 5.527766002543451e-06, "loss": 0.4493, "mean_token_accuracy": 0.8552193641662598, "num_tokens": 49534875.0, "step": 1305 }, { "epoch": 0.16613662383920622, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.8014183044433594e-05, "ewc_loss_parallel": 1.9431114196777344e-05, "grad_norm": 2.1116862297058105, "learning_rate": 5.53200508690123e-06, "loss": 0.5186, "mean_token_accuracy": 0.8418022394180298, "num_tokens": 49573537.0, "step": 1306 }, { "epoch": 0.16626383411779672, "ewc_loss": 0.0211181640625, "ewc_loss_diag": 2.8014183044433594e-05, "ewc_loss_parallel": 1.8358230590820312e-05, "grad_norm": 1.90589439868927, "learning_rate": 5.536244171259009e-06, "loss": 0.4825, "mean_token_accuracy": 0.8502906560897827, "num_tokens": 49615190.0, "step": 1307 }, { "epoch": 0.16639104439638722, "ewc_loss": 0.0223388671875, "ewc_loss_diag": 2.8014183044433594e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 4.148524284362793, "learning_rate": 5.540483255616788e-06, "loss": 0.5814, "mean_token_accuracy": 0.8203314542770386, "num_tokens": 49649788.0, "step": 1308 }, { "epoch": 0.16651825467497774, "ewc_loss": 0.03564453125, "ewc_loss_diag": 2.8133392333984375e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.6881678104400635, "learning_rate": 5.544722339974567e-06, "loss": 0.5052, "mean_token_accuracy": 0.8468950986862183, "num_tokens": 49683728.0, "step": 1309 }, { "epoch": 0.16664546495356825, "ewc_loss": 0.0206298828125, "ewc_loss_diag": 2.8133392333984375e-05, "ewc_loss_parallel": 1.7762184143066406e-05, "grad_norm": 1.796877145767212, "learning_rate": 5.548961424332344e-06, "loss": 0.4833, "mean_token_accuracy": 0.8429431319236755, "num_tokens": 49719331.0, "step": 1310 }, { "epoch": 0.16677267523215875, "ewc_loss": 0.02685546875, "ewc_loss_diag": 2.8133392333984375e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.3439455032348633, "learning_rate": 5.553200508690123e-06, "loss": 0.4479, "mean_token_accuracy": 0.8610774278640747, "num_tokens": 49760158.0, "step": 1311 }, { "epoch": 0.16689988551074927, "ewc_loss": 0.0242919921875, "ewc_loss_diag": 2.8133392333984375e-05, "ewc_loss_parallel": 2.1457672119140625e-05, "grad_norm": 1.990382432937622, "learning_rate": 5.557439593047902e-06, "loss": 0.5466, "mean_token_accuracy": 0.8294022083282471, "num_tokens": 49795285.0, "step": 1312 }, { "epoch": 0.16702709578933977, "ewc_loss": 0.0257568359375, "ewc_loss_diag": 2.8252601623535156e-05, "ewc_loss_parallel": 2.300739288330078e-05, "grad_norm": 2.251720905303955, "learning_rate": 5.561678677405681e-06, "loss": 0.5027, "mean_token_accuracy": 0.8450021743774414, "num_tokens": 49839071.0, "step": 1313 }, { "epoch": 0.16715430606793028, "ewc_loss": 0.02490234375, "ewc_loss_diag": 2.8252601623535156e-05, "ewc_loss_parallel": 2.205371856689453e-05, "grad_norm": 2.0511727333068848, "learning_rate": 5.5659177617634595e-06, "loss": 0.4803, "mean_token_accuracy": 0.8519135117530823, "num_tokens": 49870063.0, "step": 1314 }, { "epoch": 0.1672815163465208, "ewc_loss": 0.0242919921875, "ewc_loss_diag": 2.8252601623535156e-05, "ewc_loss_parallel": 2.1457672119140625e-05, "grad_norm": 2.1869654655456543, "learning_rate": 5.570156846121238e-06, "loss": 0.5161, "mean_token_accuracy": 0.8383403420448303, "num_tokens": 49908232.0, "step": 1315 }, { "epoch": 0.1674087266251113, "ewc_loss": 0.0242919921875, "ewc_loss_diag": 2.8252601623535156e-05, "ewc_loss_parallel": 2.1457672119140625e-05, "grad_norm": 2.0537915229797363, "learning_rate": 5.574395930479017e-06, "loss": 0.5383, "mean_token_accuracy": 0.8345918655395508, "num_tokens": 49944676.0, "step": 1316 }, { "epoch": 0.16753593690370183, "ewc_loss": 0.0235595703125, "ewc_loss_diag": 2.8252601623535156e-05, "ewc_loss_parallel": 2.0742416381835938e-05, "grad_norm": 2.1302623748779297, "learning_rate": 5.578635014836796e-06, "loss": 0.4814, "mean_token_accuracy": 0.8524182438850403, "num_tokens": 49978472.0, "step": 1317 }, { "epoch": 0.16766314718229233, "ewc_loss": 0.02392578125, "ewc_loss_diag": 2.8371810913085938e-05, "ewc_loss_parallel": 2.110004425048828e-05, "grad_norm": 2.0649023056030273, "learning_rate": 5.582874099194574e-06, "loss": 0.5322, "mean_token_accuracy": 0.8364549875259399, "num_tokens": 50010564.0, "step": 1318 }, { "epoch": 0.16779035746088283, "ewc_loss": 0.02294921875, "ewc_loss_diag": 2.8371810913085938e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 2.069204330444336, "learning_rate": 5.587113183552353e-06, "loss": 0.5191, "mean_token_accuracy": 0.8383880853652954, "num_tokens": 50049538.0, "step": 1319 }, { "epoch": 0.16791756773947336, "ewc_loss": 0.023193359375, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.014021396636963, "learning_rate": 5.591352267910132e-06, "loss": 0.4761, "mean_token_accuracy": 0.8523344993591309, "num_tokens": 50090779.0, "step": 1320 }, { "epoch": 0.16804477801806386, "ewc_loss": 0.022705078125, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.057628631591797, "learning_rate": 5.595591352267911e-06, "loss": 0.5227, "mean_token_accuracy": 0.8397912979125977, "num_tokens": 50129527.0, "step": 1321 }, { "epoch": 0.16817198829665436, "ewc_loss": 0.022705078125, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 2.0175740718841553, "learning_rate": 5.59983043662569e-06, "loss": 0.4859, "mean_token_accuracy": 0.8484375476837158, "num_tokens": 50167950.0, "step": 1322 }, { "epoch": 0.1682991985752449, "ewc_loss": 0.022705078125, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 2.0444109439849854, "learning_rate": 5.604069520983469e-06, "loss": 0.4884, "mean_token_accuracy": 0.8492610454559326, "num_tokens": 50205079.0, "step": 1323 }, { "epoch": 0.1684264088538354, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.86102294921875e-05, "ewc_loss_parallel": 1.9311904907226562e-05, "grad_norm": 2.0165905952453613, "learning_rate": 5.6083086053412475e-06, "loss": 0.4784, "mean_token_accuracy": 0.849602222442627, "num_tokens": 50237909.0, "step": 1324 }, { "epoch": 0.1685536191324259, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.872943878173828e-05, "ewc_loss_parallel": 1.9311904907226562e-05, "grad_norm": 2.038926601409912, "learning_rate": 5.612547689699025e-06, "loss": 0.54, "mean_token_accuracy": 0.8357502222061157, "num_tokens": 50272825.0, "step": 1325 }, { "epoch": 0.16868082941101642, "ewc_loss": 0.02197265625, "ewc_loss_diag": 2.872943878173828e-05, "ewc_loss_parallel": 1.9073486328125e-05, "grad_norm": 1.988911509513855, "learning_rate": 5.6167867740568036e-06, "loss": 0.5244, "mean_token_accuracy": 0.8400802612304688, "num_tokens": 50312964.0, "step": 1326 }, { "epoch": 0.16880803968960692, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.872943878173828e-05, "ewc_loss_parallel": 1.9311904907226562e-05, "grad_norm": 2.0209879875183105, "learning_rate": 5.6210258584145824e-06, "loss": 0.5256, "mean_token_accuracy": 0.835117518901825, "num_tokens": 50355912.0, "step": 1327 }, { "epoch": 0.16893524996819742, "ewc_loss": 0.02197265625, "ewc_loss_diag": 2.872943878173828e-05, "ewc_loss_parallel": 1.9073486328125e-05, "grad_norm": 1.9850984811782837, "learning_rate": 5.625264942772361e-06, "loss": 0.5087, "mean_token_accuracy": 0.8449791073799133, "num_tokens": 50395875.0, "step": 1328 }, { "epoch": 0.16906246024678795, "ewc_loss": 0.02197265625, "ewc_loss_diag": 2.872943878173828e-05, "ewc_loss_parallel": 1.919269561767578e-05, "grad_norm": 2.031914472579956, "learning_rate": 5.62950402713014e-06, "loss": 0.5611, "mean_token_accuracy": 0.8323343992233276, "num_tokens": 50435628.0, "step": 1329 }, { "epoch": 0.16918967052537845, "ewc_loss": 0.021728515625, "ewc_loss_diag": 2.872943878173828e-05, "ewc_loss_parallel": 1.895427703857422e-05, "grad_norm": 1.9895564317703247, "learning_rate": 5.633743111487919e-06, "loss": 0.5641, "mean_token_accuracy": 0.8223012685775757, "num_tokens": 50474812.0, "step": 1330 }, { "epoch": 0.16931688080396895, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.8848648071289062e-05, "ewc_loss_parallel": 1.919269561767578e-05, "grad_norm": 2.032536268234253, "learning_rate": 5.637982195845698e-06, "loss": 0.4444, "mean_token_accuracy": 0.864655077457428, "num_tokens": 50510058.0, "step": 1331 }, { "epoch": 0.16944409108255948, "ewc_loss": 0.0218505859375, "ewc_loss_diag": 2.8848648071289062e-05, "ewc_loss_parallel": 1.895427703857422e-05, "grad_norm": 1.9845292568206787, "learning_rate": 5.642221280203477e-06, "loss": 0.4462, "mean_token_accuracy": 0.8610249161720276, "num_tokens": 50546282.0, "step": 1332 }, { "epoch": 0.16957130136114998, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.8848648071289062e-05, "ewc_loss_parallel": 1.919269561767578e-05, "grad_norm": 2.0625007152557373, "learning_rate": 5.646460364561255e-06, "loss": 0.4615, "mean_token_accuracy": 0.8537216186523438, "num_tokens": 50577457.0, "step": 1333 }, { "epoch": 0.16969851163974048, "ewc_loss": 0.0216064453125, "ewc_loss_diag": 2.8848648071289062e-05, "ewc_loss_parallel": 1.8715858459472656e-05, "grad_norm": 1.9748079776763916, "learning_rate": 5.650699448919034e-06, "loss": 0.5595, "mean_token_accuracy": 0.8267068862915039, "num_tokens": 50619754.0, "step": 1334 }, { "epoch": 0.169825721918331, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.919269561767578e-05, "grad_norm": 2.0571234226226807, "learning_rate": 5.654938533276813e-06, "loss": 0.4954, "mean_token_accuracy": 0.8452955484390259, "num_tokens": 50660674.0, "step": 1335 }, { "epoch": 0.1699529321969215, "ewc_loss": 0.021728515625, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.8835067749023438e-05, "grad_norm": 1.9654494524002075, "learning_rate": 5.6591776176345916e-06, "loss": 0.4883, "mean_token_accuracy": 0.8470029830932617, "num_tokens": 50698404.0, "step": 1336 }, { "epoch": 0.170080142475512, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.919269561767578e-05, "grad_norm": 2.0608959197998047, "learning_rate": 5.6634167019923704e-06, "loss": 0.4967, "mean_token_accuracy": 0.8446415662765503, "num_tokens": 50736329.0, "step": 1337 }, { "epoch": 0.17020735275410254, "ewc_loss": 0.021728515625, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.8835067749023438e-05, "grad_norm": 1.9868218898773193, "learning_rate": 5.667655786350149e-06, "loss": 0.5543, "mean_token_accuracy": 0.820986270904541, "num_tokens": 50770065.0, "step": 1338 }, { "epoch": 0.17033456303269304, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.919269561767578e-05, "grad_norm": 2.0565879344940186, "learning_rate": 5.671894870707928e-06, "loss": 0.4906, "mean_token_accuracy": 0.8504403233528137, "num_tokens": 50811603.0, "step": 1339 }, { "epoch": 0.17046177331128357, "ewc_loss": 0.021728515625, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.8835067749023438e-05, "grad_norm": 1.9653762578964233, "learning_rate": 5.676133955065705e-06, "loss": 0.4312, "mean_token_accuracy": 0.8619799017906189, "num_tokens": 50846141.0, "step": 1340 }, { "epoch": 0.17058898358987407, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.9311904907226562e-05, "grad_norm": 2.0629730224609375, "learning_rate": 5.680373039423484e-06, "loss": 0.4794, "mean_token_accuracy": 0.8510947823524475, "num_tokens": 50885511.0, "step": 1341 }, { "epoch": 0.17071619386846457, "ewc_loss": 0.021728515625, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.8835067749023438e-05, "grad_norm": 1.981071949005127, "learning_rate": 5.684612123781263e-06, "loss": 0.5234, "mean_token_accuracy": 0.8339061737060547, "num_tokens": 50926374.0, "step": 1342 }, { "epoch": 0.1708434041470551, "ewc_loss": 0.0220947265625, "ewc_loss_diag": 2.8967857360839844e-05, "ewc_loss_parallel": 1.919269561767578e-05, "grad_norm": 2.0725715160369873, "learning_rate": 5.688851208139042e-06, "loss": 0.5364, "mean_token_accuracy": 0.8296979665756226, "num_tokens": 50965453.0, "step": 1343 }, { "epoch": 0.1709706144256456, "ewc_loss": 0.0216064453125, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.8715858459472656e-05, "grad_norm": 1.9648728370666504, "learning_rate": 5.693090292496821e-06, "loss": 0.5269, "mean_token_accuracy": 0.8408743143081665, "num_tokens": 51007921.0, "step": 1344 }, { "epoch": 0.1710978247042361, "ewc_loss": 0.0224609375, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 2.080451011657715, "learning_rate": 5.6973293768546e-06, "loss": 0.5042, "mean_token_accuracy": 0.8424428105354309, "num_tokens": 51046982.0, "step": 1345 }, { "epoch": 0.17122503498282662, "ewc_loss": 0.0216064453125, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.8715858459472656e-05, "grad_norm": 1.9707231521606445, "learning_rate": 5.701568461212379e-06, "loss": 0.441, "mean_token_accuracy": 0.8623793721199036, "num_tokens": 51083192.0, "step": 1346 }, { "epoch": 0.17135224526141712, "ewc_loss": 0.0224609375, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 2.0807197093963623, "learning_rate": 5.705807545570158e-06, "loss": 0.4519, "mean_token_accuracy": 0.8584145307540894, "num_tokens": 51125258.0, "step": 1347 }, { "epoch": 0.17147945554000762, "ewc_loss": 0.021728515625, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.8835067749023438e-05, "grad_norm": 1.9691355228424072, "learning_rate": 5.710046629927936e-06, "loss": 0.4871, "mean_token_accuracy": 0.8476909399032593, "num_tokens": 51164036.0, "step": 1348 }, { "epoch": 0.17160666581859815, "ewc_loss": 0.0224609375, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 2.094848155975342, "learning_rate": 5.7142857142857145e-06, "loss": 0.4536, "mean_token_accuracy": 0.8567363023757935, "num_tokens": 51201288.0, "step": 1349 }, { "epoch": 0.17173387609718865, "ewc_loss": 0.0216064453125, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.8715858459472656e-05, "grad_norm": 1.978142499923706, "learning_rate": 5.718524798643493e-06, "loss": 0.5216, "mean_token_accuracy": 0.8377020955085754, "num_tokens": 51235931.0, "step": 1350 }, { "epoch": 0.17186108637577915, "ewc_loss": 0.0225830078125, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.9669532775878906e-05, "grad_norm": 2.0975425243377686, "learning_rate": 5.722763883001272e-06, "loss": 0.5293, "mean_token_accuracy": 0.8375422954559326, "num_tokens": 51277188.0, "step": 1351 }, { "epoch": 0.17198829665436968, "ewc_loss": 0.0218505859375, "ewc_loss_diag": 2.9206275939941406e-05, "ewc_loss_parallel": 1.895427703857422e-05, "grad_norm": 1.9789221286773682, "learning_rate": 5.727002967359051e-06, "loss": 0.4961, "mean_token_accuracy": 0.8438878655433655, "num_tokens": 51313829.0, "step": 1352 }, { "epoch": 0.17211550693296018, "ewc_loss": 0.0225830078125, "ewc_loss_diag": 2.9325485229492188e-05, "ewc_loss_parallel": 1.9669532775878906e-05, "grad_norm": 2.117011785507202, "learning_rate": 5.73124205171683e-06, "loss": 0.5582, "mean_token_accuracy": 0.8235225677490234, "num_tokens": 51355678.0, "step": 1353 }, { "epoch": 0.17224271721155068, "ewc_loss": 0.0218505859375, "ewc_loss_diag": 2.9325485229492188e-05, "ewc_loss_parallel": 1.895427703857422e-05, "grad_norm": 1.9775259494781494, "learning_rate": 5.735481136074609e-06, "loss": 0.4681, "mean_token_accuracy": 0.8526328802108765, "num_tokens": 51390125.0, "step": 1354 }, { "epoch": 0.1723699274901412, "ewc_loss": 0.0225830078125, "ewc_loss_diag": 2.9325485229492188e-05, "ewc_loss_parallel": 1.9669532775878906e-05, "grad_norm": 2.1141273975372314, "learning_rate": 5.739720220432386e-06, "loss": 0.4692, "mean_token_accuracy": 0.8505713939666748, "num_tokens": 51426384.0, "step": 1355 }, { "epoch": 0.1724971377687317, "ewc_loss": 0.0218505859375, "ewc_loss_diag": 2.9325485229492188e-05, "ewc_loss_parallel": 1.895427703857422e-05, "grad_norm": 1.9744938611984253, "learning_rate": 5.743959304790165e-06, "loss": 0.5631, "mean_token_accuracy": 0.8294382095336914, "num_tokens": 51467087.0, "step": 1356 }, { "epoch": 0.1726243480473222, "ewc_loss": 0.022705078125, "ewc_loss_diag": 2.944469451904297e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 2.1283833980560303, "learning_rate": 5.748198389147944e-06, "loss": 0.5199, "mean_token_accuracy": 0.839650571346283, "num_tokens": 51505126.0, "step": 1357 }, { "epoch": 0.17275155832591274, "ewc_loss": 0.02197265625, "ewc_loss_diag": 2.944469451904297e-05, "ewc_loss_parallel": 1.9073486328125e-05, "grad_norm": 1.987120509147644, "learning_rate": 5.752437473505723e-06, "loss": 0.5416, "mean_token_accuracy": 0.8316553831100464, "num_tokens": 51546623.0, "step": 1358 }, { "epoch": 0.17287876860450324, "ewc_loss": 0.0228271484375, "ewc_loss_diag": 2.944469451904297e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.1335983276367188, "learning_rate": 5.756676557863502e-06, "loss": 0.5135, "mean_token_accuracy": 0.8393502831459045, "num_tokens": 51583175.0, "step": 1359 }, { "epoch": 0.17300597888309374, "ewc_loss": 0.02197265625, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 1.9073486328125e-05, "grad_norm": 1.9638651609420776, "learning_rate": 5.7609156422212805e-06, "loss": 0.4691, "mean_token_accuracy": 0.8516343832015991, "num_tokens": 51621454.0, "step": 1360 }, { "epoch": 0.17313318916168427, "ewc_loss": 0.0230712890625, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 2.14953351020813, "learning_rate": 5.7651547265790594e-06, "loss": 0.4938, "mean_token_accuracy": 0.8467224836349487, "num_tokens": 51659482.0, "step": 1361 }, { "epoch": 0.17326039944027477, "ewc_loss": 0.02197265625, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 1.9073486328125e-05, "grad_norm": 1.9763115644454956, "learning_rate": 5.769393810936838e-06, "loss": 0.4951, "mean_token_accuracy": 0.8453032970428467, "num_tokens": 51695478.0, "step": 1362 }, { "epoch": 0.17338760971886527, "ewc_loss": 0.0230712890625, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 2.15844988822937, "learning_rate": 5.773632895294616e-06, "loss": 0.5422, "mean_token_accuracy": 0.8288507461547852, "num_tokens": 51736436.0, "step": 1363 }, { "epoch": 0.1735148199974558, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 1.9311904907226562e-05, "grad_norm": 1.9777376651763916, "learning_rate": 5.777871979652395e-06, "loss": 0.4733, "mean_token_accuracy": 0.8520653247833252, "num_tokens": 51772502.0, "step": 1364 }, { "epoch": 0.1736420302760463, "ewc_loss": 0.0230712890625, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 2.1558239459991455, "learning_rate": 5.782111064010174e-06, "loss": 0.5078, "mean_token_accuracy": 0.84043288230896, "num_tokens": 51816343.0, "step": 1365 }, { "epoch": 0.17376924055463683, "ewc_loss": 0.022216796875, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 1.9311904907226562e-05, "grad_norm": 1.9867275953292847, "learning_rate": 5.786350148367953e-06, "loss": 0.4994, "mean_token_accuracy": 0.8445444107055664, "num_tokens": 51861317.0, "step": 1366 }, { "epoch": 0.17389645083322733, "ewc_loss": 0.0233154296875, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.1651127338409424, "learning_rate": 5.790589232725732e-06, "loss": 0.5039, "mean_token_accuracy": 0.8405166268348694, "num_tokens": 51902994.0, "step": 1367 }, { "epoch": 0.17402366111181783, "ewc_loss": 0.0224609375, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 1.9964063167572021, "learning_rate": 5.794828317083511e-06, "loss": 0.4879, "mean_token_accuracy": 0.8488590121269226, "num_tokens": 51936374.0, "step": 1368 }, { "epoch": 0.17415087139040836, "ewc_loss": 0.0230712890625, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 2.177612781524658, "learning_rate": 5.79906740144129e-06, "loss": 0.488, "mean_token_accuracy": 0.8476104736328125, "num_tokens": 51970235.0, "step": 1369 }, { "epoch": 0.17427808166899886, "ewc_loss": 0.0225830078125, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 1.9669532775878906e-05, "grad_norm": 1.9848179817199707, "learning_rate": 5.8033064857990685e-06, "loss": 0.476, "mean_token_accuracy": 0.853035569190979, "num_tokens": 52006220.0, "step": 1370 }, { "epoch": 0.17440529194758936, "ewc_loss": 0.0233154296875, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.163853168487549, "learning_rate": 5.807545570156847e-06, "loss": 0.459, "mean_token_accuracy": 0.8561360836029053, "num_tokens": 52046634.0, "step": 1371 }, { "epoch": 0.1745325022261799, "ewc_loss": 0.0224609375, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 2.018763542175293, "learning_rate": 5.811784654514625e-06, "loss": 0.5472, "mean_token_accuracy": 0.8300408124923706, "num_tokens": 52079768.0, "step": 1372 }, { "epoch": 0.1746597125047704, "ewc_loss": 0.0233154296875, "ewc_loss_diag": 2.968311309814453e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.152498960494995, "learning_rate": 5.8160237388724035e-06, "loss": 0.4732, "mean_token_accuracy": 0.8506737947463989, "num_tokens": 52118160.0, "step": 1373 }, { "epoch": 0.1747869227833609, "ewc_loss": 0.0224609375, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 1.9550323486328125e-05, "grad_norm": 1.995518445968628, "learning_rate": 5.820262823230182e-06, "loss": 0.5129, "mean_token_accuracy": 0.840765118598938, "num_tokens": 52160219.0, "step": 1374 }, { "epoch": 0.17491413306195142, "ewc_loss": 0.0233154296875, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.169384717941284, "learning_rate": 5.824501907587961e-06, "loss": 0.4695, "mean_token_accuracy": 0.8522417545318604, "num_tokens": 52194746.0, "step": 1375 }, { "epoch": 0.17504134334054192, "ewc_loss": 0.022705078125, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 2.012943744659424, "learning_rate": 5.82874099194574e-06, "loss": 0.5605, "mean_token_accuracy": 0.8232378959655762, "num_tokens": 52232625.0, "step": 1376 }, { "epoch": 0.17516855361913242, "ewc_loss": 0.023193359375, "ewc_loss_diag": 2.956390380859375e-05, "ewc_loss_parallel": 2.0265579223632812e-05, "grad_norm": 2.1615631580352783, "learning_rate": 5.832980076303519e-06, "loss": 0.5597, "mean_token_accuracy": 0.8263150453567505, "num_tokens": 52273030.0, "step": 1377 }, { "epoch": 0.17529576389772294, "ewc_loss": 0.0228271484375, "ewc_loss_diag": 2.9802322387695312e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.006899118423462, "learning_rate": 5.837219160661297e-06, "loss": 0.451, "mean_token_accuracy": 0.8571464419364929, "num_tokens": 52306490.0, "step": 1378 }, { "epoch": 0.17542297417631345, "ewc_loss": 0.023193359375, "ewc_loss_diag": 2.9802322387695312e-05, "ewc_loss_parallel": 2.0265579223632812e-05, "grad_norm": 2.1510543823242188, "learning_rate": 5.841458245019076e-06, "loss": 0.4463, "mean_token_accuracy": 0.8598677515983582, "num_tokens": 52346717.0, "step": 1379 }, { "epoch": 0.17555018445490395, "ewc_loss": 0.022705078125, "ewc_loss_diag": 2.9802322387695312e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 2.0270936489105225, "learning_rate": 5.845697329376855e-06, "loss": 0.495, "mean_token_accuracy": 0.8468868732452393, "num_tokens": 52378165.0, "step": 1380 }, { "epoch": 0.17567739473349447, "ewc_loss": 0.0233154296875, "ewc_loss_diag": 2.9802322387695312e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.157592296600342, "learning_rate": 5.849936413734634e-06, "loss": 0.5593, "mean_token_accuracy": 0.8280501961708069, "num_tokens": 52418719.0, "step": 1381 }, { "epoch": 0.17580460501208497, "ewc_loss": 0.0228271484375, "ewc_loss_diag": 2.9802322387695312e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 1.995666742324829, "learning_rate": 5.854175498092413e-06, "loss": 0.4683, "mean_token_accuracy": 0.854155421257019, "num_tokens": 52459094.0, "step": 1382 }, { "epoch": 0.17593181529067548, "ewc_loss": 0.0234375, "ewc_loss_diag": 2.9921531677246094e-05, "ewc_loss_parallel": 2.0503997802734375e-05, "grad_norm": 2.150564193725586, "learning_rate": 5.8584145824501915e-06, "loss": 0.5045, "mean_token_accuracy": 0.8432486057281494, "num_tokens": 52503117.0, "step": 1383 }, { "epoch": 0.176059025569266, "ewc_loss": 0.02294921875, "ewc_loss_diag": 2.9921531677246094e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.0118861198425293, "learning_rate": 5.86265366680797e-06, "loss": 0.5265, "mean_token_accuracy": 0.8355504870414734, "num_tokens": 52542643.0, "step": 1384 }, { "epoch": 0.1761862358478565, "ewc_loss": 0.023681640625, "ewc_loss_diag": 2.9921531677246094e-05, "ewc_loss_parallel": 2.0623207092285156e-05, "grad_norm": 2.1491992473602295, "learning_rate": 5.866892751165749e-06, "loss": 0.5043, "mean_token_accuracy": 0.8436613082885742, "num_tokens": 52585885.0, "step": 1385 }, { "epoch": 0.176313446126447, "ewc_loss": 0.02294921875, "ewc_loss_diag": 3.0159950256347656e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.0210964679718018, "learning_rate": 5.871131835523527e-06, "loss": 0.5511, "mean_token_accuracy": 0.8295687437057495, "num_tokens": 52621629.0, "step": 1386 }, { "epoch": 0.17644065640503753, "ewc_loss": 0.0233154296875, "ewc_loss_diag": 3.0159950256347656e-05, "ewc_loss_parallel": 2.0265579223632812e-05, "grad_norm": 2.144099235534668, "learning_rate": 5.875370919881306e-06, "loss": 0.5053, "mean_token_accuracy": 0.838748574256897, "num_tokens": 52667056.0, "step": 1387 }, { "epoch": 0.17656786668362803, "ewc_loss": 0.022705078125, "ewc_loss_diag": 3.0159950256347656e-05, "ewc_loss_parallel": 1.9669532775878906e-05, "grad_norm": 2.0259594917297363, "learning_rate": 5.879610004239085e-06, "loss": 0.5612, "mean_token_accuracy": 0.8290172219276428, "num_tokens": 52705857.0, "step": 1388 }, { "epoch": 0.17669507696221853, "ewc_loss": 0.023193359375, "ewc_loss_diag": 3.0159950256347656e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 2.154154062271118, "learning_rate": 5.883849088596864e-06, "loss": 0.5467, "mean_token_accuracy": 0.8299360275268555, "num_tokens": 52748908.0, "step": 1389 }, { "epoch": 0.17682228724080906, "ewc_loss": 0.0228271484375, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 2.0257325172424316, "learning_rate": 5.888088172954642e-06, "loss": 0.5207, "mean_token_accuracy": 0.8408173322677612, "num_tokens": 52783600.0, "step": 1390 }, { "epoch": 0.17694949751939956, "ewc_loss": 0.023193359375, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 2.1418538093566895, "learning_rate": 5.892327257312421e-06, "loss": 0.4448, "mean_token_accuracy": 0.8620931506156921, "num_tokens": 52819873.0, "step": 1391 }, { "epoch": 0.1770767077979901, "ewc_loss": 0.02294921875, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.0094258785247803, "learning_rate": 5.8965663416702e-06, "loss": 0.4468, "mean_token_accuracy": 0.8567726612091064, "num_tokens": 52858883.0, "step": 1392 }, { "epoch": 0.1772039180765806, "ewc_loss": 0.0234375, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.1631176471710205, "learning_rate": 5.900805426027978e-06, "loss": 0.4998, "mean_token_accuracy": 0.8456031680107117, "num_tokens": 52895501.0, "step": 1393 }, { "epoch": 0.1773311283551711, "ewc_loss": 0.02294921875, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.0815742015838623, "learning_rate": 5.905044510385757e-06, "loss": 0.4469, "mean_token_accuracy": 0.8595335483551025, "num_tokens": 52934678.0, "step": 1394 }, { "epoch": 0.17745833863376162, "ewc_loss": 0.0233154296875, "ewc_loss_diag": 3.0279159545898438e-05, "ewc_loss_parallel": 2.0265579223632812e-05, "grad_norm": 2.1332480907440186, "learning_rate": 5.909283594743536e-06, "loss": 0.552, "mean_token_accuracy": 0.8262417316436768, "num_tokens": 52972540.0, "step": 1395 }, { "epoch": 0.17758554891235212, "ewc_loss": 0.02294921875, "ewc_loss_diag": 3.039836883544922e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.0205373764038086, "learning_rate": 5.9135226791013145e-06, "loss": 0.4883, "mean_token_accuracy": 0.8435526490211487, "num_tokens": 53010819.0, "step": 1396 }, { "epoch": 0.17771275919094262, "ewc_loss": 0.0230712890625, "ewc_loss_diag": 3.039836883544922e-05, "ewc_loss_parallel": 2.002716064453125e-05, "grad_norm": 2.132535219192505, "learning_rate": 5.917761763459093e-06, "loss": 0.5513, "mean_token_accuracy": 0.8286332488059998, "num_tokens": 53050142.0, "step": 1397 }, { "epoch": 0.17783996946953315, "ewc_loss": 0.0228271484375, "ewc_loss_diag": 3.039836883544922e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 2.03643536567688, "learning_rate": 5.922000847816872e-06, "loss": 0.4803, "mean_token_accuracy": 0.8517681360244751, "num_tokens": 53087306.0, "step": 1398 }, { "epoch": 0.17796717974812365, "ewc_loss": 0.0230712890625, "ewc_loss_diag": 3.039836883544922e-05, "ewc_loss_parallel": 2.002716064453125e-05, "grad_norm": 2.1288437843322754, "learning_rate": 5.926239932174651e-06, "loss": 0.5385, "mean_token_accuracy": 0.8322631120681763, "num_tokens": 53131850.0, "step": 1399 }, { "epoch": 0.17809439002671415, "ewc_loss": 0.0228271484375, "ewc_loss_diag": 3.039836883544922e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 2.0214223861694336, "learning_rate": 5.93047901653243e-06, "loss": 0.4922, "mean_token_accuracy": 0.8440941572189331, "num_tokens": 53168902.0, "step": 1400 }, { "epoch": 0.17822160030530468, "ewc_loss": 0.0234375, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.138279914855957, "learning_rate": 5.934718100890208e-06, "loss": 0.47, "mean_token_accuracy": 0.8514183163642883, "num_tokens": 53208497.0, "step": 1401 }, { "epoch": 0.17834881058389518, "ewc_loss": 0.02294921875, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.0351264476776123, "learning_rate": 5.938957185247987e-06, "loss": 0.5235, "mean_token_accuracy": 0.8348922729492188, "num_tokens": 53245618.0, "step": 1402 }, { "epoch": 0.17847602086248568, "ewc_loss": 0.0234375, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.1592893600463867, "learning_rate": 5.943196269605766e-06, "loss": 0.4856, "mean_token_accuracy": 0.8463709354400635, "num_tokens": 53280315.0, "step": 1403 }, { "epoch": 0.1786032311410762, "ewc_loss": 0.02294921875, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 1.990795135498047e-05, "grad_norm": 2.03798508644104, "learning_rate": 5.947435353963545e-06, "loss": 0.524, "mean_token_accuracy": 0.8390506505966187, "num_tokens": 53313291.0, "step": 1404 }, { "epoch": 0.1787304414196667, "ewc_loss": 0.0235595703125, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 2.0503997802734375e-05, "grad_norm": 2.1541264057159424, "learning_rate": 5.951674438321324e-06, "loss": 0.5105, "mean_token_accuracy": 0.8421649932861328, "num_tokens": 53347764.0, "step": 1405 }, { "epoch": 0.1788576516982572, "ewc_loss": 0.0230712890625, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 2.002716064453125e-05, "grad_norm": 2.029672145843506, "learning_rate": 5.9559135226791025e-06, "loss": 0.51, "mean_token_accuracy": 0.8377708792686462, "num_tokens": 53385184.0, "step": 1406 }, { "epoch": 0.17898486197684774, "ewc_loss": 0.023681640625, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 2.0623207092285156e-05, "grad_norm": 2.166577100753784, "learning_rate": 5.9601526070368805e-06, "loss": 0.5845, "mean_token_accuracy": 0.8168587684631348, "num_tokens": 53425850.0, "step": 1407 }, { "epoch": 0.17911207225543824, "ewc_loss": 0.0228271484375, "ewc_loss_diag": 3.0517578125e-05, "ewc_loss_parallel": 1.9788742065429688e-05, "grad_norm": 1.9983898401260376, "learning_rate": 5.964391691394659e-06, "loss": 0.5177, "mean_token_accuracy": 0.840107798576355, "num_tokens": 53472444.0, "step": 1408 }, { "epoch": 0.17923928253402874, "ewc_loss": 0.02392578125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.086162567138672e-05, "grad_norm": 2.2064719200134277, "learning_rate": 5.968630775752437e-06, "loss": 0.5564, "mean_token_accuracy": 0.8290106058120728, "num_tokens": 53506996.0, "step": 1409 }, { "epoch": 0.17936649281261927, "ewc_loss": 0.0230712890625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.002716064453125e-05, "grad_norm": 2.0240185260772705, "learning_rate": 5.972869860110216e-06, "loss": 0.5298, "mean_token_accuracy": 0.8328846096992493, "num_tokens": 53542005.0, "step": 1410 }, { "epoch": 0.17949370309120977, "ewc_loss": 0.0245361328125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.1457672119140625e-05, "grad_norm": 2.217346668243408, "learning_rate": 5.977108944467995e-06, "loss": 0.5683, "mean_token_accuracy": 0.8300877809524536, "num_tokens": 53582130.0, "step": 1411 }, { "epoch": 0.17962091336980027, "ewc_loss": 0.023193359375, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.014636993408203e-05, "grad_norm": 1.9961596727371216, "learning_rate": 5.981348028825774e-06, "loss": 0.4853, "mean_token_accuracy": 0.8523786067962646, "num_tokens": 53624528.0, "step": 1412 }, { "epoch": 0.1797481236483908, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.255453586578369, "learning_rate": 5.985587113183553e-06, "loss": 0.5347, "mean_token_accuracy": 0.835761547088623, "num_tokens": 53663701.0, "step": 1413 }, { "epoch": 0.1798753339269813, "ewc_loss": 0.0234375, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.029101610183716, "learning_rate": 5.989826197541332e-06, "loss": 0.5222, "mean_token_accuracy": 0.8405789136886597, "num_tokens": 53700655.0, "step": 1414 }, { "epoch": 0.18000254420557182, "ewc_loss": 0.0252685546875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2172927856445312e-05, "grad_norm": 2.2480854988098145, "learning_rate": 5.994065281899111e-06, "loss": 0.5449, "mean_token_accuracy": 0.8350675702095032, "num_tokens": 53747387.0, "step": 1415 }, { "epoch": 0.18012975448416232, "ewc_loss": 0.0235595703125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.0503997802734375e-05, "grad_norm": 2.0098628997802734, "learning_rate": 5.998304366256889e-06, "loss": 0.511, "mean_token_accuracy": 0.8413547277450562, "num_tokens": 53789416.0, "step": 1416 }, { "epoch": 0.18025696476275282, "ewc_loss": 0.0255126953125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2411346435546875e-05, "grad_norm": 2.273516893386841, "learning_rate": 6.002543450614668e-06, "loss": 0.4377, "mean_token_accuracy": 0.86672043800354, "num_tokens": 53824300.0, "step": 1417 }, { "epoch": 0.18038417504134335, "ewc_loss": 0.0235595703125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.0503997802734375e-05, "grad_norm": 2.043551445007324, "learning_rate": 6.0067825349724465e-06, "loss": 0.5106, "mean_token_accuracy": 0.8420634865760803, "num_tokens": 53861366.0, "step": 1418 }, { "epoch": 0.18051138531993385, "ewc_loss": 0.025634765625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2530555725097656e-05, "grad_norm": 2.291078567504883, "learning_rate": 6.011021619330225e-06, "loss": 0.5362, "mean_token_accuracy": 0.8354690074920654, "num_tokens": 53899655.0, "step": 1419 }, { "epoch": 0.18063859559852435, "ewc_loss": 0.02392578125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.086162567138672e-05, "grad_norm": 2.0585174560546875, "learning_rate": 6.015260703688004e-06, "loss": 0.5127, "mean_token_accuracy": 0.8375749588012695, "num_tokens": 53935563.0, "step": 1420 }, { "epoch": 0.18076580587711488, "ewc_loss": 0.0255126953125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2411346435546875e-05, "grad_norm": 2.291182518005371, "learning_rate": 6.019499788045783e-06, "loss": 0.513, "mean_token_accuracy": 0.8416054248809814, "num_tokens": 53970490.0, "step": 1421 }, { "epoch": 0.18089301615570538, "ewc_loss": 0.0240478515625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.09808349609375e-05, "grad_norm": 2.073827028274536, "learning_rate": 6.023738872403562e-06, "loss": 0.477, "mean_token_accuracy": 0.8492918014526367, "num_tokens": 54000460.0, "step": 1422 }, { "epoch": 0.18102022643429588, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2292137145996094e-05, "grad_norm": 2.274125337600708, "learning_rate": 6.027977956761341e-06, "loss": 0.5047, "mean_token_accuracy": 0.8482643365859985, "num_tokens": 54038007.0, "step": 1423 }, { "epoch": 0.1811474367128864, "ewc_loss": 0.0240478515625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.09808349609375e-05, "grad_norm": 2.0648746490478516, "learning_rate": 6.032217041119118e-06, "loss": 0.5141, "mean_token_accuracy": 0.8365116715431213, "num_tokens": 54074515.0, "step": 1424 }, { "epoch": 0.1812746469914769, "ewc_loss": 0.0252685546875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2172927856445312e-05, "grad_norm": 2.278289318084717, "learning_rate": 6.036456125476897e-06, "loss": 0.5506, "mean_token_accuracy": 0.8299583196640015, "num_tokens": 54109355.0, "step": 1425 }, { "epoch": 0.1814018572700674, "ewc_loss": 0.024169921875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.110004425048828e-05, "grad_norm": 2.058753490447998, "learning_rate": 6.040695209834676e-06, "loss": 0.45, "mean_token_accuracy": 0.8562173247337341, "num_tokens": 54150966.0, "step": 1426 }, { "epoch": 0.18152906754865794, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2292137145996094e-05, "grad_norm": 2.291120767593384, "learning_rate": 6.044934294192455e-06, "loss": 0.5456, "mean_token_accuracy": 0.8306008577346802, "num_tokens": 54188398.0, "step": 1427 }, { "epoch": 0.18165627782724844, "ewc_loss": 0.024169921875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.110004425048828e-05, "grad_norm": 2.054537296295166, "learning_rate": 6.049173378550234e-06, "loss": 0.4998, "mean_token_accuracy": 0.8438212871551514, "num_tokens": 54233669.0, "step": 1428 }, { "epoch": 0.18178348810583894, "ewc_loss": 0.0255126953125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2411346435546875e-05, "grad_norm": 2.288463830947876, "learning_rate": 6.0534124629080126e-06, "loss": 0.4837, "mean_token_accuracy": 0.8494654893875122, "num_tokens": 54272377.0, "step": 1429 }, { "epoch": 0.18191069838442947, "ewc_loss": 0.0242919921875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.1219253540039062e-05, "grad_norm": 2.076622247695923, "learning_rate": 6.0576515472657915e-06, "loss": 0.552, "mean_token_accuracy": 0.8317487239837646, "num_tokens": 54313831.0, "step": 1430 }, { "epoch": 0.18203790866301997, "ewc_loss": 0.0255126953125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2411346435546875e-05, "grad_norm": 2.2670252323150635, "learning_rate": 6.0618906316235695e-06, "loss": 0.4583, "mean_token_accuracy": 0.8553481101989746, "num_tokens": 54356724.0, "step": 1431 }, { "epoch": 0.18216511894161047, "ewc_loss": 0.024169921875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.110004425048828e-05, "grad_norm": 2.070610761642456, "learning_rate": 6.066129715981348e-06, "loss": 0.4911, "mean_token_accuracy": 0.8460094928741455, "num_tokens": 54398407.0, "step": 1432 }, { "epoch": 0.182292329220201, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.2292137145996094e-05, "grad_norm": 2.2744462490081787, "learning_rate": 6.070368800339127e-06, "loss": 0.4506, "mean_token_accuracy": 0.8610601425170898, "num_tokens": 54430023.0, "step": 1433 }, { "epoch": 0.1824195394987915, "ewc_loss": 0.024169921875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.110004425048828e-05, "grad_norm": 2.086636543273926, "learning_rate": 6.074607884696906e-06, "loss": 0.4806, "mean_token_accuracy": 0.848090410232544, "num_tokens": 54467154.0, "step": 1434 }, { "epoch": 0.182546749777382, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.274519681930542, "learning_rate": 6.078846969054685e-06, "loss": 0.5112, "mean_token_accuracy": 0.842680811882019, "num_tokens": 54501139.0, "step": 1435 }, { "epoch": 0.18267396005597253, "ewc_loss": 0.024169921875, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.110004425048828e-05, "grad_norm": 2.0804009437561035, "learning_rate": 6.083086053412464e-06, "loss": 0.4911, "mean_token_accuracy": 0.8477599620819092, "num_tokens": 54541084.0, "step": 1436 }, { "epoch": 0.18280117033456303, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.261181116104126, "learning_rate": 6.087325137770243e-06, "loss": 0.473, "mean_token_accuracy": 0.8528319001197815, "num_tokens": 54576745.0, "step": 1437 }, { "epoch": 0.18292838061315353, "ewc_loss": 0.02392578125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.086162567138672e-05, "grad_norm": 2.0815958976745605, "learning_rate": 6.091564222128022e-06, "loss": 0.5499, "mean_token_accuracy": 0.8338112235069275, "num_tokens": 54619039.0, "step": 1438 }, { "epoch": 0.18305559089174406, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.244192361831665, "learning_rate": 6.095803306485799e-06, "loss": 0.4928, "mean_token_accuracy": 0.8451873064041138, "num_tokens": 54660009.0, "step": 1439 }, { "epoch": 0.18318280117033456, "ewc_loss": 0.02392578125, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.086162567138672e-05, "grad_norm": 2.074846029281616, "learning_rate": 6.100042390843578e-06, "loss": 0.5253, "mean_token_accuracy": 0.8381652235984802, "num_tokens": 54701253.0, "step": 1440 }, { "epoch": 0.1833100114489251, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.246932029724121, "learning_rate": 6.104281475201357e-06, "loss": 0.4808, "mean_token_accuracy": 0.8497534394264221, "num_tokens": 54738406.0, "step": 1441 }, { "epoch": 0.1834372217275156, "ewc_loss": 0.023681640625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.0623207092285156e-05, "grad_norm": 2.1064093112945557, "learning_rate": 6.1085205595591355e-06, "loss": 0.5327, "mean_token_accuracy": 0.83543860912323, "num_tokens": 54771808.0, "step": 1442 }, { "epoch": 0.1835644320061061, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.2228920459747314, "learning_rate": 6.112759643916914e-06, "loss": 0.4329, "mean_token_accuracy": 0.8639798164367676, "num_tokens": 54807337.0, "step": 1443 }, { "epoch": 0.18369164228469662, "ewc_loss": 0.023681640625, "ewc_loss_diag": 3.075599670410156e-05, "ewc_loss_parallel": 2.0623207092285156e-05, "grad_norm": 2.101574182510376, "learning_rate": 6.116998728274693e-06, "loss": 0.518, "mean_token_accuracy": 0.8407729864120483, "num_tokens": 54844932.0, "step": 1444 }, { "epoch": 0.18381885256328712, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.207092523574829, "learning_rate": 6.121237812632472e-06, "loss": 0.5221, "mean_token_accuracy": 0.8349883556365967, "num_tokens": 54884492.0, "step": 1445 }, { "epoch": 0.18394606284187762, "ewc_loss": 0.0238037109375, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.0623207092285156e-05, "grad_norm": 2.079221725463867, "learning_rate": 6.12547689699025e-06, "loss": 0.4714, "mean_token_accuracy": 0.8564850091934204, "num_tokens": 54924164.0, "step": 1446 }, { "epoch": 0.18407327312046814, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.209364891052246, "learning_rate": 6.129715981348029e-06, "loss": 0.5041, "mean_token_accuracy": 0.8469940423965454, "num_tokens": 54961676.0, "step": 1447 }, { "epoch": 0.18420048339905865, "ewc_loss": 0.02392578125, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.0742416381835938e-05, "grad_norm": 2.0851597785949707, "learning_rate": 6.133955065705808e-06, "loss": 0.5285, "mean_token_accuracy": 0.833773672580719, "num_tokens": 55000889.0, "step": 1448 }, { "epoch": 0.18432769367764915, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.2057251930236816, "learning_rate": 6.138194150063587e-06, "loss": 0.4947, "mean_token_accuracy": 0.8474034070968628, "num_tokens": 55039471.0, "step": 1449 }, { "epoch": 0.18445490395623967, "ewc_loss": 0.0235595703125, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.071819543838501, "learning_rate": 6.142433234421366e-06, "loss": 0.5269, "mean_token_accuracy": 0.8340802192687988, "num_tokens": 55078408.0, "step": 1450 }, { "epoch": 0.18458211423483017, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.224419355392456, "learning_rate": 6.146672318779145e-06, "loss": 0.5295, "mean_token_accuracy": 0.835878312587738, "num_tokens": 55116068.0, "step": 1451 }, { "epoch": 0.18470932451342068, "ewc_loss": 0.0235595703125, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.0384788513183594e-05, "grad_norm": 2.0912296772003174, "learning_rate": 6.1509114031369235e-06, "loss": 0.5146, "mean_token_accuracy": 0.8431380987167358, "num_tokens": 55149933.0, "step": 1452 }, { "epoch": 0.1848365347920112, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.21872878074646, "learning_rate": 6.155150487494702e-06, "loss": 0.4454, "mean_token_accuracy": 0.860839307308197, "num_tokens": 55184674.0, "step": 1453 }, { "epoch": 0.1849637450706017, "ewc_loss": 0.023681640625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.0503997802734375e-05, "grad_norm": 2.0555806159973145, "learning_rate": 6.15938957185248e-06, "loss": 0.4782, "mean_token_accuracy": 0.8507328629493713, "num_tokens": 55227918.0, "step": 1454 }, { "epoch": 0.1850909553491922, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.2172927856445312e-05, "grad_norm": 2.2230818271636963, "learning_rate": 6.1636286562102585e-06, "loss": 0.4898, "mean_token_accuracy": 0.847848653793335, "num_tokens": 55274312.0, "step": 1455 }, { "epoch": 0.18521816562778273, "ewc_loss": 0.02392578125, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.0742416381835938e-05, "grad_norm": 2.07938551902771, "learning_rate": 6.167867740568037e-06, "loss": 0.5561, "mean_token_accuracy": 0.826935887336731, "num_tokens": 55310583.0, "step": 1456 }, { "epoch": 0.18534537590637323, "ewc_loss": 0.0255126953125, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.2292137145996094e-05, "grad_norm": 2.2332491874694824, "learning_rate": 6.172106824925816e-06, "loss": 0.4862, "mean_token_accuracy": 0.846149206161499, "num_tokens": 55352365.0, "step": 1457 }, { "epoch": 0.18547258618496373, "ewc_loss": 0.0238037109375, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.0623207092285156e-05, "grad_norm": 2.0688962936401367, "learning_rate": 6.176345909283595e-06, "loss": 0.4998, "mean_token_accuracy": 0.8430348038673401, "num_tokens": 55391268.0, "step": 1458 }, { "epoch": 0.18559979646355426, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.2172927856445312e-05, "grad_norm": 2.2514688968658447, "learning_rate": 6.180584993641374e-06, "loss": 0.4797, "mean_token_accuracy": 0.8500507473945618, "num_tokens": 55428842.0, "step": 1459 }, { "epoch": 0.18572700674214476, "ewc_loss": 0.023681640625, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.0503997802734375e-05, "grad_norm": 2.0710432529449463, "learning_rate": 6.184824077999153e-06, "loss": 0.4587, "mean_token_accuracy": 0.855787992477417, "num_tokens": 55464341.0, "step": 1460 }, { "epoch": 0.18585421702073526, "ewc_loss": 0.0255126953125, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.2292137145996094e-05, "grad_norm": 2.287105083465576, "learning_rate": 6.189063162356932e-06, "loss": 0.5563, "mean_token_accuracy": 0.8270854949951172, "num_tokens": 55498985.0, "step": 1461 }, { "epoch": 0.1859814272993258, "ewc_loss": 0.02392578125, "ewc_loss_diag": 3.123283386230469e-05, "ewc_loss_parallel": 2.0742416381835938e-05, "grad_norm": 2.057574987411499, "learning_rate": 6.19330224671471e-06, "loss": 0.4828, "mean_token_accuracy": 0.8480968475341797, "num_tokens": 55537762.0, "step": 1462 }, { "epoch": 0.1861086375779163, "ewc_loss": 0.02587890625, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.2649765014648438e-05, "grad_norm": 2.314716339111328, "learning_rate": 6.197541331072489e-06, "loss": 0.5295, "mean_token_accuracy": 0.8350274562835693, "num_tokens": 55572571.0, "step": 1463 }, { "epoch": 0.1862358478565068, "ewc_loss": 0.0244140625, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.1219253540039062e-05, "grad_norm": 2.0543088912963867, "learning_rate": 6.201780415430268e-06, "loss": 0.4922, "mean_token_accuracy": 0.8489307165145874, "num_tokens": 55611858.0, "step": 1464 }, { "epoch": 0.18636305813509732, "ewc_loss": 0.0264892578125, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.337512254714966, "learning_rate": 6.2060194997880465e-06, "loss": 0.5331, "mean_token_accuracy": 0.8357667326927185, "num_tokens": 55649168.0, "step": 1465 }, { "epoch": 0.18649026841368782, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.0649993419647217, "learning_rate": 6.210258584145825e-06, "loss": 0.4717, "mean_token_accuracy": 0.8502360582351685, "num_tokens": 55686127.0, "step": 1466 }, { "epoch": 0.18661747869227835, "ewc_loss": 0.0269775390625, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.365090847015381, "learning_rate": 6.214497668503604e-06, "loss": 0.5315, "mean_token_accuracy": 0.8335297107696533, "num_tokens": 55724062.0, "step": 1467 }, { "epoch": 0.18674468897086885, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.147125244140625e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.1099398136138916, "learning_rate": 6.218736752861383e-06, "loss": 0.4867, "mean_token_accuracy": 0.8488485813140869, "num_tokens": 55752137.0, "step": 1468 }, { "epoch": 0.18687189924945935, "ewc_loss": 0.0274658203125, "ewc_loss_diag": 3.170967102050781e-05, "ewc_loss_parallel": 2.4318695068359375e-05, "grad_norm": 2.357975721359253, "learning_rate": 6.22297583721916e-06, "loss": 0.5266, "mean_token_accuracy": 0.8443860411643982, "num_tokens": 55793171.0, "step": 1469 }, { "epoch": 0.18699910952804988, "ewc_loss": 0.025146484375, "ewc_loss_diag": 3.170967102050781e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.1097607612609863, "learning_rate": 6.227214921576939e-06, "loss": 0.5514, "mean_token_accuracy": 0.8349721431732178, "num_tokens": 55831118.0, "step": 1470 }, { "epoch": 0.18712631980664038, "ewc_loss": 0.0272216796875, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.348240852355957, "learning_rate": 6.231454005934718e-06, "loss": 0.45, "mean_token_accuracy": 0.8605268597602844, "num_tokens": 55865568.0, "step": 1471 }, { "epoch": 0.18725353008523088, "ewc_loss": 0.025146484375, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.1174089908599854, "learning_rate": 6.235693090292497e-06, "loss": 0.5649, "mean_token_accuracy": 0.830661416053772, "num_tokens": 55904624.0, "step": 1472 }, { "epoch": 0.1873807403638214, "ewc_loss": 0.0272216796875, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.3250627517700195, "learning_rate": 6.239932174650276e-06, "loss": 0.4478, "mean_token_accuracy": 0.8613288402557373, "num_tokens": 55944714.0, "step": 1473 }, { "epoch": 0.1875079506424119, "ewc_loss": 0.025146484375, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.1385772228240967, "learning_rate": 6.244171259008055e-06, "loss": 0.5302, "mean_token_accuracy": 0.8327070474624634, "num_tokens": 55979444.0, "step": 1474 }, { "epoch": 0.1876351609210024, "ewc_loss": 0.0267333984375, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 2.306697368621826, "learning_rate": 6.248410343365834e-06, "loss": 0.4685, "mean_token_accuracy": 0.8550898432731628, "num_tokens": 56017171.0, "step": 1475 }, { "epoch": 0.18776237119959294, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.1191701889038086, "learning_rate": 6.2526494277236125e-06, "loss": 0.4844, "mean_token_accuracy": 0.8514900207519531, "num_tokens": 56060562.0, "step": 1476 }, { "epoch": 0.18788958147818344, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.3008952140808105, "learning_rate": 6.2568885120813906e-06, "loss": 0.4832, "mean_token_accuracy": 0.8493474721908569, "num_tokens": 56098366.0, "step": 1477 }, { "epoch": 0.18801679175677394, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.1136016845703125, "learning_rate": 6.2611275964391694e-06, "loss": 0.5422, "mean_token_accuracy": 0.8296784162521362, "num_tokens": 56143323.0, "step": 1478 }, { "epoch": 0.18814400203536447, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.2921459674835205, "learning_rate": 6.265366680796948e-06, "loss": 0.4411, "mean_token_accuracy": 0.8608164191246033, "num_tokens": 56183846.0, "step": 1479 }, { "epoch": 0.18827121231395497, "ewc_loss": 0.0247802734375, "ewc_loss_diag": 3.1948089599609375e-05, "ewc_loss_parallel": 2.1576881408691406e-05, "grad_norm": 2.1066384315490723, "learning_rate": 6.269605765154727e-06, "loss": 0.4983, "mean_token_accuracy": 0.8450321555137634, "num_tokens": 56227412.0, "step": 1480 }, { "epoch": 0.18839842259254547, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.299158811569214, "learning_rate": 6.273844849512506e-06, "loss": 0.4982, "mean_token_accuracy": 0.8454793095588684, "num_tokens": 56267649.0, "step": 1481 }, { "epoch": 0.188525632871136, "ewc_loss": 0.0247802734375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1576881408691406e-05, "grad_norm": 2.09240984916687, "learning_rate": 6.278083933870285e-06, "loss": 0.4251, "mean_token_accuracy": 0.8674886226654053, "num_tokens": 56310802.0, "step": 1482 }, { "epoch": 0.1886528431497265, "ewc_loss": 0.0264892578125, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.3257508277893066, "learning_rate": 6.282323018228064e-06, "loss": 0.467, "mean_token_accuracy": 0.8575607538223267, "num_tokens": 56345972.0, "step": 1483 }, { "epoch": 0.188780053428317, "ewc_loss": 0.0247802734375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1576881408691406e-05, "grad_norm": 2.099496364593506, "learning_rate": 6.286562102585841e-06, "loss": 0.5484, "mean_token_accuracy": 0.8339653015136719, "num_tokens": 56391778.0, "step": 1484 }, { "epoch": 0.18890726370690752, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.325120449066162, "learning_rate": 6.29080118694362e-06, "loss": 0.4335, "mean_token_accuracy": 0.8625361919403076, "num_tokens": 56426306.0, "step": 1485 }, { "epoch": 0.18903447398549802, "ewc_loss": 0.0247802734375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1576881408691406e-05, "grad_norm": 2.0987777709960938, "learning_rate": 6.295040271301399e-06, "loss": 0.4372, "mean_token_accuracy": 0.8622990846633911, "num_tokens": 56461974.0, "step": 1486 }, { "epoch": 0.18916168426408853, "ewc_loss": 0.0267333984375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 2.323667526245117, "learning_rate": 6.299279355659178e-06, "loss": 0.4861, "mean_token_accuracy": 0.8509048223495483, "num_tokens": 56507205.0, "step": 1487 }, { "epoch": 0.18928889454267905, "ewc_loss": 0.024658203125, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1457672119140625e-05, "grad_norm": 2.117539882659912, "learning_rate": 6.303518440016957e-06, "loss": 0.495, "mean_token_accuracy": 0.8427411317825317, "num_tokens": 56542076.0, "step": 1488 }, { "epoch": 0.18941610482126955, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 2.343256950378418, "learning_rate": 6.3077575243747355e-06, "loss": 0.4919, "mean_token_accuracy": 0.8517066240310669, "num_tokens": 56575980.0, "step": 1489 }, { "epoch": 0.18954331509986008, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.116096258163452, "learning_rate": 6.311996608732514e-06, "loss": 0.5222, "mean_token_accuracy": 0.8421658277511597, "num_tokens": 56613410.0, "step": 1490 }, { "epoch": 0.18967052537845058, "ewc_loss": 0.0269775390625, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.3466670513153076, "learning_rate": 6.316235693090293e-06, "loss": 0.5159, "mean_token_accuracy": 0.838103175163269, "num_tokens": 56649105.0, "step": 1491 }, { "epoch": 0.18979773565704108, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.1293249130249023, "learning_rate": 6.320474777448071e-06, "loss": 0.5172, "mean_token_accuracy": 0.8370205163955688, "num_tokens": 56686049.0, "step": 1492 }, { "epoch": 0.1899249459356316, "ewc_loss": 0.0267333984375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 2.3520290851593018, "learning_rate": 6.32471386180585e-06, "loss": 0.4822, "mean_token_accuracy": 0.8483353853225708, "num_tokens": 56725659.0, "step": 1493 }, { "epoch": 0.1900521562142221, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.159233570098877, "learning_rate": 6.328952946163629e-06, "loss": 0.57, "mean_token_accuracy": 0.8272123336791992, "num_tokens": 56760612.0, "step": 1494 }, { "epoch": 0.1901793664928126, "ewc_loss": 0.0267333984375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 2.3117775917053223, "learning_rate": 6.333192030521408e-06, "loss": 0.4478, "mean_token_accuracy": 0.8589745759963989, "num_tokens": 56796081.0, "step": 1495 }, { "epoch": 0.19030657677140314, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.1307482719421387, "learning_rate": 6.337431114879187e-06, "loss": 0.5012, "mean_token_accuracy": 0.8456240296363831, "num_tokens": 56833477.0, "step": 1496 }, { "epoch": 0.19043378704999364, "ewc_loss": 0.0263671875, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.3245811462402344e-05, "grad_norm": 2.310285806655884, "learning_rate": 6.341670199236966e-06, "loss": 0.4818, "mean_token_accuracy": 0.8542121648788452, "num_tokens": 56870898.0, "step": 1497 }, { "epoch": 0.19056099732858414, "ewc_loss": 0.0247802734375, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.1576881408691406e-05, "grad_norm": 2.1067118644714355, "learning_rate": 6.345909283594745e-06, "loss": 0.4497, "mean_token_accuracy": 0.8585980534553528, "num_tokens": 56916980.0, "step": 1498 }, { "epoch": 0.19068820760717467, "ewc_loss": 0.0263671875, "ewc_loss_diag": 3.218650817871094e-05, "ewc_loss_parallel": 2.3245811462402344e-05, "grad_norm": 2.3258838653564453, "learning_rate": 6.3501483679525235e-06, "loss": 0.5013, "mean_token_accuracy": 0.8396015167236328, "num_tokens": 56952564.0, "step": 1499 }, { "epoch": 0.19081541788576517, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.24249267578125e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.1290929317474365, "learning_rate": 6.354387452310301e-06, "loss": 0.4619, "mean_token_accuracy": 0.8531335592269897, "num_tokens": 56987381.0, "step": 1500 }, { "epoch": 0.19094262816435567, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.333678960800171, "learning_rate": 6.3586265366680796e-06, "loss": 0.4684, "mean_token_accuracy": 0.8537403345108032, "num_tokens": 57023402.0, "step": 1501 }, { "epoch": 0.1910698384429462, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.1025516986846924, "learning_rate": 6.3628656210258584e-06, "loss": 0.4589, "mean_token_accuracy": 0.8576639890670776, "num_tokens": 57066464.0, "step": 1502 }, { "epoch": 0.1911970487215367, "ewc_loss": 0.0272216796875, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.396106719970703e-05, "grad_norm": 2.343616247177124, "learning_rate": 6.367104705383637e-06, "loss": 0.4774, "mean_token_accuracy": 0.8505622744560242, "num_tokens": 57104495.0, "step": 1503 }, { "epoch": 0.1913242590001272, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.1234934329986572, "learning_rate": 6.371343789741416e-06, "loss": 0.4944, "mean_token_accuracy": 0.8453826904296875, "num_tokens": 57146246.0, "step": 1504 }, { "epoch": 0.19145146927871773, "ewc_loss": 0.02734375, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.348071336746216, "learning_rate": 6.375582874099195e-06, "loss": 0.488, "mean_token_accuracy": 0.8469803333282471, "num_tokens": 57184861.0, "step": 1505 }, { "epoch": 0.19157867955730823, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.154707193374634, "learning_rate": 6.379821958456974e-06, "loss": 0.5416, "mean_token_accuracy": 0.8310462236404419, "num_tokens": 57219395.0, "step": 1506 }, { "epoch": 0.19170588983589873, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.335360288619995, "learning_rate": 6.384061042814752e-06, "loss": 0.5122, "mean_token_accuracy": 0.8424406051635742, "num_tokens": 57255816.0, "step": 1507 }, { "epoch": 0.19183310011448926, "ewc_loss": 0.025146484375, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.135763168334961, "learning_rate": 6.388300127172531e-06, "loss": 0.502, "mean_token_accuracy": 0.8448313474655151, "num_tokens": 57296347.0, "step": 1508 }, { "epoch": 0.19196031039307976, "ewc_loss": 0.0272216796875, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.396106719970703e-05, "grad_norm": 2.3193936347961426, "learning_rate": 6.39253921153031e-06, "loss": 0.5048, "mean_token_accuracy": 0.8416183590888977, "num_tokens": 57341789.0, "step": 1509 }, { "epoch": 0.19208752067167026, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.1327481269836426, "learning_rate": 6.396778295888089e-06, "loss": 0.4712, "mean_token_accuracy": 0.8553236722946167, "num_tokens": 57379399.0, "step": 1510 }, { "epoch": 0.1922147309502608, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 2.335329532623291, "learning_rate": 6.4010173802458676e-06, "loss": 0.51, "mean_token_accuracy": 0.8413857221603394, "num_tokens": 57418936.0, "step": 1511 }, { "epoch": 0.1923419412288513, "ewc_loss": 0.0250244140625, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.1696090698242188e-05, "grad_norm": 2.1196227073669434, "learning_rate": 6.4052564646036464e-06, "loss": 0.4934, "mean_token_accuracy": 0.8452178239822388, "num_tokens": 57461772.0, "step": 1512 }, { "epoch": 0.1924691515074418, "ewc_loss": 0.02734375, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.342376947402954, "learning_rate": 6.409495548961425e-06, "loss": 0.4735, "mean_token_accuracy": 0.8520296216011047, "num_tokens": 57498388.0, "step": 1513 }, { "epoch": 0.19259636178603232, "ewc_loss": 0.02490234375, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.1576881408691406e-05, "grad_norm": 2.1282663345336914, "learning_rate": 6.413734633319204e-06, "loss": 0.4796, "mean_token_accuracy": 0.8496549129486084, "num_tokens": 57534123.0, "step": 1514 }, { "epoch": 0.19272357206462282, "ewc_loss": 0.0274658203125, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.4199485778808594e-05, "grad_norm": 2.3542096614837646, "learning_rate": 6.417973717676982e-06, "loss": 0.4585, "mean_token_accuracy": 0.8582430481910706, "num_tokens": 57568351.0, "step": 1515 }, { "epoch": 0.19285078234321335, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.205371856689453e-05, "grad_norm": 2.112133026123047, "learning_rate": 6.422212802034761e-06, "loss": 0.5282, "mean_token_accuracy": 0.8343706130981445, "num_tokens": 57617842.0, "step": 1516 }, { "epoch": 0.19297799262180385, "ewc_loss": 0.027587890625, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.4318695068359375e-05, "grad_norm": 2.373356342315674, "learning_rate": 6.426451886392539e-06, "loss": 0.5479, "mean_token_accuracy": 0.8354145884513855, "num_tokens": 57657658.0, "step": 1517 }, { "epoch": 0.19310520290039435, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.266334533691406e-05, "ewc_loss_parallel": 2.205371856689453e-05, "grad_norm": 2.124025583267212, "learning_rate": 6.430690970750318e-06, "loss": 0.4704, "mean_token_accuracy": 0.8523989319801331, "num_tokens": 57695964.0, "step": 1518 }, { "epoch": 0.19323241317898487, "ewc_loss": 0.02783203125, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.4557113647460938e-05, "grad_norm": 2.413546085357666, "learning_rate": 6.434930055108097e-06, "loss": 0.5222, "mean_token_accuracy": 0.8399870991706848, "num_tokens": 57725722.0, "step": 1519 }, { "epoch": 0.19335962345757537, "ewc_loss": 0.0260009765625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.2649765014648438e-05, "grad_norm": 2.148036479949951, "learning_rate": 6.439169139465876e-06, "loss": 0.5323, "mean_token_accuracy": 0.8338438272476196, "num_tokens": 57761944.0, "step": 1520 }, { "epoch": 0.19348683373616588, "ewc_loss": 0.0279541015625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.467632293701172e-05, "grad_norm": 2.4029312133789062, "learning_rate": 6.443408223823655e-06, "loss": 0.4931, "mean_token_accuracy": 0.8453001976013184, "num_tokens": 57796853.0, "step": 1521 }, { "epoch": 0.1936140440147564, "ewc_loss": 0.0260009765625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.2649765014648438e-05, "grad_norm": 2.1449289321899414, "learning_rate": 6.447647308181433e-06, "loss": 0.4934, "mean_token_accuracy": 0.8447316288948059, "num_tokens": 57835295.0, "step": 1522 }, { "epoch": 0.1937412542933469, "ewc_loss": 0.0279541015625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.467632293701172e-05, "grad_norm": 2.3900644779205322, "learning_rate": 6.451886392539212e-06, "loss": 0.4718, "mean_token_accuracy": 0.8528844118118286, "num_tokens": 57876294.0, "step": 1523 }, { "epoch": 0.1938684645719374, "ewc_loss": 0.0262451171875, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.288818359375e-05, "grad_norm": 2.154268741607666, "learning_rate": 6.4561254768969905e-06, "loss": 0.4999, "mean_token_accuracy": 0.8474951982498169, "num_tokens": 57915332.0, "step": 1524 }, { "epoch": 0.19399567485052793, "ewc_loss": 0.0279541015625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.467632293701172e-05, "grad_norm": 2.3915977478027344, "learning_rate": 6.460364561254769e-06, "loss": 0.4581, "mean_token_accuracy": 0.8586639761924744, "num_tokens": 57955624.0, "step": 1525 }, { "epoch": 0.19412288512911843, "ewc_loss": 0.0262451171875, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.300739288330078e-05, "grad_norm": 2.1615030765533447, "learning_rate": 6.464603645612548e-06, "loss": 0.4499, "mean_token_accuracy": 0.8603402376174927, "num_tokens": 57993692.0, "step": 1526 }, { "epoch": 0.19425009540770893, "ewc_loss": 0.0279541015625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.467632293701172e-05, "grad_norm": 2.3782520294189453, "learning_rate": 6.468842729970327e-06, "loss": 0.4936, "mean_token_accuracy": 0.850874662399292, "num_tokens": 58035813.0, "step": 1527 }, { "epoch": 0.19437730568629946, "ewc_loss": 0.026123046875, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.276897430419922e-05, "grad_norm": 2.1808223724365234, "learning_rate": 6.473081814328106e-06, "loss": 0.5301, "mean_token_accuracy": 0.8424539566040039, "num_tokens": 58077118.0, "step": 1528 }, { "epoch": 0.19450451596488996, "ewc_loss": 0.02783203125, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.4557113647460938e-05, "grad_norm": 2.362556219100952, "learning_rate": 6.477320898685885e-06, "loss": 0.5066, "mean_token_accuracy": 0.8437638282775879, "num_tokens": 58120853.0, "step": 1529 }, { "epoch": 0.19463172624348046, "ewc_loss": 0.0262451171875, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.288818359375e-05, "grad_norm": 2.178051233291626, "learning_rate": 6.481559983043663e-06, "loss": 0.5184, "mean_token_accuracy": 0.8360483646392822, "num_tokens": 58160279.0, "step": 1530 }, { "epoch": 0.194758936522071, "ewc_loss": 0.02783203125, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.4557113647460938e-05, "grad_norm": 2.346517324447632, "learning_rate": 6.485799067401442e-06, "loss": 0.4672, "mean_token_accuracy": 0.8568795323371887, "num_tokens": 58200902.0, "step": 1531 }, { "epoch": 0.1948861468006615, "ewc_loss": 0.02587890625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.2530555725097656e-05, "grad_norm": 2.1728861331939697, "learning_rate": 6.490038151759221e-06, "loss": 0.4298, "mean_token_accuracy": 0.8664776086807251, "num_tokens": 58237179.0, "step": 1532 }, { "epoch": 0.195013357079252, "ewc_loss": 0.02734375, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.341996669769287, "learning_rate": 6.494277236117e-06, "loss": 0.4344, "mean_token_accuracy": 0.8635362982749939, "num_tokens": 58271072.0, "step": 1533 }, { "epoch": 0.19514056735784252, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.205371856689453e-05, "grad_norm": 2.173530340194702, "learning_rate": 6.4985163204747785e-06, "loss": 0.5788, "mean_token_accuracy": 0.823434054851532, "num_tokens": 58317186.0, "step": 1534 }, { "epoch": 0.19526777763643302, "ewc_loss": 0.0272216796875, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.396106719970703e-05, "grad_norm": 2.3403267860412598, "learning_rate": 6.5027554048325566e-06, "loss": 0.4856, "mean_token_accuracy": 0.8467469215393066, "num_tokens": 58353396.0, "step": 1535 }, { "epoch": 0.19539498791502352, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.205371856689453e-05, "grad_norm": 2.1872708797454834, "learning_rate": 6.5069944891903354e-06, "loss": 0.558, "mean_token_accuracy": 0.8328934907913208, "num_tokens": 58392466.0, "step": 1536 }, { "epoch": 0.19552219819361405, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.3076746463775635, "learning_rate": 6.511233573548114e-06, "loss": 0.4446, "mean_token_accuracy": 0.8608694076538086, "num_tokens": 58432745.0, "step": 1537 }, { "epoch": 0.19564940847220455, "ewc_loss": 0.0252685546875, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.1905503273010254, "learning_rate": 6.515472657905892e-06, "loss": 0.4949, "mean_token_accuracy": 0.8455528616905212, "num_tokens": 58464288.0, "step": 1538 }, { "epoch": 0.19577661875079505, "ewc_loss": 0.0263671875, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.3126602172851562e-05, "grad_norm": 2.308560371398926, "learning_rate": 6.519711742263671e-06, "loss": 0.5486, "mean_token_accuracy": 0.8316050171852112, "num_tokens": 58501232.0, "step": 1539 }, { "epoch": 0.19590382902938558, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.205371856689453e-05, "grad_norm": 2.1707561016082764, "learning_rate": 6.52395082662145e-06, "loss": 0.5457, "mean_token_accuracy": 0.828779935836792, "num_tokens": 58539749.0, "step": 1540 }, { "epoch": 0.19603103930797608, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.299044370651245, "learning_rate": 6.528189910979229e-06, "loss": 0.4752, "mean_token_accuracy": 0.853363037109375, "num_tokens": 58576436.0, "step": 1541 }, { "epoch": 0.1961582495865666, "ewc_loss": 0.025390625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.205371856689453e-05, "grad_norm": 2.148325204849243, "learning_rate": 6.532428995337008e-06, "loss": 0.4823, "mean_token_accuracy": 0.8516387939453125, "num_tokens": 58618677.0, "step": 1542 }, { "epoch": 0.1962854598651571, "ewc_loss": 0.0269775390625, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 2.3170325756073, "learning_rate": 6.536668079694787e-06, "loss": 0.5072, "mean_token_accuracy": 0.8396083116531372, "num_tokens": 58654404.0, "step": 1543 }, { "epoch": 0.1964126701437476, "ewc_loss": 0.0252685546875, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.1466453075408936, "learning_rate": 6.540907164052566e-06, "loss": 0.4997, "mean_token_accuracy": 0.8434759378433228, "num_tokens": 58697312.0, "step": 1544 }, { "epoch": 0.19653988042233814, "ewc_loss": 0.0274658203125, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.4199485778808594e-05, "grad_norm": 2.333571434020996, "learning_rate": 6.545146248410344e-06, "loss": 0.51, "mean_token_accuracy": 0.8418271541595459, "num_tokens": 58738940.0, "step": 1545 }, { "epoch": 0.19666709070092864, "ewc_loss": 0.025146484375, "ewc_loss_diag": 3.2901763916015625e-05, "ewc_loss_parallel": 2.181529998779297e-05, "grad_norm": 2.149052858352661, "learning_rate": 6.549385332768123e-06, "loss": 0.5146, "mean_token_accuracy": 0.8425419330596924, "num_tokens": 58779626.0, "step": 1546 }, { "epoch": 0.19679430097951914, "ewc_loss": 0.0274658203125, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.4199485778808594e-05, "grad_norm": 2.3569343090057373, "learning_rate": 6.5536244171259015e-06, "loss": 0.5156, "mean_token_accuracy": 0.8388569355010986, "num_tokens": 58818938.0, "step": 1547 }, { "epoch": 0.19692151125810967, "ewc_loss": 0.0252685546875, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.193450927734375e-05, "grad_norm": 2.1493799686431885, "learning_rate": 6.55786350148368e-06, "loss": 0.5288, "mean_token_accuracy": 0.8342807292938232, "num_tokens": 58855973.0, "step": 1548 }, { "epoch": 0.19704872153670017, "ewc_loss": 0.027587890625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.4318695068359375e-05, "grad_norm": 2.3681318759918213, "learning_rate": 6.562102585841459e-06, "loss": 0.4725, "mean_token_accuracy": 0.8494454622268677, "num_tokens": 58894624.0, "step": 1549 }, { "epoch": 0.19717593181529067, "ewc_loss": 0.0255126953125, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.2172927856445312e-05, "grad_norm": 2.1232669353485107, "learning_rate": 6.566341670199238e-06, "loss": 0.4808, "mean_token_accuracy": 0.8508559465408325, "num_tokens": 58934440.0, "step": 1550 }, { "epoch": 0.1973031420938812, "ewc_loss": 0.028076171875, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.47955322265625e-05, "grad_norm": 2.4011800289154053, "learning_rate": 6.570580754557017e-06, "loss": 0.4858, "mean_token_accuracy": 0.8508672714233398, "num_tokens": 58972150.0, "step": 1551 }, { "epoch": 0.1974303523724717, "ewc_loss": 0.0260009765625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.2649765014648438e-05, "grad_norm": 2.132985830307007, "learning_rate": 6.574819838914795e-06, "loss": 0.4598, "mean_token_accuracy": 0.8553540110588074, "num_tokens": 59006179.0, "step": 1552 }, { "epoch": 0.1975575626510622, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.4328725337982178, "learning_rate": 6.579058923272573e-06, "loss": 0.5162, "mean_token_accuracy": 0.8395674228668213, "num_tokens": 59046583.0, "step": 1553 }, { "epoch": 0.19768477292965272, "ewc_loss": 0.02587890625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.2530555725097656e-05, "grad_norm": 2.148043155670166, "learning_rate": 6.583298007630352e-06, "loss": 0.5032, "mean_token_accuracy": 0.843909740447998, "num_tokens": 59078023.0, "step": 1554 }, { "epoch": 0.19781198320824323, "ewc_loss": 0.0291748046875, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.586841583251953e-05, "grad_norm": 2.4598615169525146, "learning_rate": 6.587537091988131e-06, "loss": 0.5121, "mean_token_accuracy": 0.8410018086433411, "num_tokens": 59121160.0, "step": 1555 }, { "epoch": 0.19793919348683373, "ewc_loss": 0.0262451171875, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.300739288330078e-05, "grad_norm": 2.1383211612701416, "learning_rate": 6.59177617634591e-06, "loss": 0.4791, "mean_token_accuracy": 0.8481625914573669, "num_tokens": 59164564.0, "step": 1556 }, { "epoch": 0.19806640376542425, "ewc_loss": 0.030029296875, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.6702880859375e-05, "grad_norm": 2.476555109024048, "learning_rate": 6.596015260703689e-06, "loss": 0.4859, "mean_token_accuracy": 0.8526875972747803, "num_tokens": 59211558.0, "step": 1557 }, { "epoch": 0.19819361404401475, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.172159194946289, "learning_rate": 6.6002543450614675e-06, "loss": 0.4772, "mean_token_accuracy": 0.8514338135719299, "num_tokens": 59251260.0, "step": 1558 }, { "epoch": 0.19832082432260525, "ewc_loss": 0.0299072265625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.658367156982422e-05, "grad_norm": 2.496006727218628, "learning_rate": 6.604493429419246e-06, "loss": 0.5413, "mean_token_accuracy": 0.8351198434829712, "num_tokens": 59293949.0, "step": 1559 }, { "epoch": 0.19844803460119578, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.195225477218628, "learning_rate": 6.6087325137770244e-06, "loss": 0.5502, "mean_token_accuracy": 0.8383879661560059, "num_tokens": 59341549.0, "step": 1560 }, { "epoch": 0.19857524487978628, "ewc_loss": 0.0301513671875, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.682209014892578e-05, "grad_norm": 2.4891068935394287, "learning_rate": 6.612971598134803e-06, "loss": 0.493, "mean_token_accuracy": 0.8467992544174194, "num_tokens": 59379355.0, "step": 1561 }, { "epoch": 0.19870245515837678, "ewc_loss": 0.027587890625, "ewc_loss_diag": 3.314018249511719e-05, "ewc_loss_parallel": 2.4318695068359375e-05, "grad_norm": 2.229902744293213, "learning_rate": 6.617210682492582e-06, "loss": 0.5369, "mean_token_accuracy": 0.8318372964859009, "num_tokens": 59416340.0, "step": 1562 }, { "epoch": 0.1988296654369673, "ewc_loss": 0.029541015625, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.6226043701171875e-05, "grad_norm": 2.468390703201294, "learning_rate": 6.621449766850361e-06, "loss": 0.4653, "mean_token_accuracy": 0.8559190630912781, "num_tokens": 59453100.0, "step": 1563 }, { "epoch": 0.1989568757155578, "ewc_loss": 0.0274658203125, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.4199485778808594e-05, "grad_norm": 2.221693754196167, "learning_rate": 6.62568885120814e-06, "loss": 0.5061, "mean_token_accuracy": 0.8429961800575256, "num_tokens": 59491164.0, "step": 1564 }, { "epoch": 0.19908408599414834, "ewc_loss": 0.029541015625, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.6226043701171875e-05, "grad_norm": 2.461503267288208, "learning_rate": 6.629927935565919e-06, "loss": 0.5143, "mean_token_accuracy": 0.8417251706123352, "num_tokens": 59525752.0, "step": 1565 }, { "epoch": 0.19921129627273884, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.237682342529297, "learning_rate": 6.634167019923698e-06, "loss": 0.5493, "mean_token_accuracy": 0.8354941606521606, "num_tokens": 59565206.0, "step": 1566 }, { "epoch": 0.19933850655132934, "ewc_loss": 0.0289306640625, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 2.439265251159668, "learning_rate": 6.638406104281477e-06, "loss": 0.4902, "mean_token_accuracy": 0.8481454253196716, "num_tokens": 59597578.0, "step": 1567 }, { "epoch": 0.19946571682991987, "ewc_loss": 0.0269775390625, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 2.2108335494995117, "learning_rate": 6.642645188639254e-06, "loss": 0.4629, "mean_token_accuracy": 0.8548903465270996, "num_tokens": 59640073.0, "step": 1568 }, { "epoch": 0.19959292710851037, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.4213814735412598, "learning_rate": 6.646884272997033e-06, "loss": 0.518, "mean_token_accuracy": 0.8385869264602661, "num_tokens": 59680259.0, "step": 1569 }, { "epoch": 0.19972013738710087, "ewc_loss": 0.0269775390625, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 2.224785804748535, "learning_rate": 6.651123357354812e-06, "loss": 0.5015, "mean_token_accuracy": 0.8444411754608154, "num_tokens": 59716372.0, "step": 1570 }, { "epoch": 0.1998473476656914, "ewc_loss": 0.0284423828125, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.5153160095214844e-05, "grad_norm": 2.429809808731079, "learning_rate": 6.6553624417125905e-06, "loss": 0.5219, "mean_token_accuracy": 0.8374418020248413, "num_tokens": 59751294.0, "step": 1571 }, { "epoch": 0.1999745579442819, "ewc_loss": 0.0267333984375, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.211115837097168, "learning_rate": 6.659601526070369e-06, "loss": 0.553, "mean_token_accuracy": 0.8352688550949097, "num_tokens": 59794395.0, "step": 1572 }, { "epoch": 0.2001017682228724, "ewc_loss": 0.0286865234375, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.5391578674316406e-05, "grad_norm": 2.4461092948913574, "learning_rate": 6.663840610428148e-06, "loss": 0.5816, "mean_token_accuracy": 0.8261787295341492, "num_tokens": 59825503.0, "step": 1573 }, { "epoch": 0.20022897850146293, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.2153384685516357, "learning_rate": 6.668079694785927e-06, "loss": 0.4641, "mean_token_accuracy": 0.8572162389755249, "num_tokens": 59864634.0, "step": 1574 }, { "epoch": 0.20035618878005343, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.4267265796661377, "learning_rate": 6.672318779143705e-06, "loss": 0.4851, "mean_token_accuracy": 0.8487992882728577, "num_tokens": 59898693.0, "step": 1575 }, { "epoch": 0.20048339905864393, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.337860107421875e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.2236878871917725, "learning_rate": 6.676557863501484e-06, "loss": 0.5056, "mean_token_accuracy": 0.8442823886871338, "num_tokens": 59930234.0, "step": 1576 }, { "epoch": 0.20061060933723446, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.4144787788391113, "learning_rate": 6.680796947859263e-06, "loss": 0.5022, "mean_token_accuracy": 0.8444501161575317, "num_tokens": 59970932.0, "step": 1577 }, { "epoch": 0.20073781961582496, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.2134850025177, "learning_rate": 6.685036032217042e-06, "loss": 0.4926, "mean_token_accuracy": 0.8460102081298828, "num_tokens": 60007282.0, "step": 1578 }, { "epoch": 0.20086502989441546, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.5391578674316406e-05, "grad_norm": 2.4037649631500244, "learning_rate": 6.689275116574821e-06, "loss": 0.4791, "mean_token_accuracy": 0.8546197414398193, "num_tokens": 60043336.0, "step": 1579 }, { "epoch": 0.200992240173006, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.2024381160736084, "learning_rate": 6.6935142009326e-06, "loss": 0.5038, "mean_token_accuracy": 0.8481305837631226, "num_tokens": 60083284.0, "step": 1580 }, { "epoch": 0.2011194504515965, "ewc_loss": 0.029052734375, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 2.3981785774230957, "learning_rate": 6.6977532852903785e-06, "loss": 0.5083, "mean_token_accuracy": 0.8412944674491882, "num_tokens": 60125115.0, "step": 1581 }, { "epoch": 0.201246660730187, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 2.201016664505005, "learning_rate": 6.701992369648157e-06, "loss": 0.4223, "mean_token_accuracy": 0.8676832914352417, "num_tokens": 60161497.0, "step": 1582 }, { "epoch": 0.20137387100877752, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.5391578674316406e-05, "grad_norm": 2.3962562084198, "learning_rate": 6.7062314540059345e-06, "loss": 0.4522, "mean_token_accuracy": 0.8593279123306274, "num_tokens": 60202282.0, "step": 1583 }, { "epoch": 0.20150108128736802, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.3245811462402344e-05, "grad_norm": 2.2066736221313477, "learning_rate": 6.7104705383637134e-06, "loss": 0.4837, "mean_token_accuracy": 0.8489723801612854, "num_tokens": 60241211.0, "step": 1584 }, { "epoch": 0.20162829156595852, "ewc_loss": 0.0283203125, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.5033950805664062e-05, "grad_norm": 2.422858476638794, "learning_rate": 6.714709622721492e-06, "loss": 0.5256, "mean_token_accuracy": 0.8379716873168945, "num_tokens": 60273677.0, "step": 1585 }, { "epoch": 0.20175550184454905, "ewc_loss": 0.026611328125, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.189764976501465, "learning_rate": 6.718948707079271e-06, "loss": 0.4814, "mean_token_accuracy": 0.8516695499420166, "num_tokens": 60313904.0, "step": 1586 }, { "epoch": 0.20188271212313955, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.4229736328125, "learning_rate": 6.72318779143705e-06, "loss": 0.4835, "mean_token_accuracy": 0.8505493402481079, "num_tokens": 60351534.0, "step": 1587 }, { "epoch": 0.20200992240173005, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.361701965332031e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.1865687370300293, "learning_rate": 6.727426875794829e-06, "loss": 0.5255, "mean_token_accuracy": 0.8397796750068665, "num_tokens": 60393362.0, "step": 1588 }, { "epoch": 0.20213713268032057, "ewc_loss": 0.029296875, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.586841583251953e-05, "grad_norm": 2.43176007270813, "learning_rate": 6.731665960152608e-06, "loss": 0.4997, "mean_token_accuracy": 0.8440254926681519, "num_tokens": 60437735.0, "step": 1589 }, { "epoch": 0.20226434295891108, "ewc_loss": 0.0267333984375, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.3365020751953125e-05, "grad_norm": 2.2100300788879395, "learning_rate": 6.735905044510387e-06, "loss": 0.5557, "mean_token_accuracy": 0.8232405781745911, "num_tokens": 60476032.0, "step": 1590 }, { "epoch": 0.2023915532375016, "ewc_loss": 0.0291748046875, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.574920654296875e-05, "grad_norm": 2.448793888092041, "learning_rate": 6.740144128868165e-06, "loss": 0.51, "mean_token_accuracy": 0.8423967957496643, "num_tokens": 60513330.0, "step": 1591 }, { "epoch": 0.2025187635160921, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.1920533180236816, "learning_rate": 6.744383213225944e-06, "loss": 0.5351, "mean_token_accuracy": 0.8350517749786377, "num_tokens": 60552829.0, "step": 1592 }, { "epoch": 0.2026459737946826, "ewc_loss": 0.0296630859375, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.6226043701171875e-05, "grad_norm": 2.463444948196411, "learning_rate": 6.7486222975837225e-06, "loss": 0.471, "mean_token_accuracy": 0.8499823808670044, "num_tokens": 60585333.0, "step": 1593 }, { "epoch": 0.20277318407327313, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.207453727722168, "learning_rate": 6.7528613819415014e-06, "loss": 0.4998, "mean_token_accuracy": 0.8422648906707764, "num_tokens": 60621091.0, "step": 1594 }, { "epoch": 0.20290039435186363, "ewc_loss": 0.02978515625, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.6345252990722656e-05, "grad_norm": 2.468820333480835, "learning_rate": 6.75710046629928e-06, "loss": 0.5047, "mean_token_accuracy": 0.8466514348983765, "num_tokens": 60661602.0, "step": 1595 }, { "epoch": 0.20302760463045413, "ewc_loss": 0.0269775390625, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.3603439331054688e-05, "grad_norm": 2.2441916465759277, "learning_rate": 6.761339550657059e-06, "loss": 0.5078, "mean_token_accuracy": 0.8386706113815308, "num_tokens": 60689085.0, "step": 1596 }, { "epoch": 0.20315481490904466, "ewc_loss": 0.0294189453125, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.5987625122070312e-05, "grad_norm": 2.4722118377685547, "learning_rate": 6.765578635014838e-06, "loss": 0.5044, "mean_token_accuracy": 0.8476608991622925, "num_tokens": 60727259.0, "step": 1597 }, { "epoch": 0.20328202518763516, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 2.215172529220581, "learning_rate": 6.769817719372615e-06, "loss": 0.4771, "mean_token_accuracy": 0.8501039743423462, "num_tokens": 60761880.0, "step": 1598 }, { "epoch": 0.20340923546622566, "ewc_loss": 0.02978515625, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.6345252990722656e-05, "grad_norm": 2.467926263809204, "learning_rate": 6.774056803730394e-06, "loss": 0.5214, "mean_token_accuracy": 0.8399991393089294, "num_tokens": 60795416.0, "step": 1599 }, { "epoch": 0.2035364457448162, "ewc_loss": 0.027587890625, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.4199485778808594e-05, "grad_norm": 2.2126305103302, "learning_rate": 6.778295888088173e-06, "loss": 0.4965, "mean_token_accuracy": 0.8465018272399902, "num_tokens": 60833373.0, "step": 1600 }, { "epoch": 0.2036636560234067, "ewc_loss": 0.0301513671875, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.6702880859375e-05, "grad_norm": 2.4811272621154785, "learning_rate": 6.782534972445952e-06, "loss": 0.5148, "mean_token_accuracy": 0.8386102318763733, "num_tokens": 60870024.0, "step": 1601 }, { "epoch": 0.2037908663019972, "ewc_loss": 0.02783203125, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.4437904357910156e-05, "grad_norm": 2.2160892486572266, "learning_rate": 6.786774056803731e-06, "loss": 0.5055, "mean_token_accuracy": 0.8454362750053406, "num_tokens": 60909790.0, "step": 1602 }, { "epoch": 0.20391807658058772, "ewc_loss": 0.0302734375, "ewc_loss_diag": 3.3855438232421875e-05, "ewc_loss_parallel": 2.682209014892578e-05, "grad_norm": 2.48765230178833, "learning_rate": 6.79101314116151e-06, "loss": 0.5155, "mean_token_accuracy": 0.8411623239517212, "num_tokens": 60948793.0, "step": 1603 }, { "epoch": 0.20404528685917822, "ewc_loss": 0.0279541015625, "ewc_loss_diag": 3.409385681152344e-05, "ewc_loss_parallel": 2.4557113647460938e-05, "grad_norm": 2.2256293296813965, "learning_rate": 6.795252225519289e-06, "loss": 0.54, "mean_token_accuracy": 0.8326965570449829, "num_tokens": 60991370.0, "step": 1604 }, { "epoch": 0.20417249713776872, "ewc_loss": 0.0303955078125, "ewc_loss_diag": 3.409385681152344e-05, "ewc_loss_parallel": 2.6941299438476562e-05, "grad_norm": 2.486698865890503, "learning_rate": 6.7994913098770675e-06, "loss": 0.5074, "mean_token_accuracy": 0.8421770334243774, "num_tokens": 61028137.0, "step": 1605 }, { "epoch": 0.20429970741635925, "ewc_loss": 0.028076171875, "ewc_loss_diag": 3.409385681152344e-05, "ewc_loss_parallel": 2.467632293701172e-05, "grad_norm": 2.357342004776001, "learning_rate": 6.8037303942348455e-06, "loss": 0.4612, "mean_token_accuracy": 0.8566191792488098, "num_tokens": 61060843.0, "step": 1606 }, { "epoch": 0.20442691769494975, "ewc_loss": 0.0294189453125, "ewc_loss_diag": 3.409385681152344e-05, "ewc_loss_parallel": 2.5987625122070312e-05, "grad_norm": 2.421097755432129, "learning_rate": 6.807969478592624e-06, "loss": 0.5074, "mean_token_accuracy": 0.8456616401672363, "num_tokens": 61096272.0, "step": 1607 }, { "epoch": 0.20455412797354025, "ewc_loss": 0.02783203125, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 2.4437904357910156e-05, "grad_norm": 2.2481322288513184, "learning_rate": 6.812208562950403e-06, "loss": 0.5273, "mean_token_accuracy": 0.8433264493942261, "num_tokens": 61139119.0, "step": 1608 }, { "epoch": 0.20468133825213078, "ewc_loss": 0.029052734375, "ewc_loss_diag": 3.409385681152344e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 2.3864943981170654, "learning_rate": 6.816447647308182e-06, "loss": 0.5605, "mean_token_accuracy": 0.8282395005226135, "num_tokens": 61183327.0, "step": 1609 }, { "epoch": 0.20480854853072128, "ewc_loss": 0.0277099609375, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 2.4318695068359375e-05, "grad_norm": 2.254577159881592, "learning_rate": 6.820686731665961e-06, "loss": 0.4931, "mean_token_accuracy": 0.8475564122200012, "num_tokens": 61218787.0, "step": 1610 }, { "epoch": 0.20493575880931178, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 2.5391578674316406e-05, "grad_norm": 2.3930773735046387, "learning_rate": 6.82492581602374e-06, "loss": 0.5006, "mean_token_accuracy": 0.8462328314781189, "num_tokens": 61256517.0, "step": 1611 }, { "epoch": 0.2050629690879023, "ewc_loss": 0.0274658203125, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.230808973312378, "learning_rate": 6.829164900381519e-06, "loss": 0.4878, "mean_token_accuracy": 0.849016010761261, "num_tokens": 61301065.0, "step": 1612 }, { "epoch": 0.2051901793664928, "ewc_loss": 0.0289306640625, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.4029951095581055, "learning_rate": 6.833403984739296e-06, "loss": 0.4933, "mean_token_accuracy": 0.8461768627166748, "num_tokens": 61336368.0, "step": 1613 }, { "epoch": 0.2053173896450833, "ewc_loss": 0.0272216796875, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.2425999641418457, "learning_rate": 6.837643069097075e-06, "loss": 0.5707, "mean_token_accuracy": 0.8249925374984741, "num_tokens": 61373742.0, "step": 1614 }, { "epoch": 0.20544459992367384, "ewc_loss": 0.0289306640625, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.41662859916687, "learning_rate": 6.841882153454854e-06, "loss": 0.4817, "mean_token_accuracy": 0.8491209745407104, "num_tokens": 61408454.0, "step": 1615 }, { "epoch": 0.20557181020226434, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.4332275390625e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 2.305227041244507, "learning_rate": 6.846121237812633e-06, "loss": 0.4499, "mean_token_accuracy": 0.8594872951507568, "num_tokens": 61445420.0, "step": 1616 }, { "epoch": 0.20569902048085487, "ewc_loss": 0.0286865234375, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 2.5272369384765625e-05, "grad_norm": 2.3534250259399414, "learning_rate": 6.8503603221704115e-06, "loss": 0.4818, "mean_token_accuracy": 0.850998044013977, "num_tokens": 61487607.0, "step": 1617 }, { "epoch": 0.20582623075944537, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 2.2592358589172363, "learning_rate": 6.85459940652819e-06, "loss": 0.4869, "mean_token_accuracy": 0.8478755950927734, "num_tokens": 61521053.0, "step": 1618 }, { "epoch": 0.20595344103803587, "ewc_loss": 0.028564453125, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 2.5153160095214844e-05, "grad_norm": 2.3400416374206543, "learning_rate": 6.858838490885969e-06, "loss": 0.5478, "mean_token_accuracy": 0.8295272588729858, "num_tokens": 61562211.0, "step": 1619 }, { "epoch": 0.2060806513166264, "ewc_loss": 0.0274658203125, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.2580270767211914, "learning_rate": 6.863077575243748e-06, "loss": 0.5584, "mean_token_accuracy": 0.8327764272689819, "num_tokens": 61600089.0, "step": 1620 }, { "epoch": 0.2062078615952169, "ewc_loss": 0.0283203125, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 2.491474151611328e-05, "grad_norm": 2.3403985500335693, "learning_rate": 6.867316659601526e-06, "loss": 0.4809, "mean_token_accuracy": 0.8489052057266235, "num_tokens": 61636226.0, "step": 1621 }, { "epoch": 0.2063350718738074, "ewc_loss": 0.02734375, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 2.396106719970703e-05, "grad_norm": 2.2373833656311035, "learning_rate": 6.871555743959305e-06, "loss": 0.5071, "mean_token_accuracy": 0.8452615141868591, "num_tokens": 61677224.0, "step": 1622 }, { "epoch": 0.20646228215239792, "ewc_loss": 0.0283203125, "ewc_loss_diag": 3.457069396972656e-05, "ewc_loss_parallel": 2.491474151611328e-05, "grad_norm": 2.3501882553100586, "learning_rate": 6.875794828317084e-06, "loss": 0.4844, "mean_token_accuracy": 0.8505039215087891, "num_tokens": 61715174.0, "step": 1623 }, { "epoch": 0.20658949243098843, "ewc_loss": 0.027099609375, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 2.372264862060547e-05, "grad_norm": 2.218190908432007, "learning_rate": 6.880033912674863e-06, "loss": 0.4707, "mean_token_accuracy": 0.8551836013793945, "num_tokens": 61760819.0, "step": 1624 }, { "epoch": 0.20671670270957893, "ewc_loss": 0.028564453125, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 2.5153160095214844e-05, "grad_norm": 2.385277032852173, "learning_rate": 6.884272997032642e-06, "loss": 0.5816, "mean_token_accuracy": 0.8229838609695435, "num_tokens": 61800502.0, "step": 1625 }, { "epoch": 0.20684391298816945, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.211361885070801, "learning_rate": 6.888512081390421e-06, "loss": 0.465, "mean_token_accuracy": 0.8564770817756653, "num_tokens": 61834922.0, "step": 1626 }, { "epoch": 0.20697112326675995, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 2.5272369384765625e-05, "grad_norm": 2.4236929416656494, "learning_rate": 6.8927511657481995e-06, "loss": 0.5261, "mean_token_accuracy": 0.8381171226501465, "num_tokens": 61874675.0, "step": 1627 }, { "epoch": 0.20709833354535045, "ewc_loss": 0.02685546875, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 2.3484230041503906e-05, "grad_norm": 2.1892025470733643, "learning_rate": 6.896990250105978e-06, "loss": 0.4685, "mean_token_accuracy": 0.8568729162216187, "num_tokens": 61914310.0, "step": 1628 }, { "epoch": 0.20722554382394098, "ewc_loss": 0.029541015625, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 2.6106834411621094e-05, "grad_norm": 2.463254928588867, "learning_rate": 6.901229334463756e-06, "loss": 0.4761, "mean_token_accuracy": 0.8501273989677429, "num_tokens": 61949142.0, "step": 1629 }, { "epoch": 0.20735275410253148, "ewc_loss": 0.02734375, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 2.396106719970703e-05, "grad_norm": 2.1765823364257812, "learning_rate": 6.9054684188215345e-06, "loss": 0.4921, "mean_token_accuracy": 0.8529344797134399, "num_tokens": 61989939.0, "step": 1630 }, { "epoch": 0.20747996438112198, "ewc_loss": 0.031005859375, "ewc_loss_diag": 3.4809112548828125e-05, "ewc_loss_parallel": 2.753734588623047e-05, "grad_norm": 2.5235097408294678, "learning_rate": 6.909707503179313e-06, "loss": 0.4992, "mean_token_accuracy": 0.8445952534675598, "num_tokens": 62029263.0, "step": 1631 }, { "epoch": 0.2076071746597125, "ewc_loss": 0.0277099609375, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.4199485778808594e-05, "grad_norm": 2.1665239334106445, "learning_rate": 6.913946587537092e-06, "loss": 0.4837, "mean_token_accuracy": 0.8510605096817017, "num_tokens": 62067963.0, "step": 1632 }, { "epoch": 0.207734384938303, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.86102294921875e-05, "grad_norm": 2.6182000637054443, "learning_rate": 6.918185671894871e-06, "loss": 0.5704, "mean_token_accuracy": 0.8254247903823853, "num_tokens": 62105313.0, "step": 1633 }, { "epoch": 0.2078615952168935, "ewc_loss": 0.028564453125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.5033950805664062e-05, "grad_norm": 2.210419178009033, "learning_rate": 6.92242475625265e-06, "loss": 0.5353, "mean_token_accuracy": 0.8469207286834717, "num_tokens": 62140836.0, "step": 1634 }, { "epoch": 0.20798880549548404, "ewc_loss": 0.03271484375, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.6506850719451904, "learning_rate": 6.926663840610429e-06, "loss": 0.4558, "mean_token_accuracy": 0.8607196807861328, "num_tokens": 62178951.0, "step": 1635 }, { "epoch": 0.20811601577407454, "ewc_loss": 0.029541015625, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.5987625122070312e-05, "grad_norm": 2.2365262508392334, "learning_rate": 6.930902924968207e-06, "loss": 0.4893, "mean_token_accuracy": 0.8467059135437012, "num_tokens": 62216941.0, "step": 1636 }, { "epoch": 0.20824322605266504, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.968311309814453e-05, "grad_norm": 2.70591402053833, "learning_rate": 6.935142009325986e-06, "loss": 0.5306, "mean_token_accuracy": 0.8389480113983154, "num_tokens": 62257676.0, "step": 1637 }, { "epoch": 0.20837043633125557, "ewc_loss": 0.030517578125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.6941299438476562e-05, "grad_norm": 2.3043861389160156, "learning_rate": 6.939381093683765e-06, "loss": 0.5393, "mean_token_accuracy": 0.834879457950592, "num_tokens": 62296878.0, "step": 1638 }, { "epoch": 0.20849764660984607, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.968311309814453e-05, "grad_norm": 2.6940114498138428, "learning_rate": 6.943620178041544e-06, "loss": 0.4631, "mean_token_accuracy": 0.8581688404083252, "num_tokens": 62335292.0, "step": 1639 }, { "epoch": 0.2086248568884366, "ewc_loss": 0.0311279296875, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.753734588623047e-05, "grad_norm": 2.3657023906707764, "learning_rate": 6.9478592623993225e-06, "loss": 0.5015, "mean_token_accuracy": 0.8522838950157166, "num_tokens": 62371182.0, "step": 1640 }, { "epoch": 0.2087520671670271, "ewc_loss": 0.032470703125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.654910087585449, "learning_rate": 6.952098346757101e-06, "loss": 0.4922, "mean_token_accuracy": 0.8483834266662598, "num_tokens": 62407803.0, "step": 1641 }, { "epoch": 0.2088792774456176, "ewc_loss": 0.03125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.765655517578125e-05, "grad_norm": 2.3836562633514404, "learning_rate": 6.95633743111488e-06, "loss": 0.4794, "mean_token_accuracy": 0.8534016013145447, "num_tokens": 62446320.0, "step": 1642 }, { "epoch": 0.20900648772420813, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.8252601623535156e-05, "grad_norm": 2.6102116107940674, "learning_rate": 6.960576515472659e-06, "loss": 0.5246, "mean_token_accuracy": 0.8406872749328613, "num_tokens": 62486919.0, "step": 1643 }, { "epoch": 0.20913369800279863, "ewc_loss": 0.031005859375, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.386854410171509, "learning_rate": 6.964815599830436e-06, "loss": 0.502, "mean_token_accuracy": 0.8484940528869629, "num_tokens": 62521511.0, "step": 1644 }, { "epoch": 0.20926090828138913, "ewc_loss": 0.03125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.765655517578125e-05, "grad_norm": 2.5328738689422607, "learning_rate": 6.969054684188215e-06, "loss": 0.559, "mean_token_accuracy": 0.824749767780304, "num_tokens": 62564702.0, "step": 1645 }, { "epoch": 0.20938811855997966, "ewc_loss": 0.030029296875, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.6464462280273438e-05, "grad_norm": 2.349655866622925, "learning_rate": 6.973293768545994e-06, "loss": 0.4688, "mean_token_accuracy": 0.8561902046203613, "num_tokens": 62605850.0, "step": 1646 }, { "epoch": 0.20951532883857016, "ewc_loss": 0.0308837890625, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 2.4857394695281982, "learning_rate": 6.977532852903773e-06, "loss": 0.4986, "mean_token_accuracy": 0.8443433046340942, "num_tokens": 62650497.0, "step": 1647 }, { "epoch": 0.20964253911716066, "ewc_loss": 0.029296875, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.574920654296875e-05, "grad_norm": 2.351646661758423, "learning_rate": 6.981771937261552e-06, "loss": 0.5503, "mean_token_accuracy": 0.8332728147506714, "num_tokens": 62685538.0, "step": 1648 }, { "epoch": 0.2097697493957512, "ewc_loss": 0.0303955078125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.682209014892578e-05, "grad_norm": 2.4560928344726562, "learning_rate": 6.986011021619331e-06, "loss": 0.5085, "mean_token_accuracy": 0.8446797728538513, "num_tokens": 62725076.0, "step": 1649 }, { "epoch": 0.2098969596743417, "ewc_loss": 0.028564453125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.5033950805664062e-05, "grad_norm": 2.33742356300354, "learning_rate": 6.99025010597711e-06, "loss": 0.5233, "mean_token_accuracy": 0.8389587998390198, "num_tokens": 62762335.0, "step": 1650 }, { "epoch": 0.2100241699529322, "ewc_loss": 0.0296630859375, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.6106834411621094e-05, "grad_norm": 2.4111392498016357, "learning_rate": 6.994489190334888e-06, "loss": 0.4828, "mean_token_accuracy": 0.850100040435791, "num_tokens": 62804507.0, "step": 1651 }, { "epoch": 0.21015138023152272, "ewc_loss": 0.0283203125, "ewc_loss_diag": 3.504753112792969e-05, "ewc_loss_parallel": 2.47955322265625e-05, "grad_norm": 2.3010544776916504, "learning_rate": 6.9987282746926666e-06, "loss": 0.4512, "mean_token_accuracy": 0.8606946468353271, "num_tokens": 62844499.0, "step": 1652 }, { "epoch": 0.21027859051011322, "ewc_loss": 0.0291748046875, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 2.423955202102661, "learning_rate": 7.0029673590504455e-06, "loss": 0.4943, "mean_token_accuracy": 0.8520065546035767, "num_tokens": 62880061.0, "step": 1653 }, { "epoch": 0.21040580078870372, "ewc_loss": 0.02783203125, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 2.4318695068359375e-05, "grad_norm": 2.2738659381866455, "learning_rate": 7.007206443408224e-06, "loss": 0.4413, "mean_token_accuracy": 0.8612407445907593, "num_tokens": 62916726.0, "step": 1654 }, { "epoch": 0.21053301106729425, "ewc_loss": 0.0294189453125, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 2.586841583251953e-05, "grad_norm": 2.43705677986145, "learning_rate": 7.011445527766003e-06, "loss": 0.4889, "mean_token_accuracy": 0.8464140892028809, "num_tokens": 62953341.0, "step": 1655 }, { "epoch": 0.21066022134588475, "ewc_loss": 0.0277099609375, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 2.4199485778808594e-05, "grad_norm": 2.2583162784576416, "learning_rate": 7.015684612123782e-06, "loss": 0.4806, "mean_token_accuracy": 0.8497603535652161, "num_tokens": 62995047.0, "step": 1656 }, { "epoch": 0.21078743162447525, "ewc_loss": 0.030029296875, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 2.6464462280273438e-05, "grad_norm": 2.464177370071411, "learning_rate": 7.019923696481561e-06, "loss": 0.5123, "mean_token_accuracy": 0.8432170152664185, "num_tokens": 63039520.0, "step": 1657 }, { "epoch": 0.21091464190306577, "ewc_loss": 0.02734375, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.2681961059570312, "learning_rate": 7.02416278083934e-06, "loss": 0.5267, "mean_token_accuracy": 0.8324614763259888, "num_tokens": 63071243.0, "step": 1658 }, { "epoch": 0.21104185218165628, "ewc_loss": 0.0302734375, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 2.6702880859375e-05, "grad_norm": 2.4723949432373047, "learning_rate": 7.028401865197118e-06, "loss": 0.5257, "mean_token_accuracy": 0.836600124835968, "num_tokens": 63114496.0, "step": 1659 }, { "epoch": 0.21116906246024678, "ewc_loss": 0.02734375, "ewc_loss_diag": 3.528594970703125e-05, "ewc_loss_parallel": 2.384185791015625e-05, "grad_norm": 2.221259355545044, "learning_rate": 7.032640949554897e-06, "loss": 0.4327, "mean_token_accuracy": 0.8605777025222778, "num_tokens": 63158111.0, "step": 1660 }, { "epoch": 0.2112962727388373, "ewc_loss": 0.0308837890625, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 2.5233867168426514, "learning_rate": 7.036880033912676e-06, "loss": 0.5349, "mean_token_accuracy": 0.8357983827590942, "num_tokens": 63198305.0, "step": 1661 }, { "epoch": 0.2114234830174278, "ewc_loss": 0.027587890625, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.4080276489257812e-05, "grad_norm": 2.2239744663238525, "learning_rate": 7.041119118270454e-06, "loss": 0.4803, "mean_token_accuracy": 0.8499369621276855, "num_tokens": 63235053.0, "step": 1662 }, { "epoch": 0.2115506932960183, "ewc_loss": 0.031494140625, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.7894973754882812e-05, "grad_norm": 2.5779287815093994, "learning_rate": 7.045358202628233e-06, "loss": 0.5329, "mean_token_accuracy": 0.8395905494689941, "num_tokens": 63275495.0, "step": 1663 }, { "epoch": 0.21167790357460883, "ewc_loss": 0.028564453125, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.5033950805664062e-05, "grad_norm": 2.231137752532959, "learning_rate": 7.0495972869860115e-06, "loss": 0.525, "mean_token_accuracy": 0.8389617204666138, "num_tokens": 63314338.0, "step": 1664 }, { "epoch": 0.21180511385319933, "ewc_loss": 0.032470703125, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.8848648071289062e-05, "grad_norm": 2.6378631591796875, "learning_rate": 7.05383637134379e-06, "loss": 0.4814, "mean_token_accuracy": 0.852098286151886, "num_tokens": 63348582.0, "step": 1665 }, { "epoch": 0.21193232413178986, "ewc_loss": 0.029052734375, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.2598283290863037, "learning_rate": 7.058075455701568e-06, "loss": 0.5224, "mean_token_accuracy": 0.8333279490470886, "num_tokens": 63388496.0, "step": 1666 }, { "epoch": 0.21205953441038036, "ewc_loss": 0.032958984375, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.668687105178833, "learning_rate": 7.062314540059347e-06, "loss": 0.4768, "mean_token_accuracy": 0.855624258518219, "num_tokens": 63426216.0, "step": 1667 }, { "epoch": 0.21218674468897086, "ewc_loss": 0.0301513671875, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.658367156982422e-05, "grad_norm": 2.2895047664642334, "learning_rate": 7.066553624417126e-06, "loss": 0.4757, "mean_token_accuracy": 0.852017879486084, "num_tokens": 63460627.0, "step": 1668 }, { "epoch": 0.2123139549675614, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.968311309814453e-05, "grad_norm": 2.671138286590576, "learning_rate": 7.070792708774905e-06, "loss": 0.5116, "mean_token_accuracy": 0.8433072566986084, "num_tokens": 63501911.0, "step": 1669 }, { "epoch": 0.2124411652461519, "ewc_loss": 0.0308837890625, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 2.3165283203125, "learning_rate": 7.075031793132684e-06, "loss": 0.5317, "mean_token_accuracy": 0.8405649662017822, "num_tokens": 63548764.0, "step": 1670 }, { "epoch": 0.2125683755247424, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.968311309814453e-05, "grad_norm": 2.669410228729248, "learning_rate": 7.079270877490463e-06, "loss": 0.4829, "mean_token_accuracy": 0.8511670231819153, "num_tokens": 63583347.0, "step": 1671 }, { "epoch": 0.21269558580333292, "ewc_loss": 0.031005859375, "ewc_loss_diag": 3.552436828613281e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.3684959411621094, "learning_rate": 7.083509961848242e-06, "loss": 0.562, "mean_token_accuracy": 0.830054759979248, "num_tokens": 63624115.0, "step": 1672 }, { "epoch": 0.21282279608192342, "ewc_loss": 0.03271484375, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 2.6305346488952637, "learning_rate": 7.087749046206021e-06, "loss": 0.4825, "mean_token_accuracy": 0.8561604619026184, "num_tokens": 63660850.0, "step": 1673 }, { "epoch": 0.21295000636051392, "ewc_loss": 0.03125, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.765655517578125e-05, "grad_norm": 2.3777005672454834, "learning_rate": 7.091988130563799e-06, "loss": 0.5265, "mean_token_accuracy": 0.837486207485199, "num_tokens": 63701530.0, "step": 1674 }, { "epoch": 0.21307721663910445, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.584233283996582, "learning_rate": 7.0962272149215775e-06, "loss": 0.5246, "mean_token_accuracy": 0.8382375240325928, "num_tokens": 63740400.0, "step": 1675 }, { "epoch": 0.21320442691769495, "ewc_loss": 0.03076171875, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.7060508728027344e-05, "grad_norm": 2.384639263153076, "learning_rate": 7.100466299279356e-06, "loss": 0.5063, "mean_token_accuracy": 0.8460714817047119, "num_tokens": 63775019.0, "step": 1676 }, { "epoch": 0.21333163719628545, "ewc_loss": 0.031494140625, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.7894973754882812e-05, "grad_norm": 2.5342469215393066, "learning_rate": 7.104705383637135e-06, "loss": 0.5282, "mean_token_accuracy": 0.8428420424461365, "num_tokens": 63813154.0, "step": 1677 }, { "epoch": 0.21345884747487598, "ewc_loss": 0.030517578125, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.6941299438476562e-05, "grad_norm": 2.358499050140381, "learning_rate": 7.108944467994914e-06, "loss": 0.4899, "mean_token_accuracy": 0.8507145047187805, "num_tokens": 63851982.0, "step": 1678 }, { "epoch": 0.21358605775346648, "ewc_loss": 0.031005859375, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.513906717300415, "learning_rate": 7.113183552352693e-06, "loss": 0.5596, "mean_token_accuracy": 0.8265404105186462, "num_tokens": 63888983.0, "step": 1679 }, { "epoch": 0.21371326803205698, "ewc_loss": 0.030029296875, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.6345252990722656e-05, "grad_norm": 2.33998966217041, "learning_rate": 7.117422636710471e-06, "loss": 0.4469, "mean_token_accuracy": 0.8618566989898682, "num_tokens": 63927131.0, "step": 1680 }, { "epoch": 0.2138404783106475, "ewc_loss": 0.03076171875, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.7179718017578125e-05, "grad_norm": 2.514127731323242, "learning_rate": 7.12166172106825e-06, "loss": 0.5818, "mean_token_accuracy": 0.8195395469665527, "num_tokens": 63962362.0, "step": 1681 }, { "epoch": 0.213967688589238, "ewc_loss": 0.029296875, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 2.3207430839538574, "learning_rate": 7.125900805426028e-06, "loss": 0.4561, "mean_token_accuracy": 0.8599961996078491, "num_tokens": 63997189.0, "step": 1682 }, { "epoch": 0.2140948988678285, "ewc_loss": 0.031005859375, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.501195192337036, "learning_rate": 7.130139889783807e-06, "loss": 0.4958, "mean_token_accuracy": 0.8470444679260254, "num_tokens": 64032763.0, "step": 1683 }, { "epoch": 0.21422210914641904, "ewc_loss": 0.029052734375, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.5391578674316406e-05, "grad_norm": 2.3044779300689697, "learning_rate": 7.134378974141586e-06, "loss": 0.4885, "mean_token_accuracy": 0.8477697372436523, "num_tokens": 64070176.0, "step": 1684 }, { "epoch": 0.21434931942500954, "ewc_loss": 0.031005859375, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.521352767944336, "learning_rate": 7.138618058499365e-06, "loss": 0.4909, "mean_token_accuracy": 0.8508887887001038, "num_tokens": 64105632.0, "step": 1685 }, { "epoch": 0.21447652970360004, "ewc_loss": 0.02880859375, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.5153160095214844e-05, "grad_norm": 2.279165506362915, "learning_rate": 7.1428571428571436e-06, "loss": 0.4609, "mean_token_accuracy": 0.8591905832290649, "num_tokens": 64148729.0, "step": 1686 }, { "epoch": 0.21460373998219057, "ewc_loss": 0.03125, "ewc_loss_diag": 3.600120544433594e-05, "ewc_loss_parallel": 2.765655517578125e-05, "grad_norm": 2.533698797225952, "learning_rate": 7.1470962272149224e-06, "loss": 0.4886, "mean_token_accuracy": 0.8499584197998047, "num_tokens": 64189466.0, "step": 1687 }, { "epoch": 0.21473095026078107, "ewc_loss": 0.0291748046875, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.2795979976654053, "learning_rate": 7.151335311572701e-06, "loss": 0.4645, "mean_token_accuracy": 0.857225775718689, "num_tokens": 64226325.0, "step": 1688 }, { "epoch": 0.21485816053937157, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.8252601623535156e-05, "grad_norm": 2.5670254230499268, "learning_rate": 7.155574395930479e-06, "loss": 0.5094, "mean_token_accuracy": 0.8406457901000977, "num_tokens": 64262178.0, "step": 1689 }, { "epoch": 0.2149853708179621, "ewc_loss": 0.0291748046875, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.279956817626953, "learning_rate": 7.159813480288258e-06, "loss": 0.4848, "mean_token_accuracy": 0.8476004004478455, "num_tokens": 64299933.0, "step": 1690 }, { "epoch": 0.2151125810965526, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.849102020263672e-05, "grad_norm": 2.592475414276123, "learning_rate": 7.164052564646037e-06, "loss": 0.4522, "mean_token_accuracy": 0.8575311899185181, "num_tokens": 64335836.0, "step": 1691 }, { "epoch": 0.21523979137514312, "ewc_loss": 0.029296875, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.562999725341797e-05, "grad_norm": 2.2686965465545654, "learning_rate": 7.168291649003816e-06, "loss": 0.4784, "mean_token_accuracy": 0.8562178611755371, "num_tokens": 64377190.0, "step": 1692 }, { "epoch": 0.21536700165373363, "ewc_loss": 0.03271484375, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 2.6283183097839355, "learning_rate": 7.172530733361595e-06, "loss": 0.5273, "mean_token_accuracy": 0.83948814868927, "num_tokens": 64414693.0, "step": 1693 }, { "epoch": 0.21549421193232413, "ewc_loss": 0.0302734375, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.658367156982422e-05, "grad_norm": 2.2827515602111816, "learning_rate": 7.176769817719374e-06, "loss": 0.4406, "mean_token_accuracy": 0.8701292276382446, "num_tokens": 64454256.0, "step": 1694 }, { "epoch": 0.21562142221091465, "ewc_loss": 0.032958984375, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.944469451904297e-05, "grad_norm": 2.661675214767456, "learning_rate": 7.181008902077153e-06, "loss": 0.497, "mean_token_accuracy": 0.8510273694992065, "num_tokens": 64491513.0, "step": 1695 }, { "epoch": 0.21574863248950515, "ewc_loss": 0.0311279296875, "ewc_loss_diag": 3.647804260253906e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.321880578994751, "learning_rate": 7.1852479864349316e-06, "loss": 0.515, "mean_token_accuracy": 0.8458344340324402, "num_tokens": 64531349.0, "step": 1696 }, { "epoch": 0.21587584276809565, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.647804260253906e-05, "ewc_loss_parallel": 2.956390380859375e-05, "grad_norm": 2.665999174118042, "learning_rate": 7.189487070792709e-06, "loss": 0.4966, "mean_token_accuracy": 0.8509142398834229, "num_tokens": 64568007.0, "step": 1697 }, { "epoch": 0.21600305304668618, "ewc_loss": 0.03125, "ewc_loss_diag": 3.647804260253906e-05, "ewc_loss_parallel": 2.753734588623047e-05, "grad_norm": 2.3464019298553467, "learning_rate": 7.193726155150488e-06, "loss": 0.4973, "mean_token_accuracy": 0.8487459421157837, "num_tokens": 64610287.0, "step": 1698 }, { "epoch": 0.21613026332527668, "ewc_loss": 0.032958984375, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.6449711322784424, "learning_rate": 7.1979652395082665e-06, "loss": 0.5005, "mean_token_accuracy": 0.8482086062431335, "num_tokens": 64648757.0, "step": 1699 }, { "epoch": 0.21625747360386718, "ewc_loss": 0.0311279296875, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.356687545776367, "learning_rate": 7.202204323866045e-06, "loss": 0.4973, "mean_token_accuracy": 0.8494628667831421, "num_tokens": 64690137.0, "step": 1700 }, { "epoch": 0.2163846838824577, "ewc_loss": 0.03271484375, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 2.619159460067749, "learning_rate": 7.206443408223824e-06, "loss": 0.4962, "mean_token_accuracy": 0.8464686870574951, "num_tokens": 64730218.0, "step": 1701 }, { "epoch": 0.2165118941610482, "ewc_loss": 0.03125, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.753734588623047e-05, "grad_norm": 2.3810884952545166, "learning_rate": 7.210682492581603e-06, "loss": 0.4737, "mean_token_accuracy": 0.8536660671234131, "num_tokens": 64762213.0, "step": 1702 }, { "epoch": 0.2166391044396387, "ewc_loss": 0.0322265625, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.86102294921875e-05, "grad_norm": 2.5766401290893555, "learning_rate": 7.214921576939382e-06, "loss": 0.5164, "mean_token_accuracy": 0.8379622101783752, "num_tokens": 64802471.0, "step": 1703 }, { "epoch": 0.21676631471822924, "ewc_loss": 0.0308837890625, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.7179718017578125e-05, "grad_norm": 2.377542734146118, "learning_rate": 7.21916066129716e-06, "loss": 0.4627, "mean_token_accuracy": 0.8566747903823853, "num_tokens": 64838139.0, "step": 1704 }, { "epoch": 0.21689352499681974, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.543691396713257, "learning_rate": 7.223399745654939e-06, "loss": 0.5037, "mean_token_accuracy": 0.8449272513389587, "num_tokens": 64877721.0, "step": 1705 }, { "epoch": 0.21702073527541024, "ewc_loss": 0.030517578125, "ewc_loss_diag": 3.62396240234375e-05, "ewc_loss_parallel": 2.682209014892578e-05, "grad_norm": 2.368238687515259, "learning_rate": 7.227638830012718e-06, "loss": 0.5191, "mean_token_accuracy": 0.8396594524383545, "num_tokens": 64916315.0, "step": 1706 }, { "epoch": 0.21714794555400077, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.7894973754882812e-05, "grad_norm": 2.5261666774749756, "learning_rate": 7.231877914370497e-06, "loss": 0.4746, "mean_token_accuracy": 0.8566188812255859, "num_tokens": 64951303.0, "step": 1707 }, { "epoch": 0.21727515583259127, "ewc_loss": 0.0299072265625, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.6226043701171875e-05, "grad_norm": 2.3638288974761963, "learning_rate": 7.236116998728276e-06, "loss": 0.5294, "mean_token_accuracy": 0.8371508121490479, "num_tokens": 64984677.0, "step": 1708 }, { "epoch": 0.21740236611118177, "ewc_loss": 0.031494140625, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.777576446533203e-05, "grad_norm": 2.4902820587158203, "learning_rate": 7.2403560830860545e-06, "loss": 0.4222, "mean_token_accuracy": 0.8701052069664001, "num_tokens": 65021960.0, "step": 1709 }, { "epoch": 0.2175295763897723, "ewc_loss": 0.030029296875, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.6345252990722656e-05, "grad_norm": 2.342451333999634, "learning_rate": 7.244595167443833e-06, "loss": 0.5228, "mean_token_accuracy": 0.8401111364364624, "num_tokens": 65065576.0, "step": 1710 }, { "epoch": 0.2176567866683628, "ewc_loss": 0.031005859375, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 2.481203556060791, "learning_rate": 7.248834251801612e-06, "loss": 0.491, "mean_token_accuracy": 0.8536051511764526, "num_tokens": 65107188.0, "step": 1711 }, { "epoch": 0.2177839969469533, "ewc_loss": 0.0296630859375, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.5987625122070312e-05, "grad_norm": 2.32999324798584, "learning_rate": 7.2530733361593895e-06, "loss": 0.4599, "mean_token_accuracy": 0.8571075201034546, "num_tokens": 65145740.0, "step": 1712 }, { "epoch": 0.21791120722554383, "ewc_loss": 0.031005859375, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 2.5048320293426514, "learning_rate": 7.257312420517168e-06, "loss": 0.4494, "mean_token_accuracy": 0.8604253530502319, "num_tokens": 65176572.0, "step": 1713 }, { "epoch": 0.21803841750413433, "ewc_loss": 0.0291748046875, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.314692974090576, "learning_rate": 7.261551504874947e-06, "loss": 0.4758, "mean_token_accuracy": 0.8495520949363708, "num_tokens": 65214839.0, "step": 1714 }, { "epoch": 0.21816562778272486, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.7894973754882812e-05, "grad_norm": 2.5164127349853516, "learning_rate": 7.265790589232726e-06, "loss": 0.4737, "mean_token_accuracy": 0.8549488186836243, "num_tokens": 65251438.0, "step": 1715 }, { "epoch": 0.21829283806131536, "ewc_loss": 0.0291748046875, "ewc_loss_diag": 3.6716461181640625e-05, "ewc_loss_parallel": 2.5510787963867188e-05, "grad_norm": 2.2945809364318848, "learning_rate": 7.270029673590505e-06, "loss": 0.5078, "mean_token_accuracy": 0.8443611860275269, "num_tokens": 65294256.0, "step": 1716 }, { "epoch": 0.21842004833990586, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.719329833984375e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.577359676361084, "learning_rate": 7.274268757948284e-06, "loss": 0.5576, "mean_token_accuracy": 0.8309982419013977, "num_tokens": 65332917.0, "step": 1717 }, { "epoch": 0.2185472586184964, "ewc_loss": 0.029541015625, "ewc_loss_diag": 3.719329833984375e-05, "ewc_loss_parallel": 2.574920654296875e-05, "grad_norm": 2.2889585494995117, "learning_rate": 7.278507842306063e-06, "loss": 0.4677, "mean_token_accuracy": 0.8543938398361206, "num_tokens": 65366678.0, "step": 1718 }, { "epoch": 0.2186744688970869, "ewc_loss": 0.0322265625, "ewc_loss_diag": 3.719329833984375e-05, "ewc_loss_parallel": 2.849102020263672e-05, "grad_norm": 2.6146576404571533, "learning_rate": 7.282746926663842e-06, "loss": 0.5585, "mean_token_accuracy": 0.8295032978057861, "num_tokens": 65404797.0, "step": 1719 }, { "epoch": 0.2188016791756774, "ewc_loss": 0.029541015625, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 2.574920654296875e-05, "grad_norm": 2.268632650375366, "learning_rate": 7.28698601102162e-06, "loss": 0.4682, "mean_token_accuracy": 0.8542028665542603, "num_tokens": 65441855.0, "step": 1720 }, { "epoch": 0.21892888945426792, "ewc_loss": 0.03369140625, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 2.9921531677246094e-05, "grad_norm": 2.6586387157440186, "learning_rate": 7.291225095379399e-06, "loss": 0.517, "mean_token_accuracy": 0.8452897071838379, "num_tokens": 65486895.0, "step": 1721 }, { "epoch": 0.21905609973285842, "ewc_loss": 0.0303955078125, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 2.658367156982422e-05, "grad_norm": 2.283259630203247, "learning_rate": 7.2954641797371775e-06, "loss": 0.496, "mean_token_accuracy": 0.8494555950164795, "num_tokens": 65524862.0, "step": 1722 }, { "epoch": 0.21918331001144892, "ewc_loss": 0.034423828125, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.7199904918670654, "learning_rate": 7.299703264094956e-06, "loss": 0.4948, "mean_token_accuracy": 0.8473896980285645, "num_tokens": 65561538.0, "step": 1723 }, { "epoch": 0.21931052029003945, "ewc_loss": 0.03125, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.326730489730835, "learning_rate": 7.303942348452735e-06, "loss": 0.5123, "mean_token_accuracy": 0.8422829508781433, "num_tokens": 65602376.0, "step": 1724 }, { "epoch": 0.21943773056862995, "ewc_loss": 0.03466796875, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.7370996475219727, "learning_rate": 7.308181432810514e-06, "loss": 0.4628, "mean_token_accuracy": 0.8547122478485107, "num_tokens": 65640817.0, "step": 1725 }, { "epoch": 0.21956494084722045, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 2.7894973754882812e-05, "grad_norm": 2.347944498062134, "learning_rate": 7.312420517168293e-06, "loss": 0.4884, "mean_token_accuracy": 0.8459470868110657, "num_tokens": 65679238.0, "step": 1726 }, { "epoch": 0.21969215112581097, "ewc_loss": 0.034912109375, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.746148109436035, "learning_rate": 7.31665960152607e-06, "loss": 0.5146, "mean_token_accuracy": 0.8390932083129883, "num_tokens": 65720141.0, "step": 1727 }, { "epoch": 0.21981936140440148, "ewc_loss": 0.032470703125, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 2.4005658626556396, "learning_rate": 7.320898685883849e-06, "loss": 0.4849, "mean_token_accuracy": 0.8495373725891113, "num_tokens": 65758511.0, "step": 1728 }, { "epoch": 0.21994657168299198, "ewc_loss": 0.03466796875, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.730208158493042, "learning_rate": 7.325137770241628e-06, "loss": 0.6084, "mean_token_accuracy": 0.817255973815918, "num_tokens": 65798485.0, "step": 1729 }, { "epoch": 0.2200737819615825, "ewc_loss": 0.032470703125, "ewc_loss_diag": 3.7670135498046875e-05, "ewc_loss_parallel": 2.8848648071289062e-05, "grad_norm": 2.41245436668396, "learning_rate": 7.329376854599407e-06, "loss": 0.5129, "mean_token_accuracy": 0.8426375389099121, "num_tokens": 65840414.0, "step": 1730 }, { "epoch": 0.220200992240173, "ewc_loss": 0.0341796875, "ewc_loss_diag": 3.790855407714844e-05, "ewc_loss_parallel": 3.039836883544922e-05, "grad_norm": 2.6966450214385986, "learning_rate": 7.333615938957186e-06, "loss": 0.493, "mean_token_accuracy": 0.8502610325813293, "num_tokens": 65878425.0, "step": 1731 }, { "epoch": 0.2203282025187635, "ewc_loss": 0.032470703125, "ewc_loss_diag": 3.790855407714844e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 2.4277243614196777, "learning_rate": 7.337855023314965e-06, "loss": 0.4957, "mean_token_accuracy": 0.8458577394485474, "num_tokens": 65916958.0, "step": 1732 }, { "epoch": 0.22045541279735403, "ewc_loss": 0.033447265625, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.956390380859375e-05, "grad_norm": 2.64628005027771, "learning_rate": 7.3420941076727435e-06, "loss": 0.548, "mean_token_accuracy": 0.8306445479393005, "num_tokens": 65958511.0, "step": 1733 }, { "epoch": 0.22058262307594453, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.814697265625e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.4362289905548096, "learning_rate": 7.346333192030522e-06, "loss": 0.4954, "mean_token_accuracy": 0.8502975106239319, "num_tokens": 65994735.0, "step": 1734 }, { "epoch": 0.22070983335453503, "ewc_loss": 0.03271484375, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.6011385917663574, "learning_rate": 7.3505722763883004e-06, "loss": 0.5514, "mean_token_accuracy": 0.8302323818206787, "num_tokens": 66036017.0, "step": 1735 }, { "epoch": 0.22083704363312556, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.8014183044433594e-05, "grad_norm": 2.416081428527832, "learning_rate": 7.354811360746079e-06, "loss": 0.4987, "mean_token_accuracy": 0.8450861573219299, "num_tokens": 66073203.0, "step": 1736 }, { "epoch": 0.22096425391171606, "ewc_loss": 0.032470703125, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.86102294921875e-05, "grad_norm": 2.568126916885376, "learning_rate": 7.359050445103858e-06, "loss": 0.5169, "mean_token_accuracy": 0.8425949811935425, "num_tokens": 66109623.0, "step": 1737 }, { "epoch": 0.22109146419030656, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.777576446533203e-05, "grad_norm": 2.403486490249634, "learning_rate": 7.363289529461637e-06, "loss": 0.4919, "mean_token_accuracy": 0.8491443991661072, "num_tokens": 66148426.0, "step": 1738 }, { "epoch": 0.2212186744688971, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.8252601623535156e-05, "grad_norm": 2.532386541366577, "learning_rate": 7.367528613819416e-06, "loss": 0.4508, "mean_token_accuracy": 0.8613927364349365, "num_tokens": 66189164.0, "step": 1739 }, { "epoch": 0.2213458847474876, "ewc_loss": 0.03125, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.7418136596679688e-05, "grad_norm": 2.388596296310425, "learning_rate": 7.371767698177195e-06, "loss": 0.4676, "mean_token_accuracy": 0.8544189929962158, "num_tokens": 66223523.0, "step": 1740 }, { "epoch": 0.22147309502607812, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.8014183044433594e-05, "grad_norm": 2.526691436767578, "learning_rate": 7.376006782534974e-06, "loss": 0.4568, "mean_token_accuracy": 0.8597429394721985, "num_tokens": 66256819.0, "step": 1741 }, { "epoch": 0.22160030530466862, "ewc_loss": 0.03076171875, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.6941299438476562e-05, "grad_norm": 2.379037618637085, "learning_rate": 7.380245866892751e-06, "loss": 0.4856, "mean_token_accuracy": 0.8497291803359985, "num_tokens": 66291619.0, "step": 1742 }, { "epoch": 0.22172751558325912, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.5154001712799072, "learning_rate": 7.38448495125053e-06, "loss": 0.49, "mean_token_accuracy": 0.8497094511985779, "num_tokens": 66334034.0, "step": 1743 }, { "epoch": 0.22185472586184965, "ewc_loss": 0.0302734375, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.6464462280273438e-05, "grad_norm": 2.3639137744903564, "learning_rate": 7.388724035608309e-06, "loss": 0.4763, "mean_token_accuracy": 0.856167197227478, "num_tokens": 66368786.0, "step": 1744 }, { "epoch": 0.22198193614044015, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.838539123535156e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.534710645675659, "learning_rate": 7.392963119966088e-06, "loss": 0.4705, "mean_token_accuracy": 0.8540331721305847, "num_tokens": 66400034.0, "step": 1745 }, { "epoch": 0.22210914641903065, "ewc_loss": 0.0302734375, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 2.6345252990722656e-05, "grad_norm": 2.3478519916534424, "learning_rate": 7.3972022043238665e-06, "loss": 0.5043, "mean_token_accuracy": 0.8466057777404785, "num_tokens": 66436954.0, "step": 1746 }, { "epoch": 0.22223635669762118, "ewc_loss": 0.0322265625, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 2.849102020263672e-05, "grad_norm": 2.5410244464874268, "learning_rate": 7.401441288681645e-06, "loss": 0.472, "mean_token_accuracy": 0.855483889579773, "num_tokens": 66476360.0, "step": 1747 }, { "epoch": 0.22236356697621168, "ewc_loss": 0.02978515625, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 2.586841583251953e-05, "grad_norm": 2.3194150924682617, "learning_rate": 7.405680373039424e-06, "loss": 0.4485, "mean_token_accuracy": 0.8635915517807007, "num_tokens": 66516205.0, "step": 1748 }, { "epoch": 0.22249077725480218, "ewc_loss": 0.032470703125, "ewc_loss_diag": 3.8623809814453125e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 2.5996882915496826, "learning_rate": 7.409919457397203e-06, "loss": 0.5141, "mean_token_accuracy": 0.8401703238487244, "num_tokens": 66548732.0, "step": 1749 }, { "epoch": 0.2226179875333927, "ewc_loss": 0.0299072265625, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 2.5987625122070312e-05, "grad_norm": 2.3157498836517334, "learning_rate": 7.414158541754981e-06, "loss": 0.5508, "mean_token_accuracy": 0.8281025886535645, "num_tokens": 66587888.0, "step": 1750 }, { "epoch": 0.2227451978119832, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.6541130542755127, "learning_rate": 7.41839762611276e-06, "loss": 0.4892, "mean_token_accuracy": 0.8590534925460815, "num_tokens": 66617846.0, "step": 1751 }, { "epoch": 0.2228724080905737, "ewc_loss": 0.0299072265625, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 2.5987625122070312e-05, "grad_norm": 2.29301118850708, "learning_rate": 7.422636710470539e-06, "loss": 0.5002, "mean_token_accuracy": 0.8455027937889099, "num_tokens": 66659703.0, "step": 1752 }, { "epoch": 0.22299961836916424, "ewc_loss": 0.0341796875, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.716341733932495, "learning_rate": 7.426875794828318e-06, "loss": 0.5171, "mean_token_accuracy": 0.8460773229598999, "num_tokens": 66697604.0, "step": 1753 }, { "epoch": 0.22312682864775474, "ewc_loss": 0.03125, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 2.312587022781372, "learning_rate": 7.431114879186097e-06, "loss": 0.5425, "mean_token_accuracy": 0.8349050283432007, "num_tokens": 66736839.0, "step": 1754 }, { "epoch": 0.22325403892634524, "ewc_loss": 0.035400390625, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.7648379802703857, "learning_rate": 7.435353963543876e-06, "loss": 0.5083, "mean_token_accuracy": 0.8462651968002319, "num_tokens": 66777477.0, "step": 1755 }, { "epoch": 0.22338124920493577, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 2.8014183044433594e-05, "grad_norm": 2.3328166007995605, "learning_rate": 7.4395930479016545e-06, "loss": 0.452, "mean_token_accuracy": 0.8583539724349976, "num_tokens": 66816294.0, "step": 1756 }, { "epoch": 0.22350845948352627, "ewc_loss": 0.036376953125, "ewc_loss_diag": 3.886222839355469e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.8235199451446533, "learning_rate": 7.443832132259432e-06, "loss": 0.5217, "mean_token_accuracy": 0.8402796983718872, "num_tokens": 66852740.0, "step": 1757 }, { "epoch": 0.22363566976211677, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 2.4057726860046387, "learning_rate": 7.4480712166172105e-06, "loss": 0.4959, "mean_token_accuracy": 0.8464776873588562, "num_tokens": 66890511.0, "step": 1758 }, { "epoch": 0.2237628800407073, "ewc_loss": 0.036376953125, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.8162121772766113, "learning_rate": 7.4523103009749894e-06, "loss": 0.4604, "mean_token_accuracy": 0.8575432896614075, "num_tokens": 66924724.0, "step": 1759 }, { "epoch": 0.2238900903192978, "ewc_loss": 0.0341796875, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.453864812850952, "learning_rate": 7.456549385332768e-06, "loss": 0.539, "mean_token_accuracy": 0.837257981300354, "num_tokens": 66964102.0, "step": 1760 }, { "epoch": 0.2240173005978883, "ewc_loss": 0.035400390625, "ewc_loss_diag": 3.910064697265625e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.759692907333374, "learning_rate": 7.460788469690547e-06, "loss": 0.4221, "mean_token_accuracy": 0.8695293068885803, "num_tokens": 67003234.0, "step": 1761 }, { "epoch": 0.22414451087647883, "ewc_loss": 0.033935546875, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.4815926551818848, "learning_rate": 7.465027554048326e-06, "loss": 0.5773, "mean_token_accuracy": 0.8224877119064331, "num_tokens": 67045952.0, "step": 1762 }, { "epoch": 0.22427172115506933, "ewc_loss": 0.03515625, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.6995604038238525, "learning_rate": 7.469266638406105e-06, "loss": 0.5622, "mean_token_accuracy": 0.8308171033859253, "num_tokens": 67090169.0, "step": 1763 }, { "epoch": 0.22439893143365983, "ewc_loss": 0.033447265625, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.956390380859375e-05, "grad_norm": 2.481863498687744, "learning_rate": 7.473505722763884e-06, "loss": 0.4803, "mean_token_accuracy": 0.8513428568840027, "num_tokens": 67125075.0, "step": 1764 }, { "epoch": 0.22452614171225035, "ewc_loss": 0.033935546875, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.9921531677246094e-05, "grad_norm": 2.656686544418335, "learning_rate": 7.477744807121662e-06, "loss": 0.5252, "mean_token_accuracy": 0.8407257199287415, "num_tokens": 67162892.0, "step": 1765 }, { "epoch": 0.22465335199084085, "ewc_loss": 0.032958984375, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.475494146347046, "learning_rate": 7.481983891479441e-06, "loss": 0.5069, "mean_token_accuracy": 0.8455229997634888, "num_tokens": 67198324.0, "step": 1766 }, { "epoch": 0.22478056226943138, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.603696823120117, "learning_rate": 7.48622297583722e-06, "loss": 0.4896, "mean_token_accuracy": 0.8465265035629272, "num_tokens": 67239548.0, "step": 1767 }, { "epoch": 0.22490777254802188, "ewc_loss": 0.031982421875, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.4488470554351807, "learning_rate": 7.4904620601949985e-06, "loss": 0.4526, "mean_token_accuracy": 0.8572245836257935, "num_tokens": 67273373.0, "step": 1768 }, { "epoch": 0.22503498282661238, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.5921361446380615, "learning_rate": 7.4947011445527774e-06, "loss": 0.5332, "mean_token_accuracy": 0.8383098840713501, "num_tokens": 67307074.0, "step": 1769 }, { "epoch": 0.2251621931052029, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.777576446533203e-05, "grad_norm": 2.422757387161255, "learning_rate": 7.498940228910556e-06, "loss": 0.5095, "mean_token_accuracy": 0.841423511505127, "num_tokens": 67347905.0, "step": 1770 }, { "epoch": 0.2252894033837934, "ewc_loss": 0.03271484375, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.8848648071289062e-05, "grad_norm": 2.5785324573516846, "learning_rate": 7.503179313268335e-06, "loss": 0.4814, "mean_token_accuracy": 0.8488345146179199, "num_tokens": 67382594.0, "step": 1771 }, { "epoch": 0.2254166136623839, "ewc_loss": 0.03173828125, "ewc_loss_diag": 3.933906555175781e-05, "ewc_loss_parallel": 2.765655517578125e-05, "grad_norm": 2.394843578338623, "learning_rate": 7.507418397626114e-06, "loss": 0.5264, "mean_token_accuracy": 0.83692467212677, "num_tokens": 67422867.0, "step": 1772 }, { "epoch": 0.22554382394097444, "ewc_loss": 0.032958984375, "ewc_loss_diag": 3.9577484130859375e-05, "ewc_loss_parallel": 2.9087066650390625e-05, "grad_norm": 2.577397346496582, "learning_rate": 7.511657481983891e-06, "loss": 0.5138, "mean_token_accuracy": 0.8389500379562378, "num_tokens": 67463524.0, "step": 1773 }, { "epoch": 0.22567103421956494, "ewc_loss": 0.03125, "ewc_loss_diag": 3.9577484130859375e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 2.3898441791534424, "learning_rate": 7.51589656634167e-06, "loss": 0.4689, "mean_token_accuracy": 0.848846971988678, "num_tokens": 67499224.0, "step": 1774 }, { "epoch": 0.22579824449815544, "ewc_loss": 0.033203125, "ewc_loss_diag": 3.9577484130859375e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.5931894779205322, "learning_rate": 7.520135650699449e-06, "loss": 0.5001, "mean_token_accuracy": 0.8441855311393738, "num_tokens": 67536745.0, "step": 1775 }, { "epoch": 0.22592545477674597, "ewc_loss": 0.03125, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 2.7179718017578125e-05, "grad_norm": 2.3748526573181152, "learning_rate": 7.524374735057228e-06, "loss": 0.5233, "mean_token_accuracy": 0.8429598808288574, "num_tokens": 67572725.0, "step": 1776 }, { "epoch": 0.22605266505533647, "ewc_loss": 0.033447265625, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 2.944469451904297e-05, "grad_norm": 2.6166749000549316, "learning_rate": 7.528613819415007e-06, "loss": 0.4803, "mean_token_accuracy": 0.8555351495742798, "num_tokens": 67607588.0, "step": 1777 }, { "epoch": 0.22617987533392697, "ewc_loss": 0.03125, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 2.7179718017578125e-05, "grad_norm": 2.360123872756958, "learning_rate": 7.532852903772786e-06, "loss": 0.5126, "mean_token_accuracy": 0.8420961499214172, "num_tokens": 67643064.0, "step": 1778 }, { "epoch": 0.2263070856125175, "ewc_loss": 0.03369140625, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 2.9802322387695312e-05, "grad_norm": 2.653169631958008, "learning_rate": 7.537091988130565e-06, "loss": 0.5167, "mean_token_accuracy": 0.8382848501205444, "num_tokens": 67677985.0, "step": 1779 }, { "epoch": 0.226434295891108, "ewc_loss": 0.03125, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 2.7298927307128906e-05, "grad_norm": 2.346500873565674, "learning_rate": 7.541331072488343e-06, "loss": 0.5024, "mean_token_accuracy": 0.8459811210632324, "num_tokens": 67713814.0, "step": 1780 }, { "epoch": 0.2265615061696985, "ewc_loss": 0.034423828125, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.6841700077056885, "learning_rate": 7.5455701568461215e-06, "loss": 0.5048, "mean_token_accuracy": 0.84478759765625, "num_tokens": 67752806.0, "step": 1781 }, { "epoch": 0.22668871644828903, "ewc_loss": 0.03173828125, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 2.777576446533203e-05, "grad_norm": 2.3372201919555664, "learning_rate": 7.5498092412039e-06, "loss": 0.4872, "mean_token_accuracy": 0.8509948253631592, "num_tokens": 67792804.0, "step": 1782 }, { "epoch": 0.22681592672687953, "ewc_loss": 0.034912109375, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.73079776763916, "learning_rate": 7.554048325561679e-06, "loss": 0.453, "mean_token_accuracy": 0.8603922128677368, "num_tokens": 67831361.0, "step": 1783 }, { "epoch": 0.22694313700547003, "ewc_loss": 0.031982421875, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 2.8014183044433594e-05, "grad_norm": 2.3667349815368652, "learning_rate": 7.558287409919458e-06, "loss": 0.5401, "mean_token_accuracy": 0.83990478515625, "num_tokens": 67868981.0, "step": 1784 }, { "epoch": 0.22707034728406056, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.00543212890625e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.774836301803589, "learning_rate": 7.562526494277237e-06, "loss": 0.5219, "mean_token_accuracy": 0.8383294343948364, "num_tokens": 67905501.0, "step": 1785 }, { "epoch": 0.22719755756265106, "ewc_loss": 0.032470703125, "ewc_loss_diag": 4.0531158447265625e-05, "ewc_loss_parallel": 2.8371810913085938e-05, "grad_norm": 2.3693037033081055, "learning_rate": 7.566765578635016e-06, "loss": 0.4875, "mean_token_accuracy": 0.8521802425384521, "num_tokens": 67946001.0, "step": 1786 }, { "epoch": 0.22732476784124156, "ewc_loss": 0.036376953125, "ewc_loss_diag": 4.0531158447265625e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.7996513843536377, "learning_rate": 7.571004662992795e-06, "loss": 0.514, "mean_token_accuracy": 0.8468915224075317, "num_tokens": 67983868.0, "step": 1787 }, { "epoch": 0.2274519781198321, "ewc_loss": 0.033203125, "ewc_loss_diag": 4.0531158447265625e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 2.4147167205810547, "learning_rate": 7.575243747350573e-06, "loss": 0.5671, "mean_token_accuracy": 0.8278622031211853, "num_tokens": 68023664.0, "step": 1788 }, { "epoch": 0.2275791883984226, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.0531158447265625e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.7970762252807617, "learning_rate": 7.579482831708351e-06, "loss": 0.5084, "mean_token_accuracy": 0.8479934930801392, "num_tokens": 68063950.0, "step": 1789 }, { "epoch": 0.2277063986770131, "ewc_loss": 0.0341796875, "ewc_loss_diag": 4.0531158447265625e-05, "ewc_loss_parallel": 3.0159950256347656e-05, "grad_norm": 2.4571542739868164, "learning_rate": 7.58372191606613e-06, "loss": 0.5027, "mean_token_accuracy": 0.8478244543075562, "num_tokens": 68097812.0, "step": 1790 }, { "epoch": 0.22783360895560362, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.0531158447265625e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.770134210586548, "learning_rate": 7.587961000423909e-06, "loss": 0.5061, "mean_token_accuracy": 0.8460695147514343, "num_tokens": 68135219.0, "step": 1791 }, { "epoch": 0.22796081923419412, "ewc_loss": 0.034423828125, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 3.039836883544922e-05, "grad_norm": 2.4895336627960205, "learning_rate": 7.5922000847816875e-06, "loss": 0.5576, "mean_token_accuracy": 0.8340162634849548, "num_tokens": 68169757.0, "step": 1792 }, { "epoch": 0.22808802951278465, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.7200918197631836, "learning_rate": 7.5964391691394664e-06, "loss": 0.5028, "mean_token_accuracy": 0.846853494644165, "num_tokens": 68204807.0, "step": 1793 }, { "epoch": 0.22821523979137515, "ewc_loss": 0.033935546875, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 2.9802322387695312e-05, "grad_norm": 2.471402645111084, "learning_rate": 7.600678253497245e-06, "loss": 0.4965, "mean_token_accuracy": 0.851520299911499, "num_tokens": 68245037.0, "step": 1794 }, { "epoch": 0.22834245006996565, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.673891544342041, "learning_rate": 7.604917337855023e-06, "loss": 0.4693, "mean_token_accuracy": 0.8557121753692627, "num_tokens": 68281938.0, "step": 1795 }, { "epoch": 0.22846966034855618, "ewc_loss": 0.033447265625, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.458047866821289, "learning_rate": 7.609156422212802e-06, "loss": 0.4506, "mean_token_accuracy": 0.8608039617538452, "num_tokens": 68319363.0, "step": 1796 }, { "epoch": 0.22859687062714668, "ewc_loss": 0.034423828125, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.652418375015259, "learning_rate": 7.613395506570581e-06, "loss": 0.4758, "mean_token_accuracy": 0.8542373180389404, "num_tokens": 68357428.0, "step": 1797 }, { "epoch": 0.22872408090573718, "ewc_loss": 0.032958984375, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 2.8848648071289062e-05, "grad_norm": 2.4632983207702637, "learning_rate": 7.61763459092836e-06, "loss": 0.4795, "mean_token_accuracy": 0.853173017501831, "num_tokens": 68390558.0, "step": 1798 }, { "epoch": 0.2288512911843277, "ewc_loss": 0.033935546875, "ewc_loss_diag": 4.076957702636719e-05, "ewc_loss_parallel": 2.9921531677246094e-05, "grad_norm": 2.6245932579040527, "learning_rate": 7.621873675286139e-06, "loss": 0.5155, "mean_token_accuracy": 0.8432433605194092, "num_tokens": 68432250.0, "step": 1799 }, { "epoch": 0.2289785014629182, "ewc_loss": 0.032470703125, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 2.8371810913085938e-05, "grad_norm": 2.440713882446289, "learning_rate": 7.626112759643918e-06, "loss": 0.5145, "mean_token_accuracy": 0.8413722515106201, "num_tokens": 68469427.0, "step": 1800 }, { "epoch": 0.2291057117415087, "ewc_loss": 0.03369140625, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 2.956390380859375e-05, "grad_norm": 2.620023250579834, "learning_rate": 7.630351844001696e-06, "loss": 0.4936, "mean_token_accuracy": 0.8490241169929504, "num_tokens": 68503442.0, "step": 1801 }, { "epoch": 0.22923292202009923, "ewc_loss": 0.0322265625, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 2.8252601623535156e-05, "grad_norm": 2.417273998260498, "learning_rate": 7.634590928359475e-06, "loss": 0.4776, "mean_token_accuracy": 0.8555459380149841, "num_tokens": 68540836.0, "step": 1802 }, { "epoch": 0.22936013229868973, "ewc_loss": 0.03369140625, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 2.968311309814453e-05, "grad_norm": 2.6184933185577393, "learning_rate": 7.638830012717254e-06, "loss": 0.5136, "mean_token_accuracy": 0.8408993482589722, "num_tokens": 68581118.0, "step": 1803 }, { "epoch": 0.22948734257728023, "ewc_loss": 0.031982421875, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 2.7894973754882812e-05, "grad_norm": 2.396287202835083, "learning_rate": 7.643069097075032e-06, "loss": 0.4807, "mean_token_accuracy": 0.8507009744644165, "num_tokens": 68620988.0, "step": 1804 }, { "epoch": 0.22961455285587076, "ewc_loss": 0.0341796875, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.6396167278289795, "learning_rate": 7.647308181432811e-06, "loss": 0.4563, "mean_token_accuracy": 0.8575798273086548, "num_tokens": 68662141.0, "step": 1805 }, { "epoch": 0.22974176313446126, "ewc_loss": 0.031982421875, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 2.777576446533203e-05, "grad_norm": 2.452364683151245, "learning_rate": 7.65154726579059e-06, "loss": 0.5401, "mean_token_accuracy": 0.8353431224822998, "num_tokens": 68700922.0, "step": 1806 }, { "epoch": 0.22986897341305176, "ewc_loss": 0.033935546875, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 2.9921531677246094e-05, "grad_norm": 2.6288111209869385, "learning_rate": 7.655786350148369e-06, "loss": 0.5182, "mean_token_accuracy": 0.842011570930481, "num_tokens": 68741137.0, "step": 1807 }, { "epoch": 0.2299961836916423, "ewc_loss": 0.031982421875, "ewc_loss_diag": 4.100799560546875e-05, "ewc_loss_parallel": 2.777576446533203e-05, "grad_norm": 2.3791987895965576, "learning_rate": 7.660025434506148e-06, "loss": 0.5172, "mean_token_accuracy": 0.8423998355865479, "num_tokens": 68785679.0, "step": 1808 }, { "epoch": 0.2301233939702328, "ewc_loss": 0.0341796875, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 3.0159950256347656e-05, "grad_norm": 2.64575457572937, "learning_rate": 7.664264518863927e-06, "loss": 0.501, "mean_token_accuracy": 0.8478381037712097, "num_tokens": 68823482.0, "step": 1809 }, { "epoch": 0.2302506042488233, "ewc_loss": 0.031982421875, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 2.777576446533203e-05, "grad_norm": 2.392975330352783, "learning_rate": 7.668503603221706e-06, "loss": 0.5556, "mean_token_accuracy": 0.8270413875579834, "num_tokens": 68863854.0, "step": 1810 }, { "epoch": 0.23037781452741382, "ewc_loss": 0.034912109375, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.668843984603882, "learning_rate": 7.672742687579483e-06, "loss": 0.4531, "mean_token_accuracy": 0.8590191602706909, "num_tokens": 68901341.0, "step": 1811 }, { "epoch": 0.23050502480600432, "ewc_loss": 0.031982421875, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 2.8014183044433594e-05, "grad_norm": 2.384253978729248, "learning_rate": 7.676981771937262e-06, "loss": 0.4308, "mean_token_accuracy": 0.8630724549293518, "num_tokens": 68931774.0, "step": 1812 }, { "epoch": 0.23063223508459482, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.69608998298645, "learning_rate": 7.68122085629504e-06, "loss": 0.4727, "mean_token_accuracy": 0.8544639348983765, "num_tokens": 68972114.0, "step": 1813 }, { "epoch": 0.23075944536318535, "ewc_loss": 0.0322265625, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 2.8252601623535156e-05, "grad_norm": 2.381324529647827, "learning_rate": 7.68545994065282e-06, "loss": 0.4925, "mean_token_accuracy": 0.8462672233581543, "num_tokens": 69007272.0, "step": 1814 }, { "epoch": 0.23088665564177585, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.124641418457031e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.7273476123809814, "learning_rate": 7.689699025010598e-06, "loss": 0.491, "mean_token_accuracy": 0.8498340249061584, "num_tokens": 69050011.0, "step": 1815 }, { "epoch": 0.23101386592036638, "ewc_loss": 0.0322265625, "ewc_loss_diag": 4.1484832763671875e-05, "ewc_loss_parallel": 2.8133392333984375e-05, "grad_norm": 2.3850655555725098, "learning_rate": 7.693938109368377e-06, "loss": 0.5085, "mean_token_accuracy": 0.8411867618560791, "num_tokens": 69087914.0, "step": 1816 }, { "epoch": 0.23114107619895688, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.1484832763671875e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.7748618125915527, "learning_rate": 7.698177193726156e-06, "loss": 0.5029, "mean_token_accuracy": 0.8451459407806396, "num_tokens": 69125301.0, "step": 1817 }, { "epoch": 0.23126828647754738, "ewc_loss": 0.032958984375, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 2.401838541030884, "learning_rate": 7.702416278083933e-06, "loss": 0.5125, "mean_token_accuracy": 0.8424204587936401, "num_tokens": 69159268.0, "step": 1818 }, { "epoch": 0.2313954967561379, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.7915165424346924, "learning_rate": 7.706655362441712e-06, "loss": 0.4976, "mean_token_accuracy": 0.8493187427520752, "num_tokens": 69202207.0, "step": 1819 }, { "epoch": 0.2315227070347284, "ewc_loss": 0.033203125, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 2.9087066650390625e-05, "grad_norm": 2.4247021675109863, "learning_rate": 7.710894446799491e-06, "loss": 0.4693, "mean_token_accuracy": 0.8532808423042297, "num_tokens": 69234757.0, "step": 1820 }, { "epoch": 0.2316499173133189, "ewc_loss": 0.037353515625, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.806922197341919, "learning_rate": 7.71513353115727e-06, "loss": 0.4775, "mean_token_accuracy": 0.8583965301513672, "num_tokens": 69274995.0, "step": 1821 }, { "epoch": 0.23177712759190944, "ewc_loss": 0.033935546875, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 2.9802322387695312e-05, "grad_norm": 2.440314292907715, "learning_rate": 7.719372615515049e-06, "loss": 0.5391, "mean_token_accuracy": 0.8351613879203796, "num_tokens": 69318706.0, "step": 1822 }, { "epoch": 0.23190433787049994, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.8040897846221924, "learning_rate": 7.723611699872828e-06, "loss": 0.4911, "mean_token_accuracy": 0.8492723107337952, "num_tokens": 69361895.0, "step": 1823 }, { "epoch": 0.23203154814909044, "ewc_loss": 0.03466796875, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.467524290084839, "learning_rate": 7.727850784230607e-06, "loss": 0.5056, "mean_token_accuracy": 0.8416314125061035, "num_tokens": 69400877.0, "step": 1824 }, { "epoch": 0.23215875842768097, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.790027379989624, "learning_rate": 7.732089868588386e-06, "loss": 0.5311, "mean_token_accuracy": 0.8418411612510681, "num_tokens": 69437875.0, "step": 1825 }, { "epoch": 0.23228596870627147, "ewc_loss": 0.03466796875, "ewc_loss_diag": 4.172325134277344e-05, "ewc_loss_parallel": 3.039836883544922e-05, "grad_norm": 2.458977460861206, "learning_rate": 7.736328952946165e-06, "loss": 0.4417, "mean_token_accuracy": 0.8654165863990784, "num_tokens": 69481770.0, "step": 1826 }, { "epoch": 0.23241317898486197, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.1961669921875e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.7835023403167725, "learning_rate": 7.740568037303943e-06, "loss": 0.4667, "mean_token_accuracy": 0.8539557456970215, "num_tokens": 69514643.0, "step": 1827 }, { "epoch": 0.2325403892634525, "ewc_loss": 0.03466796875, "ewc_loss_diag": 4.1961669921875e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.4799654483795166, "learning_rate": 7.744807121661722e-06, "loss": 0.4319, "mean_token_accuracy": 0.8698850870132446, "num_tokens": 69551906.0, "step": 1828 }, { "epoch": 0.232667599542043, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.1961669921875e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.756976842880249, "learning_rate": 7.749046206019501e-06, "loss": 0.506, "mean_token_accuracy": 0.8488266468048096, "num_tokens": 69593621.0, "step": 1829 }, { "epoch": 0.2327948098206335, "ewc_loss": 0.034423828125, "ewc_loss_diag": 4.1961669921875e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.483258008956909, "learning_rate": 7.75328529037728e-06, "loss": 0.4407, "mean_token_accuracy": 0.8640959858894348, "num_tokens": 69629586.0, "step": 1830 }, { "epoch": 0.23292202009922403, "ewc_loss": 0.035888671875, "ewc_loss_diag": 4.1961669921875e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.741234064102173, "learning_rate": 7.757524374735059e-06, "loss": 0.5457, "mean_token_accuracy": 0.8328002691268921, "num_tokens": 69667409.0, "step": 1831 }, { "epoch": 0.23304923037781453, "ewc_loss": 0.0341796875, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.4739227294921875, "learning_rate": 7.761763459092838e-06, "loss": 0.4509, "mean_token_accuracy": 0.8604638576507568, "num_tokens": 69704645.0, "step": 1832 }, { "epoch": 0.23317644065640503, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.7169604301452637, "learning_rate": 7.766002543450615e-06, "loss": 0.4818, "mean_token_accuracy": 0.8512699604034424, "num_tokens": 69742216.0, "step": 1833 }, { "epoch": 0.23330365093499555, "ewc_loss": 0.03369140625, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 2.956390380859375e-05, "grad_norm": 2.474730968475342, "learning_rate": 7.770241627808394e-06, "loss": 0.4435, "mean_token_accuracy": 0.8615831136703491, "num_tokens": 69779119.0, "step": 1834 }, { "epoch": 0.23343086121358606, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.6989219188690186, "learning_rate": 7.774480712166173e-06, "loss": 0.5215, "mean_token_accuracy": 0.840125560760498, "num_tokens": 69819980.0, "step": 1835 }, { "epoch": 0.23355807149217656, "ewc_loss": 0.033203125, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.4581141471862793, "learning_rate": 7.778719796523952e-06, "loss": 0.4997, "mean_token_accuracy": 0.8439276218414307, "num_tokens": 69859840.0, "step": 1836 }, { "epoch": 0.23368528177076708, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.7065188884735107, "learning_rate": 7.78295888088173e-06, "loss": 0.4908, "mean_token_accuracy": 0.8458065390586853, "num_tokens": 69893446.0, "step": 1837 }, { "epoch": 0.23381249204935758, "ewc_loss": 0.032958984375, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 2.444450616836548, "learning_rate": 7.78719796523951e-06, "loss": 0.4516, "mean_token_accuracy": 0.8574087023735046, "num_tokens": 69930258.0, "step": 1838 }, { "epoch": 0.23393970232794808, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.7325775623321533, "learning_rate": 7.791437049597288e-06, "loss": 0.5596, "mean_token_accuracy": 0.8294646143913269, "num_tokens": 69960691.0, "step": 1839 }, { "epoch": 0.2340669126065386, "ewc_loss": 0.032958984375, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 2.4368841648101807, "learning_rate": 7.795676133955067e-06, "loss": 0.5068, "mean_token_accuracy": 0.8435548543930054, "num_tokens": 69998503.0, "step": 1840 }, { "epoch": 0.2341941228851291, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.220008850097656e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.7375805377960205, "learning_rate": 7.799915218312844e-06, "loss": 0.5068, "mean_token_accuracy": 0.8436557054519653, "num_tokens": 70032381.0, "step": 1841 }, { "epoch": 0.23432133316371964, "ewc_loss": 0.032958984375, "ewc_loss_diag": 4.2438507080078125e-05, "ewc_loss_parallel": 2.872943878173828e-05, "grad_norm": 2.425253391265869, "learning_rate": 7.804154302670623e-06, "loss": 0.465, "mean_token_accuracy": 0.8569972515106201, "num_tokens": 70070616.0, "step": 1842 }, { "epoch": 0.23444854344231014, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.2438507080078125e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.7484447956085205, "learning_rate": 7.808393387028402e-06, "loss": 0.4909, "mean_token_accuracy": 0.8473287224769592, "num_tokens": 70111709.0, "step": 1843 }, { "epoch": 0.23457575372090064, "ewc_loss": 0.033203125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.44661808013916, "learning_rate": 7.812632471386181e-06, "loss": 0.5347, "mean_token_accuracy": 0.8328558802604675, "num_tokens": 70148516.0, "step": 1844 }, { "epoch": 0.23470296399949117, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.7658886909484863, "learning_rate": 7.81687155574396e-06, "loss": 0.5009, "mean_token_accuracy": 0.8468384146690369, "num_tokens": 70185688.0, "step": 1845 }, { "epoch": 0.23483017427808167, "ewc_loss": 0.033935546875, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 2.968311309814453e-05, "grad_norm": 2.4373013973236084, "learning_rate": 7.821110640101739e-06, "loss": 0.516, "mean_token_accuracy": 0.8394446969032288, "num_tokens": 70228901.0, "step": 1846 }, { "epoch": 0.23495738455667217, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.7779147624969482, "learning_rate": 7.825349724459518e-06, "loss": 0.5, "mean_token_accuracy": 0.8487472534179688, "num_tokens": 70268341.0, "step": 1847 }, { "epoch": 0.2350845948352627, "ewc_loss": 0.033935546875, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 2.968311309814453e-05, "grad_norm": 2.4558324813842773, "learning_rate": 7.829588808817295e-06, "loss": 0.4935, "mean_token_accuracy": 0.8472061157226562, "num_tokens": 70309285.0, "step": 1848 }, { "epoch": 0.2352118051138532, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.8026132583618164, "learning_rate": 7.833827893175074e-06, "loss": 0.5286, "mean_token_accuracy": 0.8351473808288574, "num_tokens": 70339694.0, "step": 1849 }, { "epoch": 0.2353390153924437, "ewc_loss": 0.0341796875, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.4666972160339355, "learning_rate": 7.838066977532853e-06, "loss": 0.4874, "mean_token_accuracy": 0.8496690988540649, "num_tokens": 70377684.0, "step": 1850 }, { "epoch": 0.23546622567103423, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.7829513549804688, "learning_rate": 7.842306061890632e-06, "loss": 0.5271, "mean_token_accuracy": 0.8407430052757263, "num_tokens": 70417690.0, "step": 1851 }, { "epoch": 0.23559343594962473, "ewc_loss": 0.034423828125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.0159950256347656e-05, "grad_norm": 2.486527919769287, "learning_rate": 7.84654514624841e-06, "loss": 0.5136, "mean_token_accuracy": 0.8407706022262573, "num_tokens": 70454485.0, "step": 1852 }, { "epoch": 0.23572064622821523, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.8894436359405518, "learning_rate": 7.85078423060619e-06, "loss": 0.5402, "mean_token_accuracy": 0.8383652567863464, "num_tokens": 70490337.0, "step": 1853 }, { "epoch": 0.23584785650680576, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.4707088470458984, "learning_rate": 7.855023314963968e-06, "loss": 0.5164, "mean_token_accuracy": 0.8444948196411133, "num_tokens": 70535194.0, "step": 1854 }, { "epoch": 0.23597506678539626, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.813007116317749, "learning_rate": 7.859262399321747e-06, "loss": 0.5214, "mean_token_accuracy": 0.839570164680481, "num_tokens": 70577046.0, "step": 1855 }, { "epoch": 0.23610227706398676, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.485645055770874, "learning_rate": 7.863501483679526e-06, "loss": 0.48, "mean_token_accuracy": 0.852211058139801, "num_tokens": 70616483.0, "step": 1856 }, { "epoch": 0.2362294873425773, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.8111908435821533, "learning_rate": 7.867740568037305e-06, "loss": 0.4795, "mean_token_accuracy": 0.8496026992797852, "num_tokens": 70654073.0, "step": 1857 }, { "epoch": 0.2363566976211678, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.519073009490967, "learning_rate": 7.871979652395084e-06, "loss": 0.5043, "mean_token_accuracy": 0.8422354459762573, "num_tokens": 70687530.0, "step": 1858 }, { "epoch": 0.2364839078997583, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.7604997158050537, "learning_rate": 7.876218736752863e-06, "loss": 0.462, "mean_token_accuracy": 0.8630220293998718, "num_tokens": 70729156.0, "step": 1859 }, { "epoch": 0.23661111817834882, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.514742374420166, "learning_rate": 7.880457821110642e-06, "loss": 0.5312, "mean_token_accuracy": 0.8383941650390625, "num_tokens": 70769781.0, "step": 1860 }, { "epoch": 0.23673832845693932, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.267692565917969e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.720808267593384, "learning_rate": 7.88469690546842e-06, "loss": 0.4651, "mean_token_accuracy": 0.8596046566963196, "num_tokens": 70811822.0, "step": 1861 }, { "epoch": 0.23686553873552982, "ewc_loss": 0.034912109375, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.5215790271759033, "learning_rate": 7.8889359898262e-06, "loss": 0.511, "mean_token_accuracy": 0.8404726982116699, "num_tokens": 70850470.0, "step": 1862 }, { "epoch": 0.23699274901412035, "ewc_loss": 0.035888671875, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.6871485710144043, "learning_rate": 7.893175074183978e-06, "loss": 0.5025, "mean_token_accuracy": 0.8450594544410706, "num_tokens": 70889270.0, "step": 1863 }, { "epoch": 0.23711995929271085, "ewc_loss": 0.034423828125, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.504715919494629, "learning_rate": 7.897414158541755e-06, "loss": 0.494, "mean_token_accuracy": 0.8482270240783691, "num_tokens": 70928787.0, "step": 1864 }, { "epoch": 0.23724716957130135, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.6749589443206787, "learning_rate": 7.901653242899534e-06, "loss": 0.531, "mean_token_accuracy": 0.841266930103302, "num_tokens": 70965726.0, "step": 1865 }, { "epoch": 0.23737437984989188, "ewc_loss": 0.033935546875, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 2.956390380859375e-05, "grad_norm": 2.4952504634857178, "learning_rate": 7.905892327257313e-06, "loss": 0.4926, "mean_token_accuracy": 0.8458145260810852, "num_tokens": 71002603.0, "step": 1866 }, { "epoch": 0.23750159012848238, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.6622440814971924, "learning_rate": 7.910131411615092e-06, "loss": 0.4869, "mean_token_accuracy": 0.8524428009986877, "num_tokens": 71039930.0, "step": 1867 }, { "epoch": 0.2376288004070729, "ewc_loss": 0.03369140625, "ewc_loss_diag": 4.315376281738281e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.4823060035705566, "learning_rate": 7.914370495972871e-06, "loss": 0.4921, "mean_token_accuracy": 0.8485124111175537, "num_tokens": 71077789.0, "step": 1868 }, { "epoch": 0.2377560106856634, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.3392181396484375e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.6655051708221436, "learning_rate": 7.91860958033065e-06, "loss": 0.5355, "mean_token_accuracy": 0.8376818895339966, "num_tokens": 71121325.0, "step": 1869 }, { "epoch": 0.2378832209642539, "ewc_loss": 0.033447265625, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 2.9087066650390625e-05, "grad_norm": 2.471837043762207, "learning_rate": 7.922848664688429e-06, "loss": 0.5331, "mean_token_accuracy": 0.8405195474624634, "num_tokens": 71159764.0, "step": 1870 }, { "epoch": 0.23801043124284443, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.6668550968170166, "learning_rate": 7.927087749046206e-06, "loss": 0.4412, "mean_token_accuracy": 0.8680150508880615, "num_tokens": 71196559.0, "step": 1871 }, { "epoch": 0.23813764152143493, "ewc_loss": 0.033203125, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 2.8848648071289062e-05, "grad_norm": 2.4553792476654053, "learning_rate": 7.931326833403985e-06, "loss": 0.4747, "mean_token_accuracy": 0.8501816391944885, "num_tokens": 71229925.0, "step": 1872 }, { "epoch": 0.23826485180002543, "ewc_loss": 0.035888671875, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.6931629180908203, "learning_rate": 7.935565917761764e-06, "loss": 0.4886, "mean_token_accuracy": 0.8506115078926086, "num_tokens": 71273764.0, "step": 1873 }, { "epoch": 0.23839206207861596, "ewc_loss": 0.032958984375, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 2.86102294921875e-05, "grad_norm": 2.453566789627075, "learning_rate": 7.939805002119543e-06, "loss": 0.5078, "mean_token_accuracy": 0.8430913686752319, "num_tokens": 71309090.0, "step": 1874 }, { "epoch": 0.23851927235720646, "ewc_loss": 0.035888671875, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.782752275466919, "learning_rate": 7.944044086477321e-06, "loss": 0.5073, "mean_token_accuracy": 0.8445608019828796, "num_tokens": 71348867.0, "step": 1875 }, { "epoch": 0.23864648263579696, "ewc_loss": 0.033203125, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.4125092029571533, "learning_rate": 7.9482831708351e-06, "loss": 0.4535, "mean_token_accuracy": 0.8585582375526428, "num_tokens": 71385707.0, "step": 1876 }, { "epoch": 0.2387736929143875, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.798691987991333, "learning_rate": 7.95252225519288e-06, "loss": 0.4266, "mean_token_accuracy": 0.8685786724090576, "num_tokens": 71428516.0, "step": 1877 }, { "epoch": 0.238900903192978, "ewc_loss": 0.033203125, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 2.8967857360839844e-05, "grad_norm": 2.3955695629119873, "learning_rate": 7.956761339550658e-06, "loss": 0.4482, "mean_token_accuracy": 0.862659215927124, "num_tokens": 71466913.0, "step": 1878 }, { "epoch": 0.2390281134715685, "ewc_loss": 0.03857421875, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.897841691970825, "learning_rate": 7.961000423908435e-06, "loss": 0.5068, "mean_token_accuracy": 0.8497169017791748, "num_tokens": 71506362.0, "step": 1879 }, { "epoch": 0.23915532375015902, "ewc_loss": 0.034423828125, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.4325168132781982, "learning_rate": 7.965239508266214e-06, "loss": 0.4756, "mean_token_accuracy": 0.85467129945755, "num_tokens": 71544303.0, "step": 1880 }, { "epoch": 0.23928253402874952, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.9568846225738525, "learning_rate": 7.969478592623993e-06, "loss": 0.4621, "mean_token_accuracy": 0.8591190576553345, "num_tokens": 71582295.0, "step": 1881 }, { "epoch": 0.23940974430734002, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.363059997558594e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.4755752086639404, "learning_rate": 7.973717676981772e-06, "loss": 0.4819, "mean_token_accuracy": 0.8533437252044678, "num_tokens": 71621880.0, "step": 1882 }, { "epoch": 0.23953695458593055, "ewc_loss": 0.0400390625, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.991802930831909, "learning_rate": 7.97795676133955e-06, "loss": 0.4482, "mean_token_accuracy": 0.861513614654541, "num_tokens": 71657997.0, "step": 1883 }, { "epoch": 0.23966416486452105, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.5662765502929688, "learning_rate": 7.98219584569733e-06, "loss": 0.5889, "mean_token_accuracy": 0.8190429210662842, "num_tokens": 71697622.0, "step": 1884 }, { "epoch": 0.23979137514311155, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.963313579559326, "learning_rate": 7.986434930055109e-06, "loss": 0.4536, "mean_token_accuracy": 0.8591880202293396, "num_tokens": 71733354.0, "step": 1885 }, { "epoch": 0.23991858542170208, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.5787606239318848, "learning_rate": 7.990674014412887e-06, "loss": 0.5064, "mean_token_accuracy": 0.845342755317688, "num_tokens": 71777172.0, "step": 1886 }, { "epoch": 0.24004579570029258, "ewc_loss": 0.0390625, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.9170494079589844, "learning_rate": 7.994913098770666e-06, "loss": 0.4974, "mean_token_accuracy": 0.8459442853927612, "num_tokens": 71816165.0, "step": 1887 }, { "epoch": 0.24017300597888308, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.6207616329193115, "learning_rate": 7.999152183128445e-06, "loss": 0.5246, "mean_token_accuracy": 0.8381701707839966, "num_tokens": 71852313.0, "step": 1888 }, { "epoch": 0.2403002162574736, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.8426456451416016, "learning_rate": 8.003391267486224e-06, "loss": 0.4756, "mean_token_accuracy": 0.8534834980964661, "num_tokens": 71892462.0, "step": 1889 }, { "epoch": 0.2404274265360641, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.6183557510375977, "learning_rate": 8.007630351844003e-06, "loss": 0.4842, "mean_token_accuracy": 0.8531416654586792, "num_tokens": 71924079.0, "step": 1890 }, { "epoch": 0.24055463681465464, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.7729103565216064, "learning_rate": 8.011869436201782e-06, "loss": 0.4723, "mean_token_accuracy": 0.85551917552948, "num_tokens": 71959546.0, "step": 1891 }, { "epoch": 0.24068184709324514, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.580301284790039, "learning_rate": 8.01610852055956e-06, "loss": 0.505, "mean_token_accuracy": 0.8440707921981812, "num_tokens": 72002565.0, "step": 1892 }, { "epoch": 0.24080905737183564, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.741880416870117, "learning_rate": 8.02034760491734e-06, "loss": 0.4641, "mean_token_accuracy": 0.8581715822219849, "num_tokens": 72037178.0, "step": 1893 }, { "epoch": 0.24093626765042617, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.560635805130005, "learning_rate": 8.024586689275117e-06, "loss": 0.4887, "mean_token_accuracy": 0.8499934673309326, "num_tokens": 72076562.0, "step": 1894 }, { "epoch": 0.24106347792901667, "ewc_loss": 0.036376953125, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.720348596572876, "learning_rate": 8.028825773632896e-06, "loss": 0.4983, "mean_token_accuracy": 0.850282609462738, "num_tokens": 72113379.0, "step": 1895 }, { "epoch": 0.24119068820760717, "ewc_loss": 0.034912109375, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.533844232559204, "learning_rate": 8.033064857990675e-06, "loss": 0.4253, "mean_token_accuracy": 0.8684369921684265, "num_tokens": 72149225.0, "step": 1896 }, { "epoch": 0.2413178984861977, "ewc_loss": 0.035888671875, "ewc_loss_diag": 4.38690185546875e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.708141326904297, "learning_rate": 8.037303942348453e-06, "loss": 0.4718, "mean_token_accuracy": 0.8536466360092163, "num_tokens": 72185774.0, "step": 1897 }, { "epoch": 0.2414451087647882, "ewc_loss": 0.0341796875, "ewc_loss_diag": 4.410743713378906e-05, "ewc_loss_parallel": 2.9802322387695312e-05, "grad_norm": 2.503255844116211, "learning_rate": 8.041543026706232e-06, "loss": 0.4643, "mean_token_accuracy": 0.857304573059082, "num_tokens": 72224270.0, "step": 1898 }, { "epoch": 0.2415723190433787, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.410743713378906e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.7231271266937256, "learning_rate": 8.045782111064011e-06, "loss": 0.4763, "mean_token_accuracy": 0.8527205586433411, "num_tokens": 72260629.0, "step": 1899 }, { "epoch": 0.24169952932196923, "ewc_loss": 0.033447265625, "ewc_loss_diag": 4.410743713378906e-05, "ewc_loss_parallel": 2.9087066650390625e-05, "grad_norm": 2.4844746589660645, "learning_rate": 8.05002119542179e-06, "loss": 0.5195, "mean_token_accuracy": 0.8397196531295776, "num_tokens": 72300040.0, "step": 1900 }, { "epoch": 0.24182673960055973, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.763733386993408, "learning_rate": 8.054260279779569e-06, "loss": 0.4903, "mean_token_accuracy": 0.8481860160827637, "num_tokens": 72335396.0, "step": 1901 }, { "epoch": 0.24195394987915023, "ewc_loss": 0.03369140625, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 2.9206275939941406e-05, "grad_norm": 2.491025447845459, "learning_rate": 8.058499364137346e-06, "loss": 0.5263, "mean_token_accuracy": 0.8416619896888733, "num_tokens": 72375215.0, "step": 1902 }, { "epoch": 0.24208116015774075, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.76932430267334, "learning_rate": 8.062738448495125e-06, "loss": 0.4436, "mean_token_accuracy": 0.8645385503768921, "num_tokens": 72416407.0, "step": 1903 }, { "epoch": 0.24220837043633126, "ewc_loss": 0.03369140625, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 2.9325485229492188e-05, "grad_norm": 2.4537038803100586, "learning_rate": 8.066977532852904e-06, "loss": 0.4409, "mean_token_accuracy": 0.8673131465911865, "num_tokens": 72455031.0, "step": 1904 }, { "epoch": 0.24233558071492176, "ewc_loss": 0.037841796875, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.8391451835632324, "learning_rate": 8.071216617210683e-06, "loss": 0.51, "mean_token_accuracy": 0.8445693254470825, "num_tokens": 72487186.0, "step": 1905 }, { "epoch": 0.24246279099351228, "ewc_loss": 0.0341796875, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 2.968311309814453e-05, "grad_norm": 2.442350387573242, "learning_rate": 8.075455701568462e-06, "loss": 0.4736, "mean_token_accuracy": 0.8528739809989929, "num_tokens": 72525659.0, "step": 1906 }, { "epoch": 0.24259000127210278, "ewc_loss": 0.03857421875, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.8929271697998047, "learning_rate": 8.07969478592624e-06, "loss": 0.4723, "mean_token_accuracy": 0.8580615520477295, "num_tokens": 72567912.0, "step": 1907 }, { "epoch": 0.24271721155069328, "ewc_loss": 0.034423828125, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.461268424987793, "learning_rate": 8.08393387028402e-06, "loss": 0.4264, "mean_token_accuracy": 0.8686201572418213, "num_tokens": 72599525.0, "step": 1908 }, { "epoch": 0.2428444218292838, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.9592232704162598, "learning_rate": 8.088172954641797e-06, "loss": 0.4855, "mean_token_accuracy": 0.8511242270469666, "num_tokens": 72635391.0, "step": 1909 }, { "epoch": 0.2429716321078743, "ewc_loss": 0.035888671875, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.579362392425537, "learning_rate": 8.092412038999576e-06, "loss": 0.523, "mean_token_accuracy": 0.8380569219589233, "num_tokens": 72666058.0, "step": 1910 }, { "epoch": 0.24309884238646481, "ewc_loss": 0.039306640625, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.9289636611938477, "learning_rate": 8.096651123357354e-06, "loss": 0.5338, "mean_token_accuracy": 0.8327407836914062, "num_tokens": 72699869.0, "step": 1911 }, { "epoch": 0.24322605266505534, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.546037435531616, "learning_rate": 8.100890207715133e-06, "loss": 0.5265, "mean_token_accuracy": 0.8414759039878845, "num_tokens": 72735660.0, "step": 1912 }, { "epoch": 0.24335326294364584, "ewc_loss": 0.038818359375, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.8964107036590576, "learning_rate": 8.105129292072912e-06, "loss": 0.4981, "mean_token_accuracy": 0.8486723303794861, "num_tokens": 72774334.0, "step": 1913 }, { "epoch": 0.24348047322223634, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.4345855712890625e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.5732476711273193, "learning_rate": 8.109368376430691e-06, "loss": 0.5223, "mean_token_accuracy": 0.8407135009765625, "num_tokens": 72808387.0, "step": 1914 }, { "epoch": 0.24360768350082687, "ewc_loss": 0.038330078125, "ewc_loss_diag": 4.458427429199219e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.8534140586853027, "learning_rate": 8.11360746078847e-06, "loss": 0.4634, "mean_token_accuracy": 0.8590853214263916, "num_tokens": 72846144.0, "step": 1915 }, { "epoch": 0.24373489377941737, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.458427429199219e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.5591888427734375, "learning_rate": 8.117846545146249e-06, "loss": 0.4383, "mean_token_accuracy": 0.8641219735145569, "num_tokens": 72887381.0, "step": 1916 }, { "epoch": 0.2438621040580079, "ewc_loss": 0.0380859375, "ewc_loss_diag": 4.458427429199219e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.8336868286132812, "learning_rate": 8.122085629504028e-06, "loss": 0.5239, "mean_token_accuracy": 0.8375363945960999, "num_tokens": 72927126.0, "step": 1917 }, { "epoch": 0.2439893143365984, "ewc_loss": 0.036376953125, "ewc_loss_diag": 4.458427429199219e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.5628747940063477, "learning_rate": 8.126324713861807e-06, "loss": 0.4987, "mean_token_accuracy": 0.8480358123779297, "num_tokens": 72968770.0, "step": 1918 }, { "epoch": 0.2441165246151889, "ewc_loss": 0.037841796875, "ewc_loss_diag": 4.458427429199219e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.8097639083862305, "learning_rate": 8.130563798219586e-06, "loss": 0.4772, "mean_token_accuracy": 0.8572032451629639, "num_tokens": 73008831.0, "step": 1919 }, { "epoch": 0.24424373489377943, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.556715250015259, "learning_rate": 8.134802882577364e-06, "loss": 0.5106, "mean_token_accuracy": 0.8472239971160889, "num_tokens": 73048469.0, "step": 1920 }, { "epoch": 0.24437094517236993, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.7782328128814697, "learning_rate": 8.139041966935143e-06, "loss": 0.4924, "mean_token_accuracy": 0.8515244722366333, "num_tokens": 73096067.0, "step": 1921 }, { "epoch": 0.24449815545096043, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.5586812496185303, "learning_rate": 8.143281051292922e-06, "loss": 0.5512, "mean_token_accuracy": 0.8377544283866882, "num_tokens": 73132680.0, "step": 1922 }, { "epoch": 0.24462536572955096, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.7721762657165527, "learning_rate": 8.147520135650701e-06, "loss": 0.5078, "mean_token_accuracy": 0.8439416885375977, "num_tokens": 73168436.0, "step": 1923 }, { "epoch": 0.24475257600814146, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.537442922592163, "learning_rate": 8.151759220008478e-06, "loss": 0.4827, "mean_token_accuracy": 0.8499653339385986, "num_tokens": 73210961.0, "step": 1924 }, { "epoch": 0.24487978628673196, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.7767364978790283, "learning_rate": 8.155998304366257e-06, "loss": 0.4903, "mean_token_accuracy": 0.8560566306114197, "num_tokens": 73244580.0, "step": 1925 }, { "epoch": 0.2450069965653225, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.5407955646514893, "learning_rate": 8.160237388724036e-06, "loss": 0.5273, "mean_token_accuracy": 0.8385916352272034, "num_tokens": 73278176.0, "step": 1926 }, { "epoch": 0.245134206843913, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.7715654373168945, "learning_rate": 8.164476473081815e-06, "loss": 0.5038, "mean_token_accuracy": 0.8476395606994629, "num_tokens": 73315678.0, "step": 1927 }, { "epoch": 0.2452614171225035, "ewc_loss": 0.03466796875, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.5015430450439453, "learning_rate": 8.168715557439594e-06, "loss": 0.446, "mean_token_accuracy": 0.8618432283401489, "num_tokens": 73356336.0, "step": 1928 }, { "epoch": 0.24538862740109402, "ewc_loss": 0.0380859375, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.7944085597991943, "learning_rate": 8.172954641797373e-06, "loss": 0.4721, "mean_token_accuracy": 0.8544434905052185, "num_tokens": 73396605.0, "step": 1929 }, { "epoch": 0.24551583767968452, "ewc_loss": 0.03466796875, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.0159950256347656e-05, "grad_norm": 2.4982316493988037, "learning_rate": 8.177193726155152e-06, "loss": 0.5777, "mean_token_accuracy": 0.824316143989563, "num_tokens": 73444278.0, "step": 1930 }, { "epoch": 0.24564304795827502, "ewc_loss": 0.0380859375, "ewc_loss_diag": 4.506111145019531e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.8363473415374756, "learning_rate": 8.18143281051293e-06, "loss": 0.5404, "mean_token_accuracy": 0.8371696472167969, "num_tokens": 73483755.0, "step": 1931 }, { "epoch": 0.24577025823686555, "ewc_loss": 0.03466796875, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.0159950256347656e-05, "grad_norm": 2.4927499294281006, "learning_rate": 8.185671894870708e-06, "loss": 0.5499, "mean_token_accuracy": 0.8320011496543884, "num_tokens": 73522819.0, "step": 1932 }, { "epoch": 0.24589746851545605, "ewc_loss": 0.038818359375, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.8832626342773438, "learning_rate": 8.189910979228487e-06, "loss": 0.4556, "mean_token_accuracy": 0.8575881719589233, "num_tokens": 73554973.0, "step": 1933 }, { "epoch": 0.24602467879404655, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.4969770908355713, "learning_rate": 8.194150063586265e-06, "loss": 0.5375, "mean_token_accuracy": 0.8354047536849976, "num_tokens": 73592541.0, "step": 1934 }, { "epoch": 0.24615188907263708, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.9258410930633545, "learning_rate": 8.198389147944044e-06, "loss": 0.5293, "mean_token_accuracy": 0.837729811668396, "num_tokens": 73629328.0, "step": 1935 }, { "epoch": 0.24627909935122758, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.505021572113037, "learning_rate": 8.202628232301823e-06, "loss": 0.5001, "mean_token_accuracy": 0.8498088121414185, "num_tokens": 73664944.0, "step": 1936 }, { "epoch": 0.24640630962981808, "ewc_loss": 0.040283203125, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.957298994064331, "learning_rate": 8.206867316659602e-06, "loss": 0.5388, "mean_token_accuracy": 0.8354566693305969, "num_tokens": 73712403.0, "step": 1937 }, { "epoch": 0.2465335199084086, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.51707124710083, "learning_rate": 8.211106401017381e-06, "loss": 0.4129, "mean_token_accuracy": 0.870647668838501, "num_tokens": 73752716.0, "step": 1938 }, { "epoch": 0.2466607301869991, "ewc_loss": 0.04052734375, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.9921553134918213, "learning_rate": 8.21534548537516e-06, "loss": 0.5452, "mean_token_accuracy": 0.8333276510238647, "num_tokens": 73796900.0, "step": 1939 }, { "epoch": 0.2467879404655896, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.567844867706299, "learning_rate": 8.219584569732939e-06, "loss": 0.5112, "mean_token_accuracy": 0.8395569324493408, "num_tokens": 73836426.0, "step": 1940 }, { "epoch": 0.24691515074418013, "ewc_loss": 0.04052734375, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.9901888370513916, "learning_rate": 8.223823654090718e-06, "loss": 0.4518, "mean_token_accuracy": 0.8605092763900757, "num_tokens": 73872147.0, "step": 1941 }, { "epoch": 0.24704236102277063, "ewc_loss": 0.037841796875, "ewc_loss_diag": 4.553794860839844e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.6093645095825195, "learning_rate": 8.228062738448497e-06, "loss": 0.4871, "mean_token_accuracy": 0.8476003408432007, "num_tokens": 73912878.0, "step": 1942 }, { "epoch": 0.24716957130136116, "ewc_loss": 0.040283203125, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.94401478767395, "learning_rate": 8.232301822806274e-06, "loss": 0.516, "mean_token_accuracy": 0.8442515134811401, "num_tokens": 73952152.0, "step": 1943 }, { "epoch": 0.24729678157995166, "ewc_loss": 0.0380859375, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.6415326595306396, "learning_rate": 8.236540907164053e-06, "loss": 0.5012, "mean_token_accuracy": 0.8488456010818481, "num_tokens": 73988760.0, "step": 1944 }, { "epoch": 0.24742399185854216, "ewc_loss": 0.038818359375, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.8853261470794678, "learning_rate": 8.240779991521831e-06, "loss": 0.4785, "mean_token_accuracy": 0.8527777194976807, "num_tokens": 74028107.0, "step": 1945 }, { "epoch": 0.2475512021371327, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.6373538970947266, "learning_rate": 8.24501907587961e-06, "loss": 0.5181, "mean_token_accuracy": 0.8407731652259827, "num_tokens": 74071236.0, "step": 1946 }, { "epoch": 0.2476784124157232, "ewc_loss": 0.0380859375, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.827122688293457, "learning_rate": 8.24925816023739e-06, "loss": 0.5192, "mean_token_accuracy": 0.8475095629692078, "num_tokens": 74105734.0, "step": 1947 }, { "epoch": 0.2478056226943137, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.6376798152923584, "learning_rate": 8.253497244595168e-06, "loss": 0.5176, "mean_token_accuracy": 0.841562032699585, "num_tokens": 74142569.0, "step": 1948 }, { "epoch": 0.24793283297290422, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.765653133392334, "learning_rate": 8.257736328952947e-06, "loss": 0.5371, "mean_token_accuracy": 0.8358786106109619, "num_tokens": 74185395.0, "step": 1949 }, { "epoch": 0.24806004325149472, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.6121819019317627, "learning_rate": 8.261975413310726e-06, "loss": 0.4943, "mean_token_accuracy": 0.8478246927261353, "num_tokens": 74223416.0, "step": 1950 }, { "epoch": 0.24818725353008522, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.721160650253296, "learning_rate": 8.266214497668505e-06, "loss": 0.4556, "mean_token_accuracy": 0.8622105717658997, "num_tokens": 74265356.0, "step": 1951 }, { "epoch": 0.24831446380867575, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.610926389694214, "learning_rate": 8.270453582026284e-06, "loss": 0.5295, "mean_token_accuracy": 0.8375913500785828, "num_tokens": 74303807.0, "step": 1952 }, { "epoch": 0.24844167408726625, "ewc_loss": 0.036376953125, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.7088353633880615, "learning_rate": 8.274692666384063e-06, "loss": 0.48, "mean_token_accuracy": 0.8534636497497559, "num_tokens": 74337299.0, "step": 1953 }, { "epoch": 0.24856888436585675, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.601478576660156e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.5987682342529297, "learning_rate": 8.278931750741841e-06, "loss": 0.4981, "mean_token_accuracy": 0.8454868793487549, "num_tokens": 74370817.0, "step": 1954 }, { "epoch": 0.24869609464444728, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.7056591510772705, "learning_rate": 8.283170835099619e-06, "loss": 0.4866, "mean_token_accuracy": 0.8570317625999451, "num_tokens": 74408009.0, "step": 1955 }, { "epoch": 0.24882330492303778, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.5650196075439453, "learning_rate": 8.287409919457398e-06, "loss": 0.464, "mean_token_accuracy": 0.8569497466087341, "num_tokens": 74446669.0, "step": 1956 }, { "epoch": 0.24895051520162828, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.706301689147949, "learning_rate": 8.291649003815176e-06, "loss": 0.5079, "mean_token_accuracy": 0.8450247645378113, "num_tokens": 74484264.0, "step": 1957 }, { "epoch": 0.2490777254802188, "ewc_loss": 0.034912109375, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.53971004486084, "learning_rate": 8.295888088172955e-06, "loss": 0.5246, "mean_token_accuracy": 0.8438766598701477, "num_tokens": 74523956.0, "step": 1958 }, { "epoch": 0.2492049357588093, "ewc_loss": 0.037353515625, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.753455400466919, "learning_rate": 8.300127172530734e-06, "loss": 0.4281, "mean_token_accuracy": 0.8683383464813232, "num_tokens": 74560748.0, "step": 1959 }, { "epoch": 0.2493321460373998, "ewc_loss": 0.03466796875, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 3.0159950256347656e-05, "grad_norm": 2.5186805725097656, "learning_rate": 8.304366256888513e-06, "loss": 0.4745, "mean_token_accuracy": 0.8529036641120911, "num_tokens": 74596465.0, "step": 1960 }, { "epoch": 0.24945935631599034, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.6253204345703125e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.822953939437866, "learning_rate": 8.308605341246292e-06, "loss": 0.4848, "mean_token_accuracy": 0.8516542315483093, "num_tokens": 74632788.0, "step": 1961 }, { "epoch": 0.24958656659458084, "ewc_loss": 0.03466796875, "ewc_loss_diag": 4.649162292480469e-05, "ewc_loss_parallel": 3.0040740966796875e-05, "grad_norm": 2.4736690521240234, "learning_rate": 8.312844425604069e-06, "loss": 0.4815, "mean_token_accuracy": 0.8562368750572205, "num_tokens": 74667723.0, "step": 1962 }, { "epoch": 0.24971377687317134, "ewc_loss": 0.038818359375, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.885544776916504, "learning_rate": 8.317083509961848e-06, "loss": 0.5171, "mean_token_accuracy": 0.8472284078598022, "num_tokens": 74709205.0, "step": 1963 }, { "epoch": 0.24984098715176187, "ewc_loss": 0.034912109375, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.43489146232605, "learning_rate": 8.321322594319627e-06, "loss": 0.4558, "mean_token_accuracy": 0.8577126264572144, "num_tokens": 74749437.0, "step": 1964 }, { "epoch": 0.24996819743035237, "ewc_loss": 0.040771484375, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 3.015547513961792, "learning_rate": 8.325561678677406e-06, "loss": 0.5293, "mean_token_accuracy": 0.8376636505126953, "num_tokens": 74790602.0, "step": 1965 }, { "epoch": 0.2500954077089429, "ewc_loss": 0.035888671875, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.4650425910949707, "learning_rate": 8.329800763035185e-06, "loss": 0.5312, "mean_token_accuracy": 0.8392404317855835, "num_tokens": 74830845.0, "step": 1966 }, { "epoch": 0.25022261798753337, "ewc_loss": 0.04296875, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.153907537460327, "learning_rate": 8.334039847392964e-06, "loss": 0.4872, "mean_token_accuracy": 0.8508749008178711, "num_tokens": 74864208.0, "step": 1967 }, { "epoch": 0.2503498282661239, "ewc_loss": 0.038330078125, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.5509033203125, "learning_rate": 8.338278931750742e-06, "loss": 0.4553, "mean_token_accuracy": 0.8619583249092102, "num_tokens": 74897217.0, "step": 1968 }, { "epoch": 0.2504770385447144, "ewc_loss": 0.044189453125, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.2254083156585693, "learning_rate": 8.342518016108521e-06, "loss": 0.4824, "mean_token_accuracy": 0.8522034883499146, "num_tokens": 74932105.0, "step": 1969 }, { "epoch": 0.2506042488233049, "ewc_loss": 0.03955078125, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.6916301250457764, "learning_rate": 8.3467571004663e-06, "loss": 0.4835, "mean_token_accuracy": 0.8540963530540466, "num_tokens": 74965229.0, "step": 1970 }, { "epoch": 0.2507314591018954, "ewc_loss": 0.0439453125, "ewc_loss_diag": 4.673004150390625e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.1967153549194336, "learning_rate": 8.350996184824079e-06, "loss": 0.5326, "mean_token_accuracy": 0.8391925692558289, "num_tokens": 75002290.0, "step": 1971 }, { "epoch": 0.25085866938048595, "ewc_loss": 0.040283203125, "ewc_loss_diag": 4.696846008300781e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.778306007385254, "learning_rate": 8.355235269181858e-06, "loss": 0.508, "mean_token_accuracy": 0.8435539603233337, "num_tokens": 75038113.0, "step": 1972 }, { "epoch": 0.2509858796590764, "ewc_loss": 0.04248046875, "ewc_loss_diag": 4.696846008300781e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 3.087661027908325, "learning_rate": 8.359474353539637e-06, "loss": 0.4729, "mean_token_accuracy": 0.8561677932739258, "num_tokens": 75078145.0, "step": 1973 }, { "epoch": 0.25111308993766696, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.696846008300781e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.7718863487243652, "learning_rate": 8.363713437897416e-06, "loss": 0.5591, "mean_token_accuracy": 0.8322348594665527, "num_tokens": 75126473.0, "step": 1974 }, { "epoch": 0.2512403002162575, "ewc_loss": 0.040771484375, "ewc_loss_diag": 4.696846008300781e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.992352247238159, "learning_rate": 8.367952522255195e-06, "loss": 0.4946, "mean_token_accuracy": 0.8493839502334595, "num_tokens": 75166785.0, "step": 1975 }, { "epoch": 0.25136751049484796, "ewc_loss": 0.03955078125, "ewc_loss_diag": 4.696846008300781e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.7556121349334717, "learning_rate": 8.372191606612974e-06, "loss": 0.497, "mean_token_accuracy": 0.8473262786865234, "num_tokens": 75204599.0, "step": 1976 }, { "epoch": 0.2514947207734385, "ewc_loss": 0.03955078125, "ewc_loss_diag": 4.696846008300781e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.9132513999938965, "learning_rate": 8.37643069097075e-06, "loss": 0.4528, "mean_token_accuracy": 0.859512448310852, "num_tokens": 75235295.0, "step": 1977 }, { "epoch": 0.251621931052029, "ewc_loss": 0.038330078125, "ewc_loss_diag": 4.744529724121094e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.6963531970977783, "learning_rate": 8.38066977532853e-06, "loss": 0.4559, "mean_token_accuracy": 0.8617681264877319, "num_tokens": 75275549.0, "step": 1978 }, { "epoch": 0.25174914133061954, "ewc_loss": 0.03857421875, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.8605058193206787, "learning_rate": 8.384908859686308e-06, "loss": 0.4554, "mean_token_accuracy": 0.8618342876434326, "num_tokens": 75316920.0, "step": 1979 }, { "epoch": 0.25187635160921, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.653632164001465, "learning_rate": 8.389147944044087e-06, "loss": 0.492, "mean_token_accuracy": 0.8506930470466614, "num_tokens": 75360091.0, "step": 1980 }, { "epoch": 0.25200356188780054, "ewc_loss": 0.03857421875, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.848430871963501, "learning_rate": 8.393387028401866e-06, "loss": 0.52, "mean_token_accuracy": 0.8413163423538208, "num_tokens": 75397123.0, "step": 1981 }, { "epoch": 0.25213077216639107, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.638237953186035, "learning_rate": 8.397626112759645e-06, "loss": 0.5419, "mean_token_accuracy": 0.8345204591751099, "num_tokens": 75429983.0, "step": 1982 }, { "epoch": 0.25225798244498154, "ewc_loss": 0.0380859375, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.849214553833008, "learning_rate": 8.401865197117424e-06, "loss": 0.4736, "mean_token_accuracy": 0.8568821549415588, "num_tokens": 75460322.0, "step": 1983 }, { "epoch": 0.25238519272357207, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.5815019607543945, "learning_rate": 8.406104281475203e-06, "loss": 0.4598, "mean_token_accuracy": 0.8573645949363708, "num_tokens": 75499733.0, "step": 1984 }, { "epoch": 0.2525124030021626, "ewc_loss": 0.0390625, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.8546178340911865, "learning_rate": 8.41034336583298e-06, "loss": 0.4692, "mean_token_accuracy": 0.8559515476226807, "num_tokens": 75541182.0, "step": 1985 }, { "epoch": 0.2526396132807531, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.5464537143707275, "learning_rate": 8.414582450190759e-06, "loss": 0.4813, "mean_token_accuracy": 0.8508889675140381, "num_tokens": 75585052.0, "step": 1986 }, { "epoch": 0.2527668235593436, "ewc_loss": 0.03955078125, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.926673173904419, "learning_rate": 8.418821534548538e-06, "loss": 0.5736, "mean_token_accuracy": 0.8290286064147949, "num_tokens": 75625208.0, "step": 1987 }, { "epoch": 0.25289403383793413, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.5398330688476562, "learning_rate": 8.423060618906317e-06, "loss": 0.4905, "mean_token_accuracy": 0.8504737019538879, "num_tokens": 75665205.0, "step": 1988 }, { "epoch": 0.2530212441165246, "ewc_loss": 0.04052734375, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.9788618087768555, "learning_rate": 8.427299703264096e-06, "loss": 0.4913, "mean_token_accuracy": 0.8510423898696899, "num_tokens": 75708278.0, "step": 1989 }, { "epoch": 0.25314845439511513, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.76837158203125e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.5444116592407227, "learning_rate": 8.431538787621875e-06, "loss": 0.5077, "mean_token_accuracy": 0.8447694182395935, "num_tokens": 75750565.0, "step": 1990 }, { "epoch": 0.25327566467370566, "ewc_loss": 0.04150390625, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 3.0561299324035645, "learning_rate": 8.435777871979653e-06, "loss": 0.4775, "mean_token_accuracy": 0.8545717597007751, "num_tokens": 75790519.0, "step": 1991 }, { "epoch": 0.25340287495229613, "ewc_loss": 0.03759765625, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.6043975353240967, "learning_rate": 8.440016956337432e-06, "loss": 0.502, "mean_token_accuracy": 0.8493205308914185, "num_tokens": 75828256.0, "step": 1992 }, { "epoch": 0.25353008523088666, "ewc_loss": 0.041259765625, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 3.061397075653076, "learning_rate": 8.44425604069521e-06, "loss": 0.5174, "mean_token_accuracy": 0.8476628661155701, "num_tokens": 75869238.0, "step": 1993 }, { "epoch": 0.2536572955094772, "ewc_loss": 0.038330078125, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.6369986534118652, "learning_rate": 8.448495125052988e-06, "loss": 0.5038, "mean_token_accuracy": 0.8471930027008057, "num_tokens": 75906776.0, "step": 1994 }, { "epoch": 0.25378450578806766, "ewc_loss": 0.041259765625, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 3.0358784198760986, "learning_rate": 8.452734209410767e-06, "loss": 0.4387, "mean_token_accuracy": 0.8677233457565308, "num_tokens": 75940271.0, "step": 1995 }, { "epoch": 0.2539117160666582, "ewc_loss": 0.0390625, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.685643434524536, "learning_rate": 8.456973293768546e-06, "loss": 0.4702, "mean_token_accuracy": 0.8550447821617126, "num_tokens": 75974536.0, "step": 1996 }, { "epoch": 0.2540389263452487, "ewc_loss": 0.040771484375, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.9669432640075684, "learning_rate": 8.461212378126325e-06, "loss": 0.5249, "mean_token_accuracy": 0.8418726921081543, "num_tokens": 76013809.0, "step": 1997 }, { "epoch": 0.2541661366238392, "ewc_loss": 0.0390625, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.71553635597229, "learning_rate": 8.465451462484104e-06, "loss": 0.5497, "mean_token_accuracy": 0.8347324132919312, "num_tokens": 76049559.0, "step": 1998 }, { "epoch": 0.2542933469024297, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.8895347118377686, "learning_rate": 8.469690546841883e-06, "loss": 0.4927, "mean_token_accuracy": 0.8490154147148132, "num_tokens": 76087091.0, "step": 1999 }, { "epoch": 0.25442055718102025, "ewc_loss": 0.03857421875, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.706838846206665, "learning_rate": 8.473929631199662e-06, "loss": 0.4685, "mean_token_accuracy": 0.8563010096549988, "num_tokens": 76121466.0, "step": 2000 }, { "epoch": 0.2545477674596107, "ewc_loss": 0.038818359375, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.8130664825439453, "learning_rate": 8.47816871555744e-06, "loss": 0.4893, "mean_token_accuracy": 0.8549435138702393, "num_tokens": 76158291.0, "step": 2001 }, { "epoch": 0.25467497773820125, "ewc_loss": 0.037353515625, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.696822166442871, "learning_rate": 8.48240779991522e-06, "loss": 0.501, "mean_token_accuracy": 0.8455675840377808, "num_tokens": 76197350.0, "step": 2002 }, { "epoch": 0.2548021880167918, "ewc_loss": 0.037841796875, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.7568445205688477, "learning_rate": 8.486646884272998e-06, "loss": 0.4939, "mean_token_accuracy": 0.8471484780311584, "num_tokens": 76235367.0, "step": 2003 }, { "epoch": 0.25492939829538225, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.697338581085205, "learning_rate": 8.490885968630777e-06, "loss": 0.5276, "mean_token_accuracy": 0.8377258777618408, "num_tokens": 76270462.0, "step": 2004 }, { "epoch": 0.2550566085739728, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.7127432823181152, "learning_rate": 8.495125052988556e-06, "loss": 0.5162, "mean_token_accuracy": 0.8466500639915466, "num_tokens": 76315032.0, "step": 2005 }, { "epoch": 0.2551838188525633, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.6788151264190674, "learning_rate": 8.499364137346335e-06, "loss": 0.4969, "mean_token_accuracy": 0.8510042428970337, "num_tokens": 76356337.0, "step": 2006 }, { "epoch": 0.2553110291311538, "ewc_loss": 0.036376953125, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.728543758392334, "learning_rate": 8.503603221704114e-06, "loss": 0.5186, "mean_token_accuracy": 0.8415470123291016, "num_tokens": 76391329.0, "step": 2007 }, { "epoch": 0.2554382394097443, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.6432783603668213, "learning_rate": 8.507842306061891e-06, "loss": 0.4817, "mean_token_accuracy": 0.8508964776992798, "num_tokens": 76429015.0, "step": 2008 }, { "epoch": 0.25556544968833483, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.7205333709716797, "learning_rate": 8.51208139041967e-06, "loss": 0.5169, "mean_token_accuracy": 0.8504955768585205, "num_tokens": 76470343.0, "step": 2009 }, { "epoch": 0.2556926599669253, "ewc_loss": 0.035888671875, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.0994415283203125e-05, "grad_norm": 2.6332101821899414, "learning_rate": 8.516320474777449e-06, "loss": 0.5, "mean_token_accuracy": 0.8485406041145325, "num_tokens": 76505140.0, "step": 2010 }, { "epoch": 0.25581987024551583, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.7230658531188965, "learning_rate": 8.520559559135228e-06, "loss": 0.4587, "mean_token_accuracy": 0.8634154796600342, "num_tokens": 76541701.0, "step": 2011 }, { "epoch": 0.25594708052410636, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.61851167678833, "learning_rate": 8.524798643493007e-06, "loss": 0.5182, "mean_token_accuracy": 0.847891092300415, "num_tokens": 76580208.0, "step": 2012 }, { "epoch": 0.25607429080269684, "ewc_loss": 0.037841796875, "ewc_loss_diag": 4.792213439941406e-05, "ewc_loss_parallel": 3.2901763916015625e-05, "grad_norm": 2.7336723804473877, "learning_rate": 8.529037727850785e-06, "loss": 0.4668, "mean_token_accuracy": 0.8569286465644836, "num_tokens": 76617750.0, "step": 2013 }, { "epoch": 0.25620150108128736, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.8160552978515625e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.5904407501220703, "learning_rate": 8.533276812208564e-06, "loss": 0.4355, "mean_token_accuracy": 0.8654483556747437, "num_tokens": 76652678.0, "step": 2014 }, { "epoch": 0.2563287113598779, "ewc_loss": 0.0380859375, "ewc_loss_diag": 4.8160552978515625e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.7682759761810303, "learning_rate": 8.537515896566342e-06, "loss": 0.4726, "mean_token_accuracy": 0.853954553604126, "num_tokens": 76695375.0, "step": 2015 }, { "epoch": 0.25645592163846836, "ewc_loss": 0.03564453125, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.075599670410156e-05, "grad_norm": 2.582106113433838, "learning_rate": 8.54175498092412e-06, "loss": 0.4829, "mean_token_accuracy": 0.8528826236724854, "num_tokens": 76728695.0, "step": 2016 }, { "epoch": 0.2565831319170589, "ewc_loss": 0.038330078125, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.8126282691955566, "learning_rate": 8.5459940652819e-06, "loss": 0.4604, "mean_token_accuracy": 0.8609179854393005, "num_tokens": 76761081.0, "step": 2017 }, { "epoch": 0.2567103421956494, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.5481200218200684, "learning_rate": 8.550233149639678e-06, "loss": 0.5527, "mean_token_accuracy": 0.829997181892395, "num_tokens": 76800335.0, "step": 2018 }, { "epoch": 0.2568375524742399, "ewc_loss": 0.039306640625, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.8685739040374756, "learning_rate": 8.554472233997457e-06, "loss": 0.521, "mean_token_accuracy": 0.8409310579299927, "num_tokens": 76839653.0, "step": 2019 }, { "epoch": 0.2569647627528304, "ewc_loss": 0.035400390625, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.0517578125e-05, "grad_norm": 2.520617723464966, "learning_rate": 8.558711318355236e-06, "loss": 0.5483, "mean_token_accuracy": 0.829390823841095, "num_tokens": 76876557.0, "step": 2020 }, { "epoch": 0.25709197303142095, "ewc_loss": 0.04052734375, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.9636833667755127, "learning_rate": 8.562950402713015e-06, "loss": 0.445, "mean_token_accuracy": 0.8686524629592896, "num_tokens": 76910000.0, "step": 2021 }, { "epoch": 0.2572191833100114, "ewc_loss": 0.03515625, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.0279159545898438e-05, "grad_norm": 2.4721338748931885, "learning_rate": 8.567189487070794e-06, "loss": 0.5309, "mean_token_accuracy": 0.8408204317092896, "num_tokens": 76946861.0, "step": 2022 }, { "epoch": 0.25734639358860195, "ewc_loss": 0.042236328125, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 3.1071901321411133, "learning_rate": 8.571428571428571e-06, "loss": 0.5533, "mean_token_accuracy": 0.8368265628814697, "num_tokens": 76990831.0, "step": 2023 }, { "epoch": 0.2574736038671925, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.839897155761719e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.4572834968566895, "learning_rate": 8.57566765578635e-06, "loss": 0.4509, "mean_token_accuracy": 0.8650123476982117, "num_tokens": 77033145.0, "step": 2024 }, { "epoch": 0.25760081414578295, "ewc_loss": 0.046142578125, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.2976529598236084, "learning_rate": 8.579906740144129e-06, "loss": 0.4949, "mean_token_accuracy": 0.8540672063827515, "num_tokens": 77076621.0, "step": 2025 }, { "epoch": 0.2577280244243735, "ewc_loss": 0.039306640625, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.600713014602661, "learning_rate": 8.584145824501908e-06, "loss": 0.524, "mean_token_accuracy": 0.8402363061904907, "num_tokens": 77115951.0, "step": 2026 }, { "epoch": 0.257855234702964, "ewc_loss": 0.04931640625, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.4233295917510986, "learning_rate": 8.588384908859686e-06, "loss": 0.5307, "mean_token_accuracy": 0.84681636095047, "num_tokens": 77154674.0, "step": 2027 }, { "epoch": 0.25798244498155454, "ewc_loss": 0.042236328125, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.807943820953369, "learning_rate": 8.592623993217465e-06, "loss": 0.5038, "mean_token_accuracy": 0.8496497869491577, "num_tokens": 77192970.0, "step": 2028 }, { "epoch": 0.258109655260145, "ewc_loss": 0.047607421875, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.386883020401001, "learning_rate": 8.596863077575244e-06, "loss": 0.5414, "mean_token_accuracy": 0.8414002656936646, "num_tokens": 77231258.0, "step": 2029 }, { "epoch": 0.25823686553873554, "ewc_loss": 0.0439453125, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9562084674835205, "learning_rate": 8.601102161933023e-06, "loss": 0.513, "mean_token_accuracy": 0.847496747970581, "num_tokens": 77269829.0, "step": 2030 }, { "epoch": 0.25836407581732607, "ewc_loss": 0.045166015625, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.231377363204956, "learning_rate": 8.605341246290802e-06, "loss": 0.4919, "mean_token_accuracy": 0.8516660332679749, "num_tokens": 77309611.0, "step": 2031 }, { "epoch": 0.25849128609591654, "ewc_loss": 0.042724609375, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.96000075340271, "learning_rate": 8.609580330648581e-06, "loss": 0.4851, "mean_token_accuracy": 0.8559678792953491, "num_tokens": 77341937.0, "step": 2032 }, { "epoch": 0.25861849637450707, "ewc_loss": 0.04296875, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 3.0961170196533203, "learning_rate": 8.61381941500636e-06, "loss": 0.5529, "mean_token_accuracy": 0.8382906913757324, "num_tokens": 77380484.0, "step": 2033 }, { "epoch": 0.2587457066530976, "ewc_loss": 0.040771484375, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.8832883834838867, "learning_rate": 8.618058499364139e-06, "loss": 0.5443, "mean_token_accuracy": 0.8359130620956421, "num_tokens": 77421220.0, "step": 2034 }, { "epoch": 0.25887291693168807, "ewc_loss": 0.04052734375, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.9723055362701416, "learning_rate": 8.622297583721918e-06, "loss": 0.4912, "mean_token_accuracy": 0.8545238971710205, "num_tokens": 77465136.0, "step": 2035 }, { "epoch": 0.2590001272102786, "ewc_loss": 0.0400390625, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.8148603439331055, "learning_rate": 8.626536668079696e-06, "loss": 0.5017, "mean_token_accuracy": 0.8473519086837769, "num_tokens": 77500120.0, "step": 2036 }, { "epoch": 0.2591273374888691, "ewc_loss": 0.03955078125, "ewc_loss_diag": 4.887580871582031e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.9210352897644043, "learning_rate": 8.630775752437475e-06, "loss": 0.487, "mean_token_accuracy": 0.8513863682746887, "num_tokens": 77532860.0, "step": 2037 }, { "epoch": 0.2592545477674596, "ewc_loss": 0.03857421875, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.7451884746551514, "learning_rate": 8.635014836795252e-06, "loss": 0.5234, "mean_token_accuracy": 0.8423018455505371, "num_tokens": 77569154.0, "step": 2038 }, { "epoch": 0.2593817580460501, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.893453598022461, "learning_rate": 8.639253921153031e-06, "loss": 0.4564, "mean_token_accuracy": 0.8596224784851074, "num_tokens": 77602918.0, "step": 2039 }, { "epoch": 0.25950896832464065, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.67952823638916, "learning_rate": 8.64349300551081e-06, "loss": 0.4714, "mean_token_accuracy": 0.8546251058578491, "num_tokens": 77637233.0, "step": 2040 }, { "epoch": 0.2596361786032311, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.9080920219421387, "learning_rate": 8.647732089868589e-06, "loss": 0.5555, "mean_token_accuracy": 0.83147132396698, "num_tokens": 77678344.0, "step": 2041 }, { "epoch": 0.25976338888182166, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.620311975479126, "learning_rate": 8.651971174226368e-06, "loss": 0.4626, "mean_token_accuracy": 0.8569501042366028, "num_tokens": 77712706.0, "step": 2042 }, { "epoch": 0.2598905991604122, "ewc_loss": 0.04052734375, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.9710042476654053, "learning_rate": 8.656210258584147e-06, "loss": 0.5308, "mean_token_accuracy": 0.8381951451301575, "num_tokens": 77749841.0, "step": 2043 }, { "epoch": 0.26001780943900266, "ewc_loss": 0.036376953125, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.590242624282837, "learning_rate": 8.660449342941926e-06, "loss": 0.5259, "mean_token_accuracy": 0.8382341861724854, "num_tokens": 77788464.0, "step": 2044 }, { "epoch": 0.2601450197175932, "ewc_loss": 0.041015625, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 3.040522575378418, "learning_rate": 8.664688427299705e-06, "loss": 0.5052, "mean_token_accuracy": 0.8445295095443726, "num_tokens": 77826493.0, "step": 2045 }, { "epoch": 0.2602722299961837, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.5733954906463623, "learning_rate": 8.668927511657482e-06, "loss": 0.4597, "mean_token_accuracy": 0.8575791716575623, "num_tokens": 77857814.0, "step": 2046 }, { "epoch": 0.2603994402747742, "ewc_loss": 0.043212890625, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.1519885063171387, "learning_rate": 8.67316659601526e-06, "loss": 0.5131, "mean_token_accuracy": 0.8464837670326233, "num_tokens": 77891263.0, "step": 2047 }, { "epoch": 0.2605266505533647, "ewc_loss": 0.038330078125, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.5967748165130615, "learning_rate": 8.67740568037304e-06, "loss": 0.4766, "mean_token_accuracy": 0.85853111743927, "num_tokens": 77931683.0, "step": 2048 }, { "epoch": 0.26065386083195524, "ewc_loss": 0.044921875, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.2544543743133545, "learning_rate": 8.681644764730819e-06, "loss": 0.5083, "mean_token_accuracy": 0.847226083278656, "num_tokens": 77965204.0, "step": 2049 }, { "epoch": 0.2607810711105457, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.702697277069092, "learning_rate": 8.685883849088597e-06, "loss": 0.4442, "mean_token_accuracy": 0.8663890957832336, "num_tokens": 78003427.0, "step": 2050 }, { "epoch": 0.26090828138913624, "ewc_loss": 0.045166015625, "ewc_loss_diag": 4.9114227294921875e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.252427339553833, "learning_rate": 8.690122933446376e-06, "loss": 0.5074, "mean_token_accuracy": 0.8508009314537048, "num_tokens": 78045524.0, "step": 2051 }, { "epoch": 0.26103549166772677, "ewc_loss": 0.040771484375, "ewc_loss_diag": 4.935264587402344e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.8214383125305176, "learning_rate": 8.694362017804155e-06, "loss": 0.4783, "mean_token_accuracy": 0.8546112179756165, "num_tokens": 78085485.0, "step": 2052 }, { "epoch": 0.26116270194631724, "ewc_loss": 0.0439453125, "ewc_loss_diag": 4.935264587402344e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.1588642597198486, "learning_rate": 8.698601102161932e-06, "loss": 0.5138, "mean_token_accuracy": 0.8442939519882202, "num_tokens": 78128212.0, "step": 2053 }, { "epoch": 0.26128991222490777, "ewc_loss": 0.04052734375, "ewc_loss_diag": 4.935264587402344e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.8574893474578857, "learning_rate": 8.702840186519711e-06, "loss": 0.4641, "mean_token_accuracy": 0.8600828647613525, "num_tokens": 78170959.0, "step": 2054 }, { "epoch": 0.2614171225034983, "ewc_loss": 0.04150390625, "ewc_loss_diag": 4.935264587402344e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 3.0171589851379395, "learning_rate": 8.70707927087749e-06, "loss": 0.4641, "mean_token_accuracy": 0.8600658774375916, "num_tokens": 78211181.0, "step": 2055 }, { "epoch": 0.2615443327820888, "ewc_loss": 0.04052734375, "ewc_loss_diag": 4.935264587402344e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.882458448410034, "learning_rate": 8.711318355235269e-06, "loss": 0.549, "mean_token_accuracy": 0.8307168483734131, "num_tokens": 78246281.0, "step": 2056 }, { "epoch": 0.2616715430606793, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.935264587402344e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.8890950679779053, "learning_rate": 8.715557439593048e-06, "loss": 0.4666, "mean_token_accuracy": 0.8562235236167908, "num_tokens": 78279604.0, "step": 2057 }, { "epoch": 0.26179875333926983, "ewc_loss": 0.0400390625, "ewc_loss_diag": 4.9591064453125e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.844667673110962, "learning_rate": 8.719796523950827e-06, "loss": 0.5508, "mean_token_accuracy": 0.8312005996704102, "num_tokens": 78317534.0, "step": 2058 }, { "epoch": 0.2619259636178603, "ewc_loss": 0.038330078125, "ewc_loss_diag": 4.9591064453125e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.7809362411499023, "learning_rate": 8.724035608308606e-06, "loss": 0.4284, "mean_token_accuracy": 0.8695669174194336, "num_tokens": 78349744.0, "step": 2059 }, { "epoch": 0.26205317389645083, "ewc_loss": 0.0390625, "ewc_loss_diag": 4.9591064453125e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.8211474418640137, "learning_rate": 8.728274692666385e-06, "loss": 0.5008, "mean_token_accuracy": 0.8470140695571899, "num_tokens": 78390841.0, "step": 2060 }, { "epoch": 0.26218038417504136, "ewc_loss": 0.037353515625, "ewc_loss_diag": 4.9591064453125e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.7299344539642334, "learning_rate": 8.732513777024163e-06, "loss": 0.4952, "mean_token_accuracy": 0.8488017916679382, "num_tokens": 78425604.0, "step": 2061 }, { "epoch": 0.26230759445363183, "ewc_loss": 0.038818359375, "ewc_loss_diag": 4.9591064453125e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.8103418350219727, "learning_rate": 8.736752861381942e-06, "loss": 0.5433, "mean_token_accuracy": 0.8335692286491394, "num_tokens": 78469444.0, "step": 2062 }, { "epoch": 0.26243480473222236, "ewc_loss": 0.036865234375, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.68502140045166, "learning_rate": 8.740991945739721e-06, "loss": 0.4907, "mean_token_accuracy": 0.8493028879165649, "num_tokens": 78507931.0, "step": 2063 }, { "epoch": 0.2625620150108129, "ewc_loss": 0.0390625, "ewc_loss_diag": 4.9591064453125e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.817216396331787, "learning_rate": 8.7452310300975e-06, "loss": 0.4879, "mean_token_accuracy": 0.8533240556716919, "num_tokens": 78550572.0, "step": 2064 }, { "epoch": 0.26268922528940336, "ewc_loss": 0.03662109375, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.651492118835449, "learning_rate": 8.749470114455279e-06, "loss": 0.4638, "mean_token_accuracy": 0.8586603999137878, "num_tokens": 78586975.0, "step": 2065 }, { "epoch": 0.2628164355679939, "ewc_loss": 0.0390625, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.8517560958862305, "learning_rate": 8.753709198813058e-06, "loss": 0.5637, "mean_token_accuracy": 0.8294404745101929, "num_tokens": 78630535.0, "step": 2066 }, { "epoch": 0.2629436458465844, "ewc_loss": 0.036376953125, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.6077682971954346, "learning_rate": 8.757948283170837e-06, "loss": 0.4658, "mean_token_accuracy": 0.8605085611343384, "num_tokens": 78672486.0, "step": 2067 }, { "epoch": 0.2630708561251749, "ewc_loss": 0.039794921875, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.906043529510498, "learning_rate": 8.762187367528614e-06, "loss": 0.4612, "mean_token_accuracy": 0.8592532873153687, "num_tokens": 78710208.0, "step": 2068 }, { "epoch": 0.2631980664037654, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.566059112548828, "learning_rate": 8.766426451886393e-06, "loss": 0.5078, "mean_token_accuracy": 0.8440225124359131, "num_tokens": 78754233.0, "step": 2069 }, { "epoch": 0.26332527668235595, "ewc_loss": 0.041259765625, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 3.0090086460113525, "learning_rate": 8.770665536244172e-06, "loss": 0.5471, "mean_token_accuracy": 0.8391139507293701, "num_tokens": 78794530.0, "step": 2070 }, { "epoch": 0.2634524869609464, "ewc_loss": 0.0361328125, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.123283386230469e-05, "grad_norm": 2.5313782691955566, "learning_rate": 8.77490462060195e-06, "loss": 0.4593, "mean_token_accuracy": 0.858601450920105, "num_tokens": 78828620.0, "step": 2071 }, { "epoch": 0.26357969723953695, "ewc_loss": 0.04296875, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 3.1457207202911377, "learning_rate": 8.77914370495973e-06, "loss": 0.467, "mean_token_accuracy": 0.8574737906455994, "num_tokens": 78867658.0, "step": 2072 }, { "epoch": 0.2637069075181275, "ewc_loss": 0.037109375, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.536391258239746, "learning_rate": 8.783382789317508e-06, "loss": 0.5045, "mean_token_accuracy": 0.8470554351806641, "num_tokens": 78903744.0, "step": 2073 }, { "epoch": 0.26383411779671795, "ewc_loss": 0.0458984375, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.3223042488098145, "learning_rate": 8.787621873675287e-06, "loss": 0.5355, "mean_token_accuracy": 0.836855411529541, "num_tokens": 78934098.0, "step": 2074 }, { "epoch": 0.2639613280753085, "ewc_loss": 0.039306640625, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.6229209899902344, "learning_rate": 8.791860958033066e-06, "loss": 0.4673, "mean_token_accuracy": 0.8574696183204651, "num_tokens": 78969473.0, "step": 2075 }, { "epoch": 0.264088538353899, "ewc_loss": 0.049072265625, "ewc_loss_diag": 4.982948303222656e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.4411120414733887, "learning_rate": 8.796100042390843e-06, "loss": 0.474, "mean_token_accuracy": 0.8604675531387329, "num_tokens": 79008799.0, "step": 2076 }, { "epoch": 0.2642157486324895, "ewc_loss": 0.0419921875, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.832911252975464, "learning_rate": 8.800339126748622e-06, "loss": 0.4747, "mean_token_accuracy": 0.8552873134613037, "num_tokens": 79042502.0, "step": 2077 }, { "epoch": 0.26434295891108, "ewc_loss": 0.048828125, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.405755043029785, "learning_rate": 8.804578211106401e-06, "loss": 0.5515, "mean_token_accuracy": 0.8328835368156433, "num_tokens": 79079641.0, "step": 2078 }, { "epoch": 0.26447016918967053, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.955331325531006, "learning_rate": 8.80881729546418e-06, "loss": 0.494, "mean_token_accuracy": 0.8521424531936646, "num_tokens": 79116120.0, "step": 2079 }, { "epoch": 0.26459737946826106, "ewc_loss": 0.04541015625, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.274385452270508, "learning_rate": 8.813056379821959e-06, "loss": 0.4783, "mean_token_accuracy": 0.8598260879516602, "num_tokens": 79150322.0, "step": 2080 }, { "epoch": 0.26472458974685154, "ewc_loss": 0.04296875, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.9719433784484863, "learning_rate": 8.817295464179738e-06, "loss": 0.5023, "mean_token_accuracy": 0.8468030095100403, "num_tokens": 79184944.0, "step": 2081 }, { "epoch": 0.26485180002544206, "ewc_loss": 0.0439453125, "ewc_loss_diag": 5.0067901611328125e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.1253485679626465, "learning_rate": 8.821534548537517e-06, "loss": 0.5699, "mean_token_accuracy": 0.8247775435447693, "num_tokens": 79226137.0, "step": 2082 }, { "epoch": 0.2649790103040326, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 2.909169912338257, "learning_rate": 8.825773632895296e-06, "loss": 0.4736, "mean_token_accuracy": 0.8558090925216675, "num_tokens": 79267721.0, "step": 2083 }, { "epoch": 0.26510622058262306, "ewc_loss": 0.0419921875, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 3.024282455444336, "learning_rate": 8.830012717253074e-06, "loss": 0.4788, "mean_token_accuracy": 0.8556594848632812, "num_tokens": 79305355.0, "step": 2084 }, { "epoch": 0.2652334308612136, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.842031717300415, "learning_rate": 8.834251801610853e-06, "loss": 0.5108, "mean_token_accuracy": 0.8438077569007874, "num_tokens": 79349714.0, "step": 2085 }, { "epoch": 0.2653606411398041, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.9968764781951904, "learning_rate": 8.838490885968632e-06, "loss": 0.4628, "mean_token_accuracy": 0.8598932027816772, "num_tokens": 79387208.0, "step": 2086 }, { "epoch": 0.2654878514183946, "ewc_loss": 0.039794921875, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.774923801422119, "learning_rate": 8.842729970326411e-06, "loss": 0.4524, "mean_token_accuracy": 0.8667389154434204, "num_tokens": 79424396.0, "step": 2087 }, { "epoch": 0.2656150616969851, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.95951247215271, "learning_rate": 8.846969054684188e-06, "loss": 0.5744, "mean_token_accuracy": 0.829383909702301, "num_tokens": 79465105.0, "step": 2088 }, { "epoch": 0.26574227197557565, "ewc_loss": 0.03857421875, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.71517014503479, "learning_rate": 8.851208139041967e-06, "loss": 0.4611, "mean_token_accuracy": 0.8586611747741699, "num_tokens": 79504000.0, "step": 2089 }, { "epoch": 0.2658694822541661, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.996065378189087, "learning_rate": 8.855447223399746e-06, "loss": 0.5234, "mean_token_accuracy": 0.8400425910949707, "num_tokens": 79538954.0, "step": 2090 }, { "epoch": 0.26599669253275665, "ewc_loss": 0.037353515625, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.670131206512451, "learning_rate": 8.859686307757525e-06, "loss": 0.5398, "mean_token_accuracy": 0.8360263109207153, "num_tokens": 79577264.0, "step": 2091 }, { "epoch": 0.2661239028113472, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 3.042400360107422, "learning_rate": 8.863925392115304e-06, "loss": 0.4893, "mean_token_accuracy": 0.8511297702789307, "num_tokens": 79612803.0, "step": 2092 }, { "epoch": 0.26625111308993765, "ewc_loss": 0.03759765625, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.24249267578125e-05, "grad_norm": 2.6143553256988525, "learning_rate": 8.868164476473083e-06, "loss": 0.4387, "mean_token_accuracy": 0.8662862777709961, "num_tokens": 79651595.0, "step": 2093 }, { "epoch": 0.2663783233685282, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.13106107711792, "learning_rate": 8.872403560830862e-06, "loss": 0.4834, "mean_token_accuracy": 0.8570878505706787, "num_tokens": 79691587.0, "step": 2094 }, { "epoch": 0.2665055336471187, "ewc_loss": 0.038330078125, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.6408495903015137, "learning_rate": 8.87664264518864e-06, "loss": 0.5295, "mean_token_accuracy": 0.83878493309021, "num_tokens": 79729803.0, "step": 2095 }, { "epoch": 0.2666327439257092, "ewc_loss": 0.044677734375, "ewc_loss_diag": 5.030632019042969e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.2249984741210938, "learning_rate": 8.88088172954642e-06, "loss": 0.4542, "mean_token_accuracy": 0.8635098338127136, "num_tokens": 79765866.0, "step": 2096 }, { "epoch": 0.2667599542042997, "ewc_loss": 0.039794921875, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.69488263130188, "learning_rate": 8.885120813904198e-06, "loss": 0.46, "mean_token_accuracy": 0.8596898317337036, "num_tokens": 79808973.0, "step": 2097 }, { "epoch": 0.26688716448289024, "ewc_loss": 0.045654296875, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.2772061824798584, "learning_rate": 8.889359898261977e-06, "loss": 0.5662, "mean_token_accuracy": 0.8317594528198242, "num_tokens": 79843677.0, "step": 2098 }, { "epoch": 0.2670143747614807, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.784522533416748, "learning_rate": 8.893598982619754e-06, "loss": 0.4754, "mean_token_accuracy": 0.8566824197769165, "num_tokens": 79882680.0, "step": 2099 }, { "epoch": 0.26714158504007124, "ewc_loss": 0.045166015625, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.2153053283691406, "learning_rate": 8.897838066977533e-06, "loss": 0.4944, "mean_token_accuracy": 0.8513991832733154, "num_tokens": 79929691.0, "step": 2100 }, { "epoch": 0.26726879531866177, "ewc_loss": 0.041259765625, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.848776340484619, "learning_rate": 8.902077151335312e-06, "loss": 0.4888, "mean_token_accuracy": 0.8507819771766663, "num_tokens": 79970515.0, "step": 2101 }, { "epoch": 0.26739600559725224, "ewc_loss": 0.043701171875, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.127136707305908, "learning_rate": 8.906316235693091e-06, "loss": 0.584, "mean_token_accuracy": 0.8188859820365906, "num_tokens": 80007550.0, "step": 2102 }, { "epoch": 0.26752321587584277, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.8794634342193604, "learning_rate": 8.91055532005087e-06, "loss": 0.4955, "mean_token_accuracy": 0.8470187187194824, "num_tokens": 80044106.0, "step": 2103 }, { "epoch": 0.2676504261544333, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 2.9917397499084473, "learning_rate": 8.914794404408649e-06, "loss": 0.4962, "mean_token_accuracy": 0.8504976630210876, "num_tokens": 80084429.0, "step": 2104 }, { "epoch": 0.26777763643302377, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.054473876953125e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.8556067943573, "learning_rate": 8.919033488766428e-06, "loss": 0.4331, "mean_token_accuracy": 0.8670982122421265, "num_tokens": 80123317.0, "step": 2105 }, { "epoch": 0.2679048467116143, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.901197671890259, "learning_rate": 8.923272573124205e-06, "loss": 0.5223, "mean_token_accuracy": 0.8449022173881531, "num_tokens": 80157658.0, "step": 2106 }, { "epoch": 0.2680320569902048, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.8425159454345703, "learning_rate": 8.927511657481984e-06, "loss": 0.531, "mean_token_accuracy": 0.8384483456611633, "num_tokens": 80200482.0, "step": 2107 }, { "epoch": 0.2681592672687953, "ewc_loss": 0.0390625, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.831467390060425, "learning_rate": 8.931750741839763e-06, "loss": 0.4562, "mean_token_accuracy": 0.8600707650184631, "num_tokens": 80231844.0, "step": 2108 }, { "epoch": 0.2682864775473858, "ewc_loss": 0.0400390625, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.8303987979888916, "learning_rate": 8.935989826197541e-06, "loss": 0.4783, "mean_token_accuracy": 0.8550312519073486, "num_tokens": 80269734.0, "step": 2109 }, { "epoch": 0.26841368782597635, "ewc_loss": 0.03857421875, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.7776033878326416, "learning_rate": 8.94022891055532e-06, "loss": 0.4469, "mean_token_accuracy": 0.8672604560852051, "num_tokens": 80306309.0, "step": 2110 }, { "epoch": 0.2685408981045668, "ewc_loss": 0.03955078125, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.828366279602051, "learning_rate": 8.9444679949131e-06, "loss": 0.5029, "mean_token_accuracy": 0.8444782495498657, "num_tokens": 80345390.0, "step": 2111 }, { "epoch": 0.26866810838315736, "ewc_loss": 0.038330078125, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.732585906982422, "learning_rate": 8.948707079270878e-06, "loss": 0.4207, "mean_token_accuracy": 0.8720102310180664, "num_tokens": 80384118.0, "step": 2112 }, { "epoch": 0.2687953186617479, "ewc_loss": 0.03955078125, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.8574347496032715, "learning_rate": 8.952946163628657e-06, "loss": 0.4794, "mean_token_accuracy": 0.8537449240684509, "num_tokens": 80416703.0, "step": 2113 }, { "epoch": 0.26892252894033836, "ewc_loss": 0.037841796875, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.6889216899871826, "learning_rate": 8.957185247986436e-06, "loss": 0.4282, "mean_token_accuracy": 0.8697580099105835, "num_tokens": 80453732.0, "step": 2114 }, { "epoch": 0.2690497392189289, "ewc_loss": 0.039794921875, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.8873157501220703, "learning_rate": 8.961424332344215e-06, "loss": 0.4851, "mean_token_accuracy": 0.8535979390144348, "num_tokens": 80495265.0, "step": 2115 }, { "epoch": 0.2691769494975194, "ewc_loss": 0.037109375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.647108793258667, "learning_rate": 8.965663416701994e-06, "loss": 0.4737, "mean_token_accuracy": 0.8538272976875305, "num_tokens": 80533789.0, "step": 2116 }, { "epoch": 0.2693041597761099, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.9616141319274902, "learning_rate": 8.969902501059773e-06, "loss": 0.4536, "mean_token_accuracy": 0.8640485405921936, "num_tokens": 80570802.0, "step": 2117 }, { "epoch": 0.2694313700547004, "ewc_loss": 0.03662109375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.147125244140625e-05, "grad_norm": 2.592545986175537, "learning_rate": 8.974141585417551e-06, "loss": 0.4197, "mean_token_accuracy": 0.870427131652832, "num_tokens": 80608031.0, "step": 2118 }, { "epoch": 0.26955858033329094, "ewc_loss": 0.042724609375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.0807888507843018, "learning_rate": 8.97838066977533e-06, "loss": 0.5139, "mean_token_accuracy": 0.8450801372528076, "num_tokens": 80646271.0, "step": 2119 }, { "epoch": 0.2696857906118814, "ewc_loss": 0.036865234375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.170967102050781e-05, "grad_norm": 2.5552732944488525, "learning_rate": 8.98261975413311e-06, "loss": 0.5054, "mean_token_accuracy": 0.8448989391326904, "num_tokens": 80683325.0, "step": 2120 }, { "epoch": 0.26981300089047194, "ewc_loss": 0.045166015625, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.2440009117126465, "learning_rate": 8.986858838490888e-06, "loss": 0.5219, "mean_token_accuracy": 0.8455431461334229, "num_tokens": 80726403.0, "step": 2121 }, { "epoch": 0.26994021116906247, "ewc_loss": 0.038818359375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.549579620361328, "learning_rate": 8.991097922848665e-06, "loss": 0.4759, "mean_token_accuracy": 0.8560380935668945, "num_tokens": 80767293.0, "step": 2122 }, { "epoch": 0.27006742144765294, "ewc_loss": 0.049560546875, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.4675004482269287, "learning_rate": 8.995337007206444e-06, "loss": 0.5156, "mean_token_accuracy": 0.8459586501121521, "num_tokens": 80806218.0, "step": 2123 }, { "epoch": 0.2701946317262435, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.1021575927734375e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.722301959991455, "learning_rate": 8.999576091564223e-06, "loss": 0.474, "mean_token_accuracy": 0.8586012125015259, "num_tokens": 80841694.0, "step": 2124 }, { "epoch": 0.270321842004834, "ewc_loss": 0.052734375, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.579801559448242, "learning_rate": 9.003815175922002e-06, "loss": 0.4912, "mean_token_accuracy": 0.8535032868385315, "num_tokens": 80874879.0, "step": 2125 }, { "epoch": 0.2704490522834245, "ewc_loss": 0.045654296875, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.958475351333618, "learning_rate": 9.00805426027978e-06, "loss": 0.5, "mean_token_accuracy": 0.8515032529830933, "num_tokens": 80916317.0, "step": 2126 }, { "epoch": 0.270576262562015, "ewc_loss": 0.051025390625, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.4980411529541016, "learning_rate": 9.01229334463756e-06, "loss": 0.5291, "mean_token_accuracy": 0.8440069556236267, "num_tokens": 80952304.0, "step": 2127 }, { "epoch": 0.27070347284060553, "ewc_loss": 0.04833984375, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.084475040435791, "learning_rate": 9.016532428995339e-06, "loss": 0.4672, "mean_token_accuracy": 0.859963059425354, "num_tokens": 80989886.0, "step": 2128 }, { "epoch": 0.27083068311919606, "ewc_loss": 0.046875, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.3226752281188965, "learning_rate": 9.020771513353116e-06, "loss": 0.5194, "mean_token_accuracy": 0.843722939491272, "num_tokens": 81031041.0, "step": 2129 }, { "epoch": 0.27095789339778653, "ewc_loss": 0.044921875, "ewc_loss_diag": 5.125999450683594e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0591964721679688, "learning_rate": 9.025010597710895e-06, "loss": 0.5102, "mean_token_accuracy": 0.8463507890701294, "num_tokens": 81068915.0, "step": 2130 }, { "epoch": 0.27108510367637706, "ewc_loss": 0.044189453125, "ewc_loss_diag": 5.1975250244140625e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.1662211418151855, "learning_rate": 9.029249682068674e-06, "loss": 0.4745, "mean_token_accuracy": 0.8584027886390686, "num_tokens": 81106044.0, "step": 2131 }, { "epoch": 0.2712123139549676, "ewc_loss": 0.042236328125, "ewc_loss_diag": 5.1975250244140625e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.97640061378479, "learning_rate": 9.033488766426452e-06, "loss": 0.4813, "mean_token_accuracy": 0.8533934354782104, "num_tokens": 81142589.0, "step": 2132 }, { "epoch": 0.27133952423355806, "ewc_loss": 0.042724609375, "ewc_loss_diag": 5.1975250244140625e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.055386543273926, "learning_rate": 9.037727850784231e-06, "loss": 0.46, "mean_token_accuracy": 0.8607677221298218, "num_tokens": 81178415.0, "step": 2133 }, { "epoch": 0.2714667345121486, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.1975250244140625e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.9103565216064453, "learning_rate": 9.04196693514201e-06, "loss": 0.5002, "mean_token_accuracy": 0.8497112393379211, "num_tokens": 81216952.0, "step": 2134 }, { "epoch": 0.2715939447907391, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.1975250244140625e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.9772229194641113, "learning_rate": 9.046206019499789e-06, "loss": 0.449, "mean_token_accuracy": 0.8657639026641846, "num_tokens": 81257879.0, "step": 2135 }, { "epoch": 0.2717211550693296, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.1975250244140625e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.8460114002227783, "learning_rate": 9.050445103857568e-06, "loss": 0.5242, "mean_token_accuracy": 0.8444960117340088, "num_tokens": 81298045.0, "step": 2136 }, { "epoch": 0.2718483653479201, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.9586520195007324, "learning_rate": 9.054684188215345e-06, "loss": 0.4384, "mean_token_accuracy": 0.86953204870224, "num_tokens": 81332342.0, "step": 2137 }, { "epoch": 0.27197557562651065, "ewc_loss": 0.0390625, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.3855438232421875e-05, "grad_norm": 2.7941150665283203, "learning_rate": 9.058923272573124e-06, "loss": 0.5121, "mean_token_accuracy": 0.8415806889533997, "num_tokens": 81367270.0, "step": 2138 }, { "epoch": 0.2721027859051011, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.9385664463043213, "learning_rate": 9.063162356930903e-06, "loss": 0.4642, "mean_token_accuracy": 0.8607640266418457, "num_tokens": 81406644.0, "step": 2139 }, { "epoch": 0.27222999618369165, "ewc_loss": 0.03857421875, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.314018249511719e-05, "grad_norm": 2.7413063049316406, "learning_rate": 9.067401441288682e-06, "loss": 0.4951, "mean_token_accuracy": 0.8525534272193909, "num_tokens": 81443225.0, "step": 2140 }, { "epoch": 0.2723572064622822, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.9710588455200195, "learning_rate": 9.07164052564646e-06, "loss": 0.4655, "mean_token_accuracy": 0.855000913143158, "num_tokens": 81482168.0, "step": 2141 }, { "epoch": 0.27248441674087265, "ewc_loss": 0.0380859375, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.693005084991455, "learning_rate": 9.07587961000424e-06, "loss": 0.505, "mean_token_accuracy": 0.8455331325531006, "num_tokens": 81518127.0, "step": 2142 }, { "epoch": 0.2726116270194632, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 3.041832685470581, "learning_rate": 9.080118694362018e-06, "loss": 0.5098, "mean_token_accuracy": 0.8427478075027466, "num_tokens": 81559857.0, "step": 2143 }, { "epoch": 0.2727388372980537, "ewc_loss": 0.03759765625, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.6501951217651367, "learning_rate": 9.084357778719797e-06, "loss": 0.4982, "mean_token_accuracy": 0.8463703393936157, "num_tokens": 81596509.0, "step": 2144 }, { "epoch": 0.2728660475766442, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.14924955368042, "learning_rate": 9.088596863077576e-06, "loss": 0.4827, "mean_token_accuracy": 0.8572466373443604, "num_tokens": 81632693.0, "step": 2145 }, { "epoch": 0.2729932578552347, "ewc_loss": 0.0380859375, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.6177453994750977, "learning_rate": 9.092835947435355e-06, "loss": 0.4892, "mean_token_accuracy": 0.8495092391967773, "num_tokens": 81670067.0, "step": 2146 }, { "epoch": 0.27312046813382523, "ewc_loss": 0.04638671875, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.3073437213897705, "learning_rate": 9.097075031793134e-06, "loss": 0.5396, "mean_token_accuracy": 0.8368617296218872, "num_tokens": 81706902.0, "step": 2147 }, { "epoch": 0.2732476784124157, "ewc_loss": 0.0400390625, "ewc_loss_diag": 5.245208740234375e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.657240867614746, "learning_rate": 9.101314116150913e-06, "loss": 0.4454, "mean_token_accuracy": 0.8676736354827881, "num_tokens": 81743340.0, "step": 2148 }, { "epoch": 0.27337488869100623, "ewc_loss": 0.0498046875, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.437174081802368, "learning_rate": 9.105553200508692e-06, "loss": 0.4711, "mean_token_accuracy": 0.8614606857299805, "num_tokens": 81779573.0, "step": 2149 }, { "epoch": 0.27350209896959676, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 2.791738748550415, "learning_rate": 9.10979228486647e-06, "loss": 0.4904, "mean_token_accuracy": 0.8532038331031799, "num_tokens": 81822004.0, "step": 2150 }, { "epoch": 0.27362930924818724, "ewc_loss": 0.050537109375, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.4752938747406006, "learning_rate": 9.11403136922425e-06, "loss": 0.5019, "mean_token_accuracy": 0.8508799076080322, "num_tokens": 81860472.0, "step": 2151 }, { "epoch": 0.27375651952677776, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.269050598144531e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.970485210418701, "learning_rate": 9.118270453582027e-06, "loss": 0.5545, "mean_token_accuracy": 0.8375764489173889, "num_tokens": 81899151.0, "step": 2152 }, { "epoch": 0.2738837298053683, "ewc_loss": 0.048095703125, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.3400845527648926, "learning_rate": 9.122509537939806e-06, "loss": 0.4835, "mean_token_accuracy": 0.8518437147140503, "num_tokens": 81941213.0, "step": 2153 }, { "epoch": 0.27401094008395877, "ewc_loss": 0.044921875, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.021608829498291, "learning_rate": 9.126748622297584e-06, "loss": 0.5262, "mean_token_accuracy": 0.8370058536529541, "num_tokens": 81978055.0, "step": 2154 }, { "epoch": 0.2741381503625493, "ewc_loss": 0.044921875, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.178738832473755, "learning_rate": 9.130987706655363e-06, "loss": 0.5168, "mean_token_accuracy": 0.8442186117172241, "num_tokens": 82015599.0, "step": 2155 }, { "epoch": 0.2742653606411398, "ewc_loss": 0.042724609375, "ewc_loss_diag": 5.2928924560546875e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.999523162841797, "learning_rate": 9.135226791013142e-06, "loss": 0.54, "mean_token_accuracy": 0.8438113927841187, "num_tokens": 82055113.0, "step": 2156 }, { "epoch": 0.2743925709197303, "ewc_loss": 0.042236328125, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 3.0532095432281494, "learning_rate": 9.139465875370921e-06, "loss": 0.4845, "mean_token_accuracy": 0.8509715795516968, "num_tokens": 82083052.0, "step": 2157 }, { "epoch": 0.2745197811983208, "ewc_loss": 0.041259765625, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.954131841659546, "learning_rate": 9.1437049597287e-06, "loss": 0.5008, "mean_token_accuracy": 0.8458957672119141, "num_tokens": 82118850.0, "step": 2158 }, { "epoch": 0.27464699147691135, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.9457530975341797, "learning_rate": 9.147944044086477e-06, "loss": 0.522, "mean_token_accuracy": 0.8403793573379517, "num_tokens": 82151261.0, "step": 2159 }, { "epoch": 0.2747742017555018, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.9068706035614014, "learning_rate": 9.152183128444256e-06, "loss": 0.524, "mean_token_accuracy": 0.8471466302871704, "num_tokens": 82192018.0, "step": 2160 }, { "epoch": 0.27490141203409235, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.879303216934204, "learning_rate": 9.156422212802035e-06, "loss": 0.5328, "mean_token_accuracy": 0.8347267508506775, "num_tokens": 82231044.0, "step": 2161 }, { "epoch": 0.2750286223126829, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.885560989379883, "learning_rate": 9.160661297159814e-06, "loss": 0.4632, "mean_token_accuracy": 0.859455943107605, "num_tokens": 82265334.0, "step": 2162 }, { "epoch": 0.27515583259127335, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.840182304382324, "learning_rate": 9.164900381517593e-06, "loss": 0.4781, "mean_token_accuracy": 0.8521716594696045, "num_tokens": 82303995.0, "step": 2163 }, { "epoch": 0.2752830428698639, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.8694231510162354, "learning_rate": 9.169139465875372e-06, "loss": 0.4863, "mean_token_accuracy": 0.8512508273124695, "num_tokens": 82345157.0, "step": 2164 }, { "epoch": 0.2754102531484544, "ewc_loss": 0.039794921875, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.8204500675201416, "learning_rate": 9.17337855023315e-06, "loss": 0.4923, "mean_token_accuracy": 0.8485038876533508, "num_tokens": 82380723.0, "step": 2165 }, { "epoch": 0.2755374634270449, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.8514604568481445, "learning_rate": 9.17761763459093e-06, "loss": 0.4515, "mean_token_accuracy": 0.8619867563247681, "num_tokens": 82418985.0, "step": 2166 }, { "epoch": 0.2756646737056354, "ewc_loss": 0.0400390625, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.7991957664489746, "learning_rate": 9.181856718948707e-06, "loss": 0.4305, "mean_token_accuracy": 0.8683664798736572, "num_tokens": 82455498.0, "step": 2167 }, { "epoch": 0.27579188398422594, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.8676533699035645, "learning_rate": 9.186095803306485e-06, "loss": 0.4741, "mean_token_accuracy": 0.8549004793167114, "num_tokens": 82491117.0, "step": 2168 }, { "epoch": 0.2759190942628164, "ewc_loss": 0.03955078125, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.7801170349121094, "learning_rate": 9.190334887664264e-06, "loss": 0.5217, "mean_token_accuracy": 0.8423551321029663, "num_tokens": 82533640.0, "step": 2169 }, { "epoch": 0.27604630454140694, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.897939920425415, "learning_rate": 9.194573972022043e-06, "loss": 0.5765, "mean_token_accuracy": 0.8326106071472168, "num_tokens": 82568812.0, "step": 2170 }, { "epoch": 0.27617351481999747, "ewc_loss": 0.0390625, "ewc_loss_diag": 5.316734313964844e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.737560749053955, "learning_rate": 9.198813056379822e-06, "loss": 0.5636, "mean_token_accuracy": 0.8294320702552795, "num_tokens": 82611777.0, "step": 2171 }, { "epoch": 0.27630072509858794, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.9387710094451904, "learning_rate": 9.203052140737601e-06, "loss": 0.5115, "mean_token_accuracy": 0.8455530405044556, "num_tokens": 82646938.0, "step": 2172 }, { "epoch": 0.27642793537717847, "ewc_loss": 0.0380859375, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 3.266334533691406e-05, "grad_norm": 2.6966042518615723, "learning_rate": 9.20729122509538e-06, "loss": 0.4486, "mean_token_accuracy": 0.8639846444129944, "num_tokens": 82684316.0, "step": 2173 }, { "epoch": 0.276555145655769, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 3.0068750381469727, "learning_rate": 9.211530309453159e-06, "loss": 0.496, "mean_token_accuracy": 0.8475829362869263, "num_tokens": 82723204.0, "step": 2174 }, { "epoch": 0.27668235593435947, "ewc_loss": 0.03759765625, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 3.218650817871094e-05, "grad_norm": 2.6351022720336914, "learning_rate": 9.215769393810938e-06, "loss": 0.4874, "mean_token_accuracy": 0.8526400327682495, "num_tokens": 82759634.0, "step": 2175 }, { "epoch": 0.27680956621295, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.144134998321533, "learning_rate": 9.220008478168717e-06, "loss": 0.5832, "mean_token_accuracy": 0.8243181109428406, "num_tokens": 82800446.0, "step": 2176 }, { "epoch": 0.2769367764915405, "ewc_loss": 0.037353515625, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 3.1948089599609375e-05, "grad_norm": 2.566309928894043, "learning_rate": 9.224247562526495e-06, "loss": 0.4918, "mean_token_accuracy": 0.8490215539932251, "num_tokens": 82840754.0, "step": 2177 }, { "epoch": 0.277063986770131, "ewc_loss": 0.047607421875, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.3606221675872803, "learning_rate": 9.228486646884274e-06, "loss": 0.5008, "mean_token_accuracy": 0.850555419921875, "num_tokens": 82881859.0, "step": 2178 }, { "epoch": 0.2771911970487215, "ewc_loss": 0.03955078125, "ewc_loss_diag": 5.340576171875e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.580639600753784, "learning_rate": 9.232725731242053e-06, "loss": 0.536, "mean_token_accuracy": 0.8356035947799683, "num_tokens": 82922146.0, "step": 2179 }, { "epoch": 0.27731840732731206, "ewc_loss": 0.0537109375, "ewc_loss_diag": 5.364418029785156e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.6369400024414062, "learning_rate": 9.236964815599832e-06, "loss": 0.5401, "mean_token_accuracy": 0.8408055901527405, "num_tokens": 82960017.0, "step": 2180 }, { "epoch": 0.2774456176059026, "ewc_loss": 0.043212890625, "ewc_loss_diag": 5.364418029785156e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.78666090965271, "learning_rate": 9.241203899957611e-06, "loss": 0.4683, "mean_token_accuracy": 0.8578048348426819, "num_tokens": 83001778.0, "step": 2181 }, { "epoch": 0.27757282788449306, "ewc_loss": 0.05859375, "ewc_loss_diag": 5.364418029785156e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.7999751567840576, "learning_rate": 9.245442984315388e-06, "loss": 0.4969, "mean_token_accuracy": 0.8531838655471802, "num_tokens": 83037023.0, "step": 2182 }, { "epoch": 0.2777000381630836, "ewc_loss": 0.052734375, "ewc_loss_diag": 5.364418029785156e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1712727546691895, "learning_rate": 9.249682068673167e-06, "loss": 0.5287, "mean_token_accuracy": 0.84278404712677, "num_tokens": 83075355.0, "step": 2183 }, { "epoch": 0.2778272484416741, "ewc_loss": 0.053466796875, "ewc_loss_diag": 5.364418029785156e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.6307103633880615, "learning_rate": 9.253921153030946e-06, "loss": 0.5651, "mean_token_accuracy": 0.8339096903800964, "num_tokens": 83116233.0, "step": 2184 }, { "epoch": 0.2779544587202646, "ewc_loss": 0.05224609375, "ewc_loss_diag": 5.364418029785156e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2662353515625, "learning_rate": 9.258160237388725e-06, "loss": 0.4802, "mean_token_accuracy": 0.8611917495727539, "num_tokens": 83152238.0, "step": 2185 }, { "epoch": 0.2780816689988551, "ewc_loss": 0.0498046875, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.425600528717041, "learning_rate": 9.262399321746504e-06, "loss": 0.4842, "mean_token_accuracy": 0.8533599972724915, "num_tokens": 83186779.0, "step": 2186 }, { "epoch": 0.27820887927744564, "ewc_loss": 0.04931640625, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.180199384689331, "learning_rate": 9.266638406104283e-06, "loss": 0.5455, "mean_token_accuracy": 0.8358173966407776, "num_tokens": 83226001.0, "step": 2187 }, { "epoch": 0.2783360895560361, "ewc_loss": 0.047119140625, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.280052900314331, "learning_rate": 9.270877490462061e-06, "loss": 0.5026, "mean_token_accuracy": 0.8496075868606567, "num_tokens": 83257196.0, "step": 2188 }, { "epoch": 0.27846329983462664, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.0810086727142334, "learning_rate": 9.27511657481984e-06, "loss": 0.505, "mean_token_accuracy": 0.8491314649581909, "num_tokens": 83293261.0, "step": 2189 }, { "epoch": 0.27859051011321717, "ewc_loss": 0.044921875, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.181148052215576, "learning_rate": 9.279355659177618e-06, "loss": 0.5188, "mean_token_accuracy": 0.848613977432251, "num_tokens": 83327096.0, "step": 2190 }, { "epoch": 0.27871772039180764, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 2.9735636711120605, "learning_rate": 9.283594743535396e-06, "loss": 0.5334, "mean_token_accuracy": 0.840409517288208, "num_tokens": 83360558.0, "step": 2191 }, { "epoch": 0.2788449306703982, "ewc_loss": 0.043701171875, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.117077112197876, "learning_rate": 9.287833827893175e-06, "loss": 0.4989, "mean_token_accuracy": 0.8467282056808472, "num_tokens": 83400563.0, "step": 2192 }, { "epoch": 0.2789721409489887, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.8785088062286377, "learning_rate": 9.292072912250954e-06, "loss": 0.5072, "mean_token_accuracy": 0.8449764251708984, "num_tokens": 83432961.0, "step": 2193 }, { "epoch": 0.2790993512275792, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 3.105289936065674, "learning_rate": 9.296311996608733e-06, "loss": 0.4439, "mean_token_accuracy": 0.8640682101249695, "num_tokens": 83472183.0, "step": 2194 }, { "epoch": 0.2792265615061697, "ewc_loss": 0.040283203125, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.7978007793426514, "learning_rate": 9.300551080966512e-06, "loss": 0.5644, "mean_token_accuracy": 0.8276404142379761, "num_tokens": 83512376.0, "step": 2195 }, { "epoch": 0.27935377178476023, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.1723010540008545, "learning_rate": 9.304790165324291e-06, "loss": 0.4894, "mean_token_accuracy": 0.850095272064209, "num_tokens": 83546141.0, "step": 2196 }, { "epoch": 0.2794809820633507, "ewc_loss": 0.039794921875, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.7284910678863525, "learning_rate": 9.30902924968207e-06, "loss": 0.4858, "mean_token_accuracy": 0.8532033562660217, "num_tokens": 83584223.0, "step": 2197 }, { "epoch": 0.27960819234194123, "ewc_loss": 0.046875, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.2950491905212402, "learning_rate": 9.313268334039847e-06, "loss": 0.5142, "mean_token_accuracy": 0.845160961151123, "num_tokens": 83621773.0, "step": 2198 }, { "epoch": 0.27973540262053176, "ewc_loss": 0.039794921875, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.706042528152466, "learning_rate": 9.317507418397626e-06, "loss": 0.4298, "mean_token_accuracy": 0.8713669776916504, "num_tokens": 83659129.0, "step": 2199 }, { "epoch": 0.27986261289912223, "ewc_loss": 0.050048828125, "ewc_loss_diag": 5.3882598876953125e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.4566128253936768, "learning_rate": 9.321746502755405e-06, "loss": 0.5477, "mean_token_accuracy": 0.8392386436462402, "num_tokens": 83699146.0, "step": 2200 }, { "epoch": 0.27998982317771276, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.833038568496704, "learning_rate": 9.325985587113184e-06, "loss": 0.5006, "mean_token_accuracy": 0.849156379699707, "num_tokens": 83733266.0, "step": 2201 }, { "epoch": 0.2801170334563033, "ewc_loss": 0.05078125, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.5251359939575195, "learning_rate": 9.330224671470962e-06, "loss": 0.5152, "mean_token_accuracy": 0.8540917634963989, "num_tokens": 83766610.0, "step": 2202 }, { "epoch": 0.28024424373489376, "ewc_loss": 0.0439453125, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.0010299682617188, "learning_rate": 9.334463755828741e-06, "loss": 0.5263, "mean_token_accuracy": 0.8384772539138794, "num_tokens": 83797891.0, "step": 2203 }, { "epoch": 0.2803714540134843, "ewc_loss": 0.0498046875, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.410841226577759, "learning_rate": 9.33870284018652e-06, "loss": 0.5318, "mean_token_accuracy": 0.8387168049812317, "num_tokens": 83837046.0, "step": 2204 }, { "epoch": 0.2804986642920748, "ewc_loss": 0.044921875, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.0677664279937744, "learning_rate": 9.342941924544299e-06, "loss": 0.5192, "mean_token_accuracy": 0.8434920310974121, "num_tokens": 83876457.0, "step": 2205 }, { "epoch": 0.2806258745706653, "ewc_loss": 0.046142578125, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.236414670944214, "learning_rate": 9.347181008902078e-06, "loss": 0.46, "mean_token_accuracy": 0.8646658658981323, "num_tokens": 83912729.0, "step": 2206 }, { "epoch": 0.2807530848492558, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.0741593837738037, "learning_rate": 9.351420093259857e-06, "loss": 0.4578, "mean_token_accuracy": 0.8626606464385986, "num_tokens": 83945958.0, "step": 2207 }, { "epoch": 0.28088029512784635, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 3.0710277557373047, "learning_rate": 9.355659177617636e-06, "loss": 0.4451, "mean_token_accuracy": 0.8612892031669617, "num_tokens": 83976767.0, "step": 2208 }, { "epoch": 0.2810075054064368, "ewc_loss": 0.043212890625, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 3.0320074558258057, "learning_rate": 9.359898261975415e-06, "loss": 0.4402, "mean_token_accuracy": 0.867220938205719, "num_tokens": 84017996.0, "step": 2209 }, { "epoch": 0.28113471568502735, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.94264554977417, "learning_rate": 9.364137346333194e-06, "loss": 0.5675, "mean_token_accuracy": 0.8318987488746643, "num_tokens": 84057528.0, "step": 2210 }, { "epoch": 0.2812619259636179, "ewc_loss": 0.04248046875, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 3.02441668510437, "learning_rate": 9.368376430690972e-06, "loss": 0.4538, "mean_token_accuracy": 0.8631348013877869, "num_tokens": 84091009.0, "step": 2211 }, { "epoch": 0.28138913624220835, "ewc_loss": 0.040771484375, "ewc_loss_diag": 5.412101745605469e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.862504005432129, "learning_rate": 9.372615515048751e-06, "loss": 0.5277, "mean_token_accuracy": 0.8391982316970825, "num_tokens": 84125042.0, "step": 2212 }, { "epoch": 0.2815163465207989, "ewc_loss": 0.0419921875, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 3.0190587043762207, "learning_rate": 9.376854599406528e-06, "loss": 0.5324, "mean_token_accuracy": 0.8401257395744324, "num_tokens": 84169340.0, "step": 2213 }, { "epoch": 0.2816435567993894, "ewc_loss": 0.040283203125, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.8014373779296875, "learning_rate": 9.381093683764307e-06, "loss": 0.4911, "mean_token_accuracy": 0.848655104637146, "num_tokens": 84205347.0, "step": 2214 }, { "epoch": 0.2817707670779799, "ewc_loss": 0.042236328125, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 3.0378408432006836, "learning_rate": 9.385332768122086e-06, "loss": 0.4806, "mean_token_accuracy": 0.8512670993804932, "num_tokens": 84245788.0, "step": 2215 }, { "epoch": 0.2818979773565704, "ewc_loss": 0.03955078125, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 3.409385681152344e-05, "grad_norm": 2.734818935394287, "learning_rate": 9.389571852479865e-06, "loss": 0.4815, "mean_token_accuracy": 0.8560864925384521, "num_tokens": 84286744.0, "step": 2216 }, { "epoch": 0.28202518763516093, "ewc_loss": 0.043701171875, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.1241259574890137, "learning_rate": 9.393810936837644e-06, "loss": 0.4924, "mean_token_accuracy": 0.8488907814025879, "num_tokens": 84326031.0, "step": 2217 }, { "epoch": 0.2821523979137514, "ewc_loss": 0.038818359375, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 3.337860107421875e-05, "grad_norm": 2.679560899734497, "learning_rate": 9.398050021195423e-06, "loss": 0.5152, "mean_token_accuracy": 0.8446236848831177, "num_tokens": 84367697.0, "step": 2218 }, { "epoch": 0.28227960819234194, "ewc_loss": 0.0458984375, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.2410337924957275, "learning_rate": 9.402289105553202e-06, "loss": 0.4689, "mean_token_accuracy": 0.8580324649810791, "num_tokens": 84410647.0, "step": 2219 }, { "epoch": 0.28240681847093246, "ewc_loss": 0.0390625, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 3.361701965332031e-05, "grad_norm": 2.663285970687866, "learning_rate": 9.406528189910979e-06, "loss": 0.487, "mean_token_accuracy": 0.8533164262771606, "num_tokens": 84446839.0, "step": 2220 }, { "epoch": 0.28253402874952294, "ewc_loss": 0.049560546875, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.4018023014068604, "learning_rate": 9.410767274268758e-06, "loss": 0.4725, "mean_token_accuracy": 0.8565544486045837, "num_tokens": 84485017.0, "step": 2221 }, { "epoch": 0.28266123902811346, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.459785461425781e-05, "ewc_loss_parallel": 3.504753112792969e-05, "grad_norm": 2.7144181728363037, "learning_rate": 9.415006358626537e-06, "loss": 0.5652, "mean_token_accuracy": 0.8258537650108337, "num_tokens": 84524518.0, "step": 2222 }, { "epoch": 0.282788449306704, "ewc_loss": 0.052734375, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.565424919128418, "learning_rate": 9.419245442984316e-06, "loss": 0.5602, "mean_token_accuracy": 0.8345211744308472, "num_tokens": 84559850.0, "step": 2223 }, { "epoch": 0.28291565958529447, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.868960380554199, "learning_rate": 9.423484527342095e-06, "loss": 0.4896, "mean_token_accuracy": 0.8495784997940063, "num_tokens": 84594332.0, "step": 2224 }, { "epoch": 0.283042869863885, "ewc_loss": 0.05419921875, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5952742099761963, "learning_rate": 9.427723611699873e-06, "loss": 0.5072, "mean_token_accuracy": 0.850199818611145, "num_tokens": 84628759.0, "step": 2225 }, { "epoch": 0.2831700801424755, "ewc_loss": 0.048828125, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0307767391204834, "learning_rate": 9.431962696057652e-06, "loss": 0.5197, "mean_token_accuracy": 0.8449074029922485, "num_tokens": 84672022.0, "step": 2226 }, { "epoch": 0.283297290421066, "ewc_loss": 0.0517578125, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.496293067932129, "learning_rate": 9.436201780415431e-06, "loss": 0.4683, "mean_token_accuracy": 0.8636415004730225, "num_tokens": 84709266.0, "step": 2227 }, { "epoch": 0.2834245006996565, "ewc_loss": 0.048828125, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1030328273773193, "learning_rate": 9.44044086477321e-06, "loss": 0.5162, "mean_token_accuracy": 0.8482571840286255, "num_tokens": 84745502.0, "step": 2228 }, { "epoch": 0.28355171097824705, "ewc_loss": 0.0498046875, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.344291925430298, "learning_rate": 9.444679949130989e-06, "loss": 0.5212, "mean_token_accuracy": 0.8442501425743103, "num_tokens": 84786061.0, "step": 2229 }, { "epoch": 0.2836789212568376, "ewc_loss": 0.046875, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0864510536193848, "learning_rate": 9.448919033488768e-06, "loss": 0.5306, "mean_token_accuracy": 0.8414848446846008, "num_tokens": 84822550.0, "step": 2230 }, { "epoch": 0.28380613153542805, "ewc_loss": 0.04638671875, "ewc_loss_diag": 5.4836273193359375e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.203526258468628, "learning_rate": 9.453158117846547e-06, "loss": 0.5184, "mean_token_accuracy": 0.8443305492401123, "num_tokens": 84860608.0, "step": 2231 }, { "epoch": 0.2839333418140186, "ewc_loss": 0.044921875, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.030385971069336, "learning_rate": 9.457397202204326e-06, "loss": 0.5439, "mean_token_accuracy": 0.8368475437164307, "num_tokens": 84903065.0, "step": 2232 }, { "epoch": 0.2840605520926091, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.1020185947418213, "learning_rate": 9.461636286562103e-06, "loss": 0.4672, "mean_token_accuracy": 0.8614925146102905, "num_tokens": 84938583.0, "step": 2233 }, { "epoch": 0.2841877623711996, "ewc_loss": 0.043212890625, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.9804067611694336, "learning_rate": 9.465875370919882e-06, "loss": 0.5486, "mean_token_accuracy": 0.8363698720932007, "num_tokens": 84975939.0, "step": 2234 }, { "epoch": 0.2843149726497901, "ewc_loss": 0.043212890625, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.0446536540985107, "learning_rate": 9.47011445527766e-06, "loss": 0.5171, "mean_token_accuracy": 0.8428019881248474, "num_tokens": 85011304.0, "step": 2235 }, { "epoch": 0.28444218292838064, "ewc_loss": 0.0419921875, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 2.9174652099609375, "learning_rate": 9.47435353963544e-06, "loss": 0.4818, "mean_token_accuracy": 0.8527349829673767, "num_tokens": 85053163.0, "step": 2236 }, { "epoch": 0.2845693932069711, "ewc_loss": 0.04296875, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 3.02921986579895, "learning_rate": 9.478592623993218e-06, "loss": 0.4829, "mean_token_accuracy": 0.851230263710022, "num_tokens": 85085919.0, "step": 2237 }, { "epoch": 0.28469660348556164, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.8604307174682617, "learning_rate": 9.482831708350997e-06, "loss": 0.4843, "mean_token_accuracy": 0.8534456491470337, "num_tokens": 85124323.0, "step": 2238 }, { "epoch": 0.28482381376415217, "ewc_loss": 0.04296875, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 3.044611692428589, "learning_rate": 9.487070792708776e-06, "loss": 0.5271, "mean_token_accuracy": 0.8439606428146362, "num_tokens": 85163735.0, "step": 2239 }, { "epoch": 0.28495102404274264, "ewc_loss": 0.041259765625, "ewc_loss_diag": 5.53131103515625e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.802492380142212, "learning_rate": 9.491309877066555e-06, "loss": 0.5137, "mean_token_accuracy": 0.8436222076416016, "num_tokens": 85206533.0, "step": 2240 }, { "epoch": 0.28507823432133317, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.5789947509765625e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 3.1060903072357178, "learning_rate": 9.495548961424334e-06, "loss": 0.5376, "mean_token_accuracy": 0.8380714654922485, "num_tokens": 85248435.0, "step": 2241 }, { "epoch": 0.2852054445999237, "ewc_loss": 0.040283203125, "ewc_loss_diag": 5.5789947509765625e-05, "ewc_loss_parallel": 3.457069396972656e-05, "grad_norm": 2.75535249710083, "learning_rate": 9.499788045782113e-06, "loss": 0.497, "mean_token_accuracy": 0.8504428863525391, "num_tokens": 85285686.0, "step": 2242 }, { "epoch": 0.28533265487851417, "ewc_loss": 0.045166015625, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.193619966506958, "learning_rate": 9.50402713013989e-06, "loss": 0.515, "mean_token_accuracy": 0.844717264175415, "num_tokens": 85327518.0, "step": 2243 }, { "epoch": 0.2854598651571047, "ewc_loss": 0.0400390625, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.4332275390625e-05, "grad_norm": 2.715941905975342, "learning_rate": 9.508266214497669e-06, "loss": 0.4623, "mean_token_accuracy": 0.8604629635810852, "num_tokens": 85364309.0, "step": 2244 }, { "epoch": 0.2855870754356952, "ewc_loss": 0.0478515625, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.3217315673828125, "learning_rate": 9.512505298855448e-06, "loss": 0.5051, "mean_token_accuracy": 0.8525391817092896, "num_tokens": 85407231.0, "step": 2245 }, { "epoch": 0.2857142857142857, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.7070701122283936, "learning_rate": 9.516744383213227e-06, "loss": 0.4401, "mean_token_accuracy": 0.8659715056419373, "num_tokens": 85443753.0, "step": 2246 }, { "epoch": 0.2858414959928762, "ewc_loss": 0.0517578125, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.487210273742676, "learning_rate": 9.520983467571005e-06, "loss": 0.5309, "mean_token_accuracy": 0.8414747714996338, "num_tokens": 85482874.0, "step": 2247 }, { "epoch": 0.28596870627146675, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.801511526107788, "learning_rate": 9.525222551928784e-06, "loss": 0.5572, "mean_token_accuracy": 0.8385993242263794, "num_tokens": 85521648.0, "step": 2248 }, { "epoch": 0.2860959165500572, "ewc_loss": 0.053466796875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.593747854232788, "learning_rate": 9.529461636286563e-06, "loss": 0.4825, "mean_token_accuracy": 0.8576446771621704, "num_tokens": 85553078.0, "step": 2249 }, { "epoch": 0.28622312682864776, "ewc_loss": 0.045654296875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.963048219680786, "learning_rate": 9.533700720644342e-06, "loss": 0.4537, "mean_token_accuracy": 0.8627949953079224, "num_tokens": 85589034.0, "step": 2250 }, { "epoch": 0.2863503371072383, "ewc_loss": 0.053466796875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.5618021488189697, "learning_rate": 9.53793980500212e-06, "loss": 0.523, "mean_token_accuracy": 0.8485108613967896, "num_tokens": 85631541.0, "step": 2251 }, { "epoch": 0.28647754738582876, "ewc_loss": 0.049560546875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.095329523086548, "learning_rate": 9.542178889359898e-06, "loss": 0.4892, "mean_token_accuracy": 0.8553940057754517, "num_tokens": 85668498.0, "step": 2252 }, { "epoch": 0.2866047576644193, "ewc_loss": 0.05029296875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.4111907482147217, "learning_rate": 9.546417973717677e-06, "loss": 0.5234, "mean_token_accuracy": 0.8459769487380981, "num_tokens": 85708419.0, "step": 2253 }, { "epoch": 0.2867319679430098, "ewc_loss": 0.048583984375, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1061060428619385, "learning_rate": 9.550657058075456e-06, "loss": 0.5295, "mean_token_accuracy": 0.8438501954078674, "num_tokens": 85751564.0, "step": 2254 }, { "epoch": 0.2868591782216003, "ewc_loss": 0.047607421875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.2537174224853516, "learning_rate": 9.554896142433235e-06, "loss": 0.5312, "mean_token_accuracy": 0.8393014073371887, "num_tokens": 85789776.0, "step": 2255 }, { "epoch": 0.2869863885001908, "ewc_loss": 0.0458984375, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0805106163024902, "learning_rate": 9.559135226791014e-06, "loss": 0.5421, "mean_token_accuracy": 0.8341079354286194, "num_tokens": 85828427.0, "step": 2256 }, { "epoch": 0.28711359877878134, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0949671268463135, "learning_rate": 9.563374311148793e-06, "loss": 0.5337, "mean_token_accuracy": 0.8400989770889282, "num_tokens": 85877598.0, "step": 2257 }, { "epoch": 0.2872408090573718, "ewc_loss": 0.044189453125, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.0544698238372803, "learning_rate": 9.567613395506572e-06, "loss": 0.4805, "mean_token_accuracy": 0.8564489483833313, "num_tokens": 85908395.0, "step": 2258 }, { "epoch": 0.28736801933596234, "ewc_loss": 0.04296875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.996060371398926, "learning_rate": 9.57185247986435e-06, "loss": 0.4668, "mean_token_accuracy": 0.8589218854904175, "num_tokens": 85945344.0, "step": 2259 }, { "epoch": 0.28749522961455287, "ewc_loss": 0.043212890625, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.0165293216705322, "learning_rate": 9.57609156422213e-06, "loss": 0.435, "mean_token_accuracy": 0.8672041296958923, "num_tokens": 85982484.0, "step": 2260 }, { "epoch": 0.28762243989314334, "ewc_loss": 0.042236328125, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.940178632736206, "learning_rate": 9.580330648579908e-06, "loss": 0.5185, "mean_token_accuracy": 0.8411810398101807, "num_tokens": 86019496.0, "step": 2261 }, { "epoch": 0.2877496501717339, "ewc_loss": 0.04296875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.9971132278442383, "learning_rate": 9.584569732937687e-06, "loss": 0.5116, "mean_token_accuracy": 0.8470748662948608, "num_tokens": 86059344.0, "step": 2262 }, { "epoch": 0.2878768604503244, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.9036073684692383, "learning_rate": 9.588808817295466e-06, "loss": 0.497, "mean_token_accuracy": 0.8460332155227661, "num_tokens": 86093090.0, "step": 2263 }, { "epoch": 0.2880040707289149, "ewc_loss": 0.042724609375, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.995349645614624, "learning_rate": 9.593047901653245e-06, "loss": 0.4944, "mean_token_accuracy": 0.8507283926010132, "num_tokens": 86127149.0, "step": 2264 }, { "epoch": 0.2881312810075054, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.8685805797576904, "learning_rate": 9.597286986011024e-06, "loss": 0.4539, "mean_token_accuracy": 0.861869215965271, "num_tokens": 86163541.0, "step": 2265 }, { "epoch": 0.28825849128609593, "ewc_loss": 0.043212890625, "ewc_loss_diag": 5.602836608886719e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.0134353637695312, "learning_rate": 9.601526070368801e-06, "loss": 0.5381, "mean_token_accuracy": 0.8391884565353394, "num_tokens": 86200124.0, "step": 2266 }, { "epoch": 0.2883857015646864, "ewc_loss": 0.041259765625, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.8375959396362305, "learning_rate": 9.60576515472658e-06, "loss": 0.4888, "mean_token_accuracy": 0.8525571823120117, "num_tokens": 86236464.0, "step": 2267 }, { "epoch": 0.28851291184327693, "ewc_loss": 0.043212890625, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.042074680328369, "learning_rate": 9.610004239084359e-06, "loss": 0.5033, "mean_token_accuracy": 0.8475383520126343, "num_tokens": 86277504.0, "step": 2268 }, { "epoch": 0.28864012212186746, "ewc_loss": 0.041259765625, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.80043363571167, "learning_rate": 9.614243323442138e-06, "loss": 0.5649, "mean_token_accuracy": 0.8278383016586304, "num_tokens": 86314404.0, "step": 2269 }, { "epoch": 0.28876733240045793, "ewc_loss": 0.0439453125, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.1087210178375244, "learning_rate": 9.618482407799916e-06, "loss": 0.4779, "mean_token_accuracy": 0.8527416586875916, "num_tokens": 86350690.0, "step": 2270 }, { "epoch": 0.28889454267904846, "ewc_loss": 0.041015625, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 3.528594970703125e-05, "grad_norm": 2.7290852069854736, "learning_rate": 9.622721492157695e-06, "loss": 0.4397, "mean_token_accuracy": 0.8670610785484314, "num_tokens": 86389294.0, "step": 2271 }, { "epoch": 0.289021752957639, "ewc_loss": 0.046630859375, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.2559497356414795, "learning_rate": 9.626960576515474e-06, "loss": 0.5003, "mean_token_accuracy": 0.8487051725387573, "num_tokens": 86420234.0, "step": 2272 }, { "epoch": 0.28914896323622946, "ewc_loss": 0.04052734375, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 3.4809112548828125e-05, "grad_norm": 2.658886194229126, "learning_rate": 9.631199660873251e-06, "loss": 0.4727, "mean_token_accuracy": 0.8534080982208252, "num_tokens": 86458435.0, "step": 2273 }, { "epoch": 0.28927617351482, "ewc_loss": 0.051025390625, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.4671103954315186, "learning_rate": 9.63543874523103e-06, "loss": 0.488, "mean_token_accuracy": 0.8554394245147705, "num_tokens": 86499082.0, "step": 2274 }, { "epoch": 0.2894033837934105, "ewc_loss": 0.041259765625, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.668295383453369, "learning_rate": 9.63967782958881e-06, "loss": 0.5114, "mean_token_accuracy": 0.8410595655441284, "num_tokens": 86536989.0, "step": 2275 }, { "epoch": 0.289530594072001, "ewc_loss": 0.058349609375, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.735705852508545, "learning_rate": 9.643916913946588e-06, "loss": 0.4829, "mean_token_accuracy": 0.8581534624099731, "num_tokens": 86572154.0, "step": 2276 }, { "epoch": 0.2896578043505915, "ewc_loss": 0.04638671875, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.867727041244507, "learning_rate": 9.648155998304367e-06, "loss": 0.4491, "mean_token_accuracy": 0.8648899793624878, "num_tokens": 86610908.0, "step": 2277 }, { "epoch": 0.28978501462918205, "ewc_loss": 0.061767578125, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.883253812789917, "learning_rate": 9.652395082662146e-06, "loss": 0.5819, "mean_token_accuracy": 0.8282077312469482, "num_tokens": 86653585.0, "step": 2278 }, { "epoch": 0.2899122249077726, "ewc_loss": 0.054931640625, "ewc_loss_diag": 5.626678466796875e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2194623947143555, "learning_rate": 9.656634167019925e-06, "loss": 0.5122, "mean_token_accuracy": 0.8467525243759155, "num_tokens": 86690794.0, "step": 2279 }, { "epoch": 0.29003943518636305, "ewc_loss": 0.057373046875, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.740966320037842, "learning_rate": 9.660873251377704e-06, "loss": 0.5324, "mean_token_accuracy": 0.8444174528121948, "num_tokens": 86729742.0, "step": 2280 }, { "epoch": 0.2901666454649536, "ewc_loss": 0.055908203125, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.338287115097046, "learning_rate": 9.66511233573548e-06, "loss": 0.5425, "mean_token_accuracy": 0.8445630073547363, "num_tokens": 86765837.0, "step": 2281 }, { "epoch": 0.2902938557435441, "ewc_loss": 0.052978515625, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.52805495262146, "learning_rate": 9.66935142009326e-06, "loss": 0.4764, "mean_token_accuracy": 0.8579511046409607, "num_tokens": 86799697.0, "step": 2282 }, { "epoch": 0.2904210660221346, "ewc_loss": 0.05224609375, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2772250175476074, "learning_rate": 9.673590504451039e-06, "loss": 0.4998, "mean_token_accuracy": 0.8508479595184326, "num_tokens": 86832264.0, "step": 2283 }, { "epoch": 0.2905482763007251, "ewc_loss": 0.0498046875, "ewc_loss_diag": 5.650520324707031e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.3557674884796143, "learning_rate": 9.677829588808817e-06, "loss": 0.4818, "mean_token_accuracy": 0.8506854772567749, "num_tokens": 86864492.0, "step": 2284 }, { "epoch": 0.29067548657931563, "ewc_loss": 0.048095703125, "ewc_loss_diag": 5.6743621826171875e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1681785583496094, "learning_rate": 9.682068673166596e-06, "loss": 0.5533, "mean_token_accuracy": 0.8392611742019653, "num_tokens": 86898186.0, "step": 2285 }, { "epoch": 0.2908026968579061, "ewc_loss": 0.04736328125, "ewc_loss_diag": 5.6743621826171875e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.249007225036621, "learning_rate": 9.686307757524375e-06, "loss": 0.504, "mean_token_accuracy": 0.8481602072715759, "num_tokens": 86935850.0, "step": 2286 }, { "epoch": 0.29092990713649663, "ewc_loss": 0.045166015625, "ewc_loss_diag": 5.6743621826171875e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.048708438873291, "learning_rate": 9.690546841882154e-06, "loss": 0.4284, "mean_token_accuracy": 0.8720294833183289, "num_tokens": 86972732.0, "step": 2287 }, { "epoch": 0.29105711741508716, "ewc_loss": 0.04638671875, "ewc_loss_diag": 5.6743621826171875e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.201148271560669, "learning_rate": 9.694785926239933e-06, "loss": 0.5078, "mean_token_accuracy": 0.8481155037879944, "num_tokens": 87012538.0, "step": 2288 }, { "epoch": 0.29118432769367764, "ewc_loss": 0.042724609375, "ewc_loss_diag": 5.6743621826171875e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.9687604904174805, "learning_rate": 9.699025010597712e-06, "loss": 0.5474, "mean_token_accuracy": 0.8346166014671326, "num_tokens": 87046917.0, "step": 2289 }, { "epoch": 0.29131153797226816, "ewc_loss": 0.0458984375, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.196913480758667, "learning_rate": 9.70326409495549e-06, "loss": 0.5129, "mean_token_accuracy": 0.8449519872665405, "num_tokens": 87086002.0, "step": 2290 }, { "epoch": 0.2914387482508587, "ewc_loss": 0.04248046875, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.8775954246520996, "learning_rate": 9.70750317931327e-06, "loss": 0.491, "mean_token_accuracy": 0.8535658121109009, "num_tokens": 87126138.0, "step": 2291 }, { "epoch": 0.29156595852944917, "ewc_loss": 0.046875, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.258957862854004, "learning_rate": 9.711742263671049e-06, "loss": 0.488, "mean_token_accuracy": 0.8528062105178833, "num_tokens": 87163062.0, "step": 2292 }, { "epoch": 0.2916931688080397, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.8039040565490723, "learning_rate": 9.715981348028827e-06, "loss": 0.4739, "mean_token_accuracy": 0.855462908744812, "num_tokens": 87203836.0, "step": 2293 }, { "epoch": 0.2918203790866302, "ewc_loss": 0.0498046875, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.3988354206085205, "learning_rate": 9.720220432386606e-06, "loss": 0.4864, "mean_token_accuracy": 0.8564329147338867, "num_tokens": 87236854.0, "step": 2294 }, { "epoch": 0.2919475893652207, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.778541326522827, "learning_rate": 9.724459516744385e-06, "loss": 0.4648, "mean_token_accuracy": 0.8611499071121216, "num_tokens": 87276316.0, "step": 2295 }, { "epoch": 0.2920747996438112, "ewc_loss": 0.0537109375, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.568844795227051, "learning_rate": 9.728698601102162e-06, "loss": 0.4721, "mean_token_accuracy": 0.8617143630981445, "num_tokens": 87314504.0, "step": 2296 }, { "epoch": 0.29220200992240175, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.909531831741333, "learning_rate": 9.732937685459941e-06, "loss": 0.4922, "mean_token_accuracy": 0.849144458770752, "num_tokens": 87348403.0, "step": 2297 }, { "epoch": 0.2923292202009922, "ewc_loss": 0.0556640625, "ewc_loss_diag": 5.7220458984375e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.63478684425354, "learning_rate": 9.73717676981772e-06, "loss": 0.5436, "mean_token_accuracy": 0.8432765007019043, "num_tokens": 87393044.0, "step": 2298 }, { "epoch": 0.29245643047958275, "ewc_loss": 0.04736328125, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.054409980773926, "learning_rate": 9.741415854175499e-06, "loss": 0.558, "mean_token_accuracy": 0.8365262746810913, "num_tokens": 87430930.0, "step": 2299 }, { "epoch": 0.2925836407581733, "ewc_loss": 0.05419921875, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.569540500640869, "learning_rate": 9.745654938533278e-06, "loss": 0.5279, "mean_token_accuracy": 0.8416910171508789, "num_tokens": 87467941.0, "step": 2300 }, { "epoch": 0.29271085103676375, "ewc_loss": 0.05029296875, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.164062738418579, "learning_rate": 9.749894022891057e-06, "loss": 0.5456, "mean_token_accuracy": 0.8394233584403992, "num_tokens": 87507453.0, "step": 2301 }, { "epoch": 0.2928380613153543, "ewc_loss": 0.0498046875, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.37717866897583, "learning_rate": 9.754133107248836e-06, "loss": 0.461, "mean_token_accuracy": 0.8617148399353027, "num_tokens": 87545983.0, "step": 2302 }, { "epoch": 0.2929652715939448, "ewc_loss": 0.04931640625, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.180974006652832, "learning_rate": 9.758372191606615e-06, "loss": 0.5453, "mean_token_accuracy": 0.8361178636550903, "num_tokens": 87582352.0, "step": 2303 }, { "epoch": 0.2930924818725353, "ewc_loss": 0.04638671875, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1861684322357178, "learning_rate": 9.762611275964392e-06, "loss": 0.4533, "mean_token_accuracy": 0.864978015422821, "num_tokens": 87616668.0, "step": 2304 }, { "epoch": 0.2932196921511258, "ewc_loss": 0.046875, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.162601947784424, "learning_rate": 9.76685036032217e-06, "loss": 0.5234, "mean_token_accuracy": 0.8426322937011719, "num_tokens": 87653008.0, "step": 2305 }, { "epoch": 0.29334690242971634, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.0211405754089355, "learning_rate": 9.77108944467995e-06, "loss": 0.4694, "mean_token_accuracy": 0.8576090335845947, "num_tokens": 87694985.0, "step": 2306 }, { "epoch": 0.2934741127083068, "ewc_loss": 0.04541015625, "ewc_loss_diag": 5.745887756347656e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.166834592819214, "learning_rate": 9.775328529037728e-06, "loss": 0.515, "mean_token_accuracy": 0.8443692922592163, "num_tokens": 87730841.0, "step": 2307 }, { "epoch": 0.29360132298689734, "ewc_loss": 0.04248046875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.9377870559692383, "learning_rate": 9.779567613395507e-06, "loss": 0.5445, "mean_token_accuracy": 0.8399220705032349, "num_tokens": 87762215.0, "step": 2308 }, { "epoch": 0.29372853326548787, "ewc_loss": 0.04541015625, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.176922559738159, "learning_rate": 9.783806697753286e-06, "loss": 0.4896, "mean_token_accuracy": 0.853495717048645, "num_tokens": 87799150.0, "step": 2309 }, { "epoch": 0.29385574354407834, "ewc_loss": 0.0419921875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.832639694213867, "learning_rate": 9.788045782111065e-06, "loss": 0.4528, "mean_token_accuracy": 0.861248254776001, "num_tokens": 87842463.0, "step": 2310 }, { "epoch": 0.29398295382266887, "ewc_loss": 0.046875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.2482616901397705, "learning_rate": 9.792284866468842e-06, "loss": 0.4499, "mean_token_accuracy": 0.8621444702148438, "num_tokens": 87880093.0, "step": 2311 }, { "epoch": 0.2941101641012594, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.7915186882019043, "learning_rate": 9.796523950826621e-06, "loss": 0.45, "mean_token_accuracy": 0.8612730503082275, "num_tokens": 87912713.0, "step": 2312 }, { "epoch": 0.29423737437984987, "ewc_loss": 0.0498046875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.3849809169769287, "learning_rate": 9.8007630351844e-06, "loss": 0.5323, "mean_token_accuracy": 0.8426889181137085, "num_tokens": 87946357.0, "step": 2313 }, { "epoch": 0.2943645846584404, "ewc_loss": 0.04150390625, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 3.5762786865234375e-05, "grad_norm": 2.7761855125427246, "learning_rate": 9.805002119542179e-06, "loss": 0.5537, "mean_token_accuracy": 0.8313146829605103, "num_tokens": 87981102.0, "step": 2314 }, { "epoch": 0.2944917949370309, "ewc_loss": 0.052978515625, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.5207107067108154, "learning_rate": 9.809241203899958e-06, "loss": 0.5006, "mean_token_accuracy": 0.8539937138557434, "num_tokens": 88022325.0, "step": 2315 }, { "epoch": 0.2946190052156214, "ewc_loss": 0.04248046875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.813528060913086, "learning_rate": 9.813480288257737e-06, "loss": 0.5124, "mean_token_accuracy": 0.8410844206809998, "num_tokens": 88057254.0, "step": 2316 }, { "epoch": 0.2947462154942119, "ewc_loss": 0.057861328125, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.666879653930664, "learning_rate": 9.817719372615516e-06, "loss": 0.4827, "mean_token_accuracy": 0.8589469194412231, "num_tokens": 88095667.0, "step": 2317 }, { "epoch": 0.29487342577280246, "ewc_loss": 0.046875, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.970189332962036, "learning_rate": 9.821958456973294e-06, "loss": 0.4419, "mean_token_accuracy": 0.8666999340057373, "num_tokens": 88129634.0, "step": 2318 }, { "epoch": 0.29500063605139293, "ewc_loss": 0.058349609375, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.690908193588257, "learning_rate": 9.826197541331073e-06, "loss": 0.4682, "mean_token_accuracy": 0.862029492855072, "num_tokens": 88164592.0, "step": 2319 }, { "epoch": 0.29512784632998346, "ewc_loss": 0.0517578125, "ewc_loss_diag": 5.7697296142578125e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1340692043304443, "learning_rate": 9.830436625688852e-06, "loss": 0.518, "mean_token_accuracy": 0.8469694256782532, "num_tokens": 88202390.0, "step": 2320 }, { "epoch": 0.295255056608574, "ewc_loss": 0.055908203125, "ewc_loss_diag": 5.817413330078125e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5722551345825195, "learning_rate": 9.834675710046631e-06, "loss": 0.4778, "mean_token_accuracy": 0.8588938117027283, "num_tokens": 88243104.0, "step": 2321 }, { "epoch": 0.29538226688716446, "ewc_loss": 0.0517578125, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.185980796813965, "learning_rate": 9.83891479440441e-06, "loss": 0.4622, "mean_token_accuracy": 0.8651919364929199, "num_tokens": 88284038.0, "step": 2322 }, { "epoch": 0.295509477165755, "ewc_loss": 0.0517578125, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.424372673034668, "learning_rate": 9.843153878762189e-06, "loss": 0.5379, "mean_token_accuracy": 0.8437752723693848, "num_tokens": 88320705.0, "step": 2323 }, { "epoch": 0.2956366874443455, "ewc_loss": 0.05126953125, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1646838188171387, "learning_rate": 9.847392963119968e-06, "loss": 0.485, "mean_token_accuracy": 0.8558777570724487, "num_tokens": 88359636.0, "step": 2324 }, { "epoch": 0.295763897722936, "ewc_loss": 0.048828125, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.270967960357666, "learning_rate": 9.851632047477747e-06, "loss": 0.5049, "mean_token_accuracy": 0.8464891314506531, "num_tokens": 88402029.0, "step": 2325 }, { "epoch": 0.2958911080015265, "ewc_loss": 0.047119140625, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.1111369132995605, "learning_rate": 9.855871131835524e-06, "loss": 0.4244, "mean_token_accuracy": 0.8713739514350891, "num_tokens": 88438245.0, "step": 2326 }, { "epoch": 0.29601831828011704, "ewc_loss": 0.04638671875, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.1707637310028076, "learning_rate": 9.860110216193303e-06, "loss": 0.4618, "mean_token_accuracy": 0.861852765083313, "num_tokens": 88477987.0, "step": 2327 }, { "epoch": 0.2961455285587075, "ewc_loss": 0.04541015625, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.066955327987671, "learning_rate": 9.864349300551082e-06, "loss": 0.452, "mean_token_accuracy": 0.8646234273910522, "num_tokens": 88510759.0, "step": 2328 }, { "epoch": 0.29627273883729804, "ewc_loss": 0.044921875, "ewc_loss_diag": 5.841255187988281e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.118950843811035, "learning_rate": 9.86858838490886e-06, "loss": 0.4666, "mean_token_accuracy": 0.8573765754699707, "num_tokens": 88544819.0, "step": 2329 }, { "epoch": 0.2963999491158886, "ewc_loss": 0.0439453125, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 3.006479263305664, "learning_rate": 9.87282746926664e-06, "loss": 0.469, "mean_token_accuracy": 0.8575915098190308, "num_tokens": 88584560.0, "step": 2330 }, { "epoch": 0.2965271593944791, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.095140218734741, "learning_rate": 9.877066553624418e-06, "loss": 0.4656, "mean_token_accuracy": 0.8582388162612915, "num_tokens": 88620733.0, "step": 2331 }, { "epoch": 0.2966543696730696, "ewc_loss": 0.043212890625, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.9419028759002686, "learning_rate": 9.881305637982197e-06, "loss": 0.4809, "mean_token_accuracy": 0.854576826095581, "num_tokens": 88661306.0, "step": 2332 }, { "epoch": 0.2967815799516601, "ewc_loss": 0.044921875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.1138598918914795, "learning_rate": 9.885544722339976e-06, "loss": 0.4633, "mean_token_accuracy": 0.8597756028175354, "num_tokens": 88696886.0, "step": 2333 }, { "epoch": 0.29690879023025063, "ewc_loss": 0.04248046875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.8953354358673096, "learning_rate": 9.889783806697753e-06, "loss": 0.4383, "mean_token_accuracy": 0.8643293976783752, "num_tokens": 88733242.0, "step": 2334 }, { "epoch": 0.2970360005088411, "ewc_loss": 0.045166015625, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.1710824966430664, "learning_rate": 9.894022891055532e-06, "loss": 0.4428, "mean_token_accuracy": 0.8665593862533569, "num_tokens": 88769599.0, "step": 2335 }, { "epoch": 0.29716321078743163, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.8355023860931396, "learning_rate": 9.898261975413311e-06, "loss": 0.4237, "mean_token_accuracy": 0.8717291355133057, "num_tokens": 88802885.0, "step": 2336 }, { "epoch": 0.29729042106602216, "ewc_loss": 0.04736328125, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.2747957706451416, "learning_rate": 9.90250105977109e-06, "loss": 0.4859, "mean_token_accuracy": 0.8556232452392578, "num_tokens": 88837061.0, "step": 2337 }, { "epoch": 0.29741763134461263, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.778639078140259, "learning_rate": 9.906740144128869e-06, "loss": 0.4899, "mean_token_accuracy": 0.8479078412055969, "num_tokens": 88874672.0, "step": 2338 }, { "epoch": 0.29754484162320316, "ewc_loss": 0.05078125, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.435248613357544, "learning_rate": 9.910979228486648e-06, "loss": 0.4501, "mean_token_accuracy": 0.8640696406364441, "num_tokens": 88910520.0, "step": 2339 }, { "epoch": 0.2976720519017937, "ewc_loss": 0.0419921875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.768256902694702, "learning_rate": 9.915218312844427e-06, "loss": 0.5122, "mean_token_accuracy": 0.8480992913246155, "num_tokens": 88949495.0, "step": 2340 }, { "epoch": 0.29779926218038416, "ewc_loss": 0.055419921875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.6257715225219727, "learning_rate": 9.919457397202205e-06, "loss": 0.5228, "mean_token_accuracy": 0.8449757695198059, "num_tokens": 88988776.0, "step": 2341 }, { "epoch": 0.2979264724589747, "ewc_loss": 0.0439453125, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.8704516887664795, "learning_rate": 9.923696481559984e-06, "loss": 0.4614, "mean_token_accuracy": 0.8618647456169128, "num_tokens": 89018983.0, "step": 2342 }, { "epoch": 0.2980536827375652, "ewc_loss": 0.059326171875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.760376453399658, "learning_rate": 9.927935565917761e-06, "loss": 0.5303, "mean_token_accuracy": 0.8443136215209961, "num_tokens": 89055834.0, "step": 2343 }, { "epoch": 0.2981808930161557, "ewc_loss": 0.05126953125, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.058979034423828, "learning_rate": 9.93217465027554e-06, "loss": 0.511, "mean_token_accuracy": 0.8523725867271423, "num_tokens": 89098612.0, "step": 2344 }, { "epoch": 0.2983081032947462, "ewc_loss": 0.058837890625, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.7485055923461914, "learning_rate": 9.93641373463332e-06, "loss": 0.5797, "mean_token_accuracy": 0.8359935283660889, "num_tokens": 89130169.0, "step": 2345 }, { "epoch": 0.29843531357333675, "ewc_loss": 0.05419921875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2250189781188965, "learning_rate": 9.940652818991098e-06, "loss": 0.4622, "mean_token_accuracy": 0.8668271899223328, "num_tokens": 89170346.0, "step": 2346 }, { "epoch": 0.2985625238519272, "ewc_loss": 0.055419921875, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.570894241333008, "learning_rate": 9.944891903348877e-06, "loss": 0.5297, "mean_token_accuracy": 0.8427762985229492, "num_tokens": 89211713.0, "step": 2347 }, { "epoch": 0.29868973413051775, "ewc_loss": 0.052978515625, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2444396018981934, "learning_rate": 9.949130987706656e-06, "loss": 0.4773, "mean_token_accuracy": 0.8602159023284912, "num_tokens": 89252842.0, "step": 2348 }, { "epoch": 0.2988169444091083, "ewc_loss": 0.05126953125, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3862898349761963, "learning_rate": 9.953370072064435e-06, "loss": 0.4578, "mean_token_accuracy": 0.8624844551086426, "num_tokens": 89291914.0, "step": 2349 }, { "epoch": 0.29894415468769875, "ewc_loss": 0.051025390625, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.210017442703247, "learning_rate": 9.957609156422214e-06, "loss": 0.5431, "mean_token_accuracy": 0.8396007418632507, "num_tokens": 89333912.0, "step": 2350 }, { "epoch": 0.2990713649662893, "ewc_loss": 0.048095703125, "ewc_loss_diag": 5.8650970458984375e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.2489099502563477, "learning_rate": 9.961848240779993e-06, "loss": 0.4829, "mean_token_accuracy": 0.8568410873413086, "num_tokens": 89363603.0, "step": 2351 }, { "epoch": 0.2991985752448798, "ewc_loss": 0.047607421875, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1587295532226562, "learning_rate": 9.966087325137771e-06, "loss": 0.4615, "mean_token_accuracy": 0.8629199266433716, "num_tokens": 89404392.0, "step": 2352 }, { "epoch": 0.2993257855234703, "ewc_loss": 0.0458984375, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.126676559448242, "learning_rate": 9.97032640949555e-06, "loss": 0.5261, "mean_token_accuracy": 0.8387860059738159, "num_tokens": 89444180.0, "step": 2353 }, { "epoch": 0.2994529958020608, "ewc_loss": 0.045654296875, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.1036808490753174, "learning_rate": 9.97456549385333e-06, "loss": 0.5324, "mean_token_accuracy": 0.8377557396888733, "num_tokens": 89483875.0, "step": 2354 }, { "epoch": 0.29958020608065133, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.0749547481536865, "learning_rate": 9.978804578211108e-06, "loss": 0.5216, "mean_token_accuracy": 0.8432931900024414, "num_tokens": 89519544.0, "step": 2355 }, { "epoch": 0.2997074163592418, "ewc_loss": 0.04443359375, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.0786426067352295, "learning_rate": 9.983043662568887e-06, "loss": 0.5107, "mean_token_accuracy": 0.8489276170730591, "num_tokens": 89552635.0, "step": 2356 }, { "epoch": 0.29983462663783234, "ewc_loss": 0.0439453125, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 3.0272433757781982, "learning_rate": 9.987282746926664e-06, "loss": 0.4731, "mean_token_accuracy": 0.859227180480957, "num_tokens": 89588514.0, "step": 2357 }, { "epoch": 0.29996183691642286, "ewc_loss": 0.043701171875, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 3.0504112243652344, "learning_rate": 9.991521831284443e-06, "loss": 0.4793, "mean_token_accuracy": 0.85601407289505, "num_tokens": 89629029.0, "step": 2358 }, { "epoch": 0.30008904719501334, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.0050508975982666, "learning_rate": 9.995760915642222e-06, "loss": 0.4235, "mean_token_accuracy": 0.8704277276992798, "num_tokens": 89659214.0, "step": 2359 }, { "epoch": 0.30021625747360386, "ewc_loss": 0.043701171875, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 3.03760027885437, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8620308637619019, "num_tokens": 89695342.0, "step": 2360 }, { "epoch": 0.3003434677521944, "ewc_loss": 0.043701171875, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.9934604167938232, "learning_rate": 1e-05, "loss": 0.4966, "mean_token_accuracy": 0.8511910438537598, "num_tokens": 89731209.0, "step": 2361 }, { "epoch": 0.30047067803078487, "ewc_loss": 0.043701171875, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 3.0379037857055664, "learning_rate": 1e-05, "loss": 0.5371, "mean_token_accuracy": 0.8362452983856201, "num_tokens": 89771516.0, "step": 2362 }, { "epoch": 0.3005978883093754, "ewc_loss": 0.04345703125, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.969170093536377, "learning_rate": 1e-05, "loss": 0.5302, "mean_token_accuracy": 0.840529203414917, "num_tokens": 89817929.0, "step": 2363 }, { "epoch": 0.3007250985879659, "ewc_loss": 0.044189453125, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.0791447162628174, "learning_rate": 1e-05, "loss": 0.605, "mean_token_accuracy": 0.8252806067466736, "num_tokens": 89850714.0, "step": 2364 }, { "epoch": 0.3008523088665564, "ewc_loss": 0.042236328125, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 2.938415765762329, "learning_rate": 1e-05, "loss": 0.5046, "mean_token_accuracy": 0.8496155142784119, "num_tokens": 89885686.0, "step": 2365 }, { "epoch": 0.3009795191451469, "ewc_loss": 0.045166015625, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0878679752349854, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8587747812271118, "num_tokens": 89926549.0, "step": 2366 }, { "epoch": 0.30110672942373745, "ewc_loss": 0.042724609375, "ewc_loss_diag": 5.888938903808594e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.8906688690185547, "learning_rate": 1e-05, "loss": 0.5289, "mean_token_accuracy": 0.838892936706543, "num_tokens": 89969031.0, "step": 2367 }, { "epoch": 0.3012339397023279, "ewc_loss": 0.0458984375, "ewc_loss_diag": 5.936622619628906e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.1730337142944336, "learning_rate": 1e-05, "loss": 0.5088, "mean_token_accuracy": 0.8454971313476562, "num_tokens": 90000348.0, "step": 2368 }, { "epoch": 0.30136114998091845, "ewc_loss": 0.041748046875, "ewc_loss_diag": 5.936622619628906e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.8326473236083984, "learning_rate": 1e-05, "loss": 0.4943, "mean_token_accuracy": 0.8508899211883545, "num_tokens": 90033945.0, "step": 2369 }, { "epoch": 0.301488360259509, "ewc_loss": 0.048095703125, "ewc_loss_diag": 5.936622619628906e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.2894859313964844, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8604340553283691, "num_tokens": 90069537.0, "step": 2370 }, { "epoch": 0.30161557053809945, "ewc_loss": 0.0419921875, "ewc_loss_diag": 5.984306335449219e-05, "ewc_loss_parallel": 3.600120544433594e-05, "grad_norm": 2.744924306869507, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8623242378234863, "num_tokens": 90107923.0, "step": 2371 }, { "epoch": 0.30174278081669, "ewc_loss": 0.052001953125, "ewc_loss_diag": 5.936622619628906e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4939486980438232, "learning_rate": 1e-05, "loss": 0.5365, "mean_token_accuracy": 0.8403735160827637, "num_tokens": 90152211.0, "step": 2372 }, { "epoch": 0.3018699910952805, "ewc_loss": 0.0419921875, "ewc_loss_diag": 5.936622619628906e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.675966739654541, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8478757739067078, "num_tokens": 90193687.0, "step": 2373 }, { "epoch": 0.301997201373871, "ewc_loss": 0.06103515625, "ewc_loss_diag": 5.984306335449219e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.8285233974456787, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.867927074432373, "num_tokens": 90232639.0, "step": 2374 }, { "epoch": 0.3021244116524615, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.008148193359375e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.84245228767395, "learning_rate": 1e-05, "loss": 0.5123, "mean_token_accuracy": 0.8450610637664795, "num_tokens": 90273344.0, "step": 2375 }, { "epoch": 0.30225162193105204, "ewc_loss": 0.0673828125, "ewc_loss_diag": 6.008148193359375e-05, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.129269599914551, "learning_rate": 1e-05, "loss": 0.528, "mean_token_accuracy": 0.8505553603172302, "num_tokens": 90309999.0, "step": 2376 }, { "epoch": 0.3023788322096425, "ewc_loss": 0.059814453125, "ewc_loss_diag": 6.008148193359375e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3106696605682373, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8547789454460144, "num_tokens": 90349049.0, "step": 2377 }, { "epoch": 0.30250604248823304, "ewc_loss": 0.064453125, "ewc_loss_diag": 6.008148193359375e-05, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 4.026847839355469, "learning_rate": 1e-05, "loss": 0.5396, "mean_token_accuracy": 0.845229983329773, "num_tokens": 90388795.0, "step": 2378 }, { "epoch": 0.30263325276682357, "ewc_loss": 0.06298828125, "ewc_loss_diag": 6.008148193359375e-05, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.567652940750122, "learning_rate": 1e-05, "loss": 0.5546, "mean_token_accuracy": 0.834720253944397, "num_tokens": 90423785.0, "step": 2379 }, { "epoch": 0.3027604630454141, "ewc_loss": 0.058349609375, "ewc_loss_diag": 6.008148193359375e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.7473578453063965, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.850249171257019, "num_tokens": 90461625.0, "step": 2380 }, { "epoch": 0.30288767332400457, "ewc_loss": 0.05908203125, "ewc_loss_diag": 6.008148193359375e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.458625555038452, "learning_rate": 1e-05, "loss": 0.488, "mean_token_accuracy": 0.8585376739501953, "num_tokens": 90500286.0, "step": 2381 }, { "epoch": 0.3030148836025951, "ewc_loss": 0.053955078125, "ewc_loss_diag": 6.008148193359375e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.565237283706665, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8546128273010254, "num_tokens": 90533463.0, "step": 2382 }, { "epoch": 0.3031420938811856, "ewc_loss": 0.05322265625, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3132219314575195, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8581942915916443, "num_tokens": 90570536.0, "step": 2383 }, { "epoch": 0.3032693041597761, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.4603073596954346, "learning_rate": 1e-05, "loss": 0.5477, "mean_token_accuracy": 0.8366376757621765, "num_tokens": 90603030.0, "step": 2384 }, { "epoch": 0.3033965144383666, "ewc_loss": 0.04833984375, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.186486005783081, "learning_rate": 1e-05, "loss": 0.5326, "mean_token_accuracy": 0.842543363571167, "num_tokens": 90638459.0, "step": 2385 }, { "epoch": 0.30352372471695716, "ewc_loss": 0.05078125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.4031472206115723, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8519283533096313, "num_tokens": 90671894.0, "step": 2386 }, { "epoch": 0.30365093499554763, "ewc_loss": 0.046142578125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0577914714813232, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8585324883460999, "num_tokens": 90708583.0, "step": 2387 }, { "epoch": 0.30377814527413816, "ewc_loss": 0.050537109375, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.4140138626098633, "learning_rate": 1e-05, "loss": 0.5275, "mean_token_accuracy": 0.842850923538208, "num_tokens": 90749914.0, "step": 2388 }, { "epoch": 0.3039053555527287, "ewc_loss": 0.044189453125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.9561634063720703, "learning_rate": 1e-05, "loss": 0.4945, "mean_token_accuracy": 0.849564790725708, "num_tokens": 90791955.0, "step": 2389 }, { "epoch": 0.30403256583131916, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.515843152999878, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.86895751953125, "num_tokens": 90827666.0, "step": 2390 }, { "epoch": 0.3041597761099097, "ewc_loss": 0.043701171875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.932584762573242, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8459147214889526, "num_tokens": 90860647.0, "step": 2391 }, { "epoch": 0.3042869863885002, "ewc_loss": 0.056396484375, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.6563189029693604, "learning_rate": 1e-05, "loss": 0.5455, "mean_token_accuracy": 0.8377803564071655, "num_tokens": 90899737.0, "step": 2392 }, { "epoch": 0.3044141966670907, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9878017902374268, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8582490086555481, "num_tokens": 90939934.0, "step": 2393 }, { "epoch": 0.3045414069456812, "ewc_loss": 0.059326171875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.7497806549072266, "learning_rate": 1e-05, "loss": 0.5265, "mean_token_accuracy": 0.8427139520645142, "num_tokens": 90971538.0, "step": 2394 }, { "epoch": 0.30466861722427174, "ewc_loss": 0.05078125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1619441509246826, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8544867038726807, "num_tokens": 91012119.0, "step": 2395 }, { "epoch": 0.3047958275028622, "ewc_loss": 0.057373046875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.6538190841674805, "learning_rate": 1e-05, "loss": 0.5162, "mean_token_accuracy": 0.8493876457214355, "num_tokens": 91051342.0, "step": 2396 }, { "epoch": 0.30492303778145274, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2769930362701416, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8562859296798706, "num_tokens": 91087650.0, "step": 2397 }, { "epoch": 0.30505024806004327, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.466869592666626, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.8531898856163025, "num_tokens": 91129786.0, "step": 2398 }, { "epoch": 0.30517745833863374, "ewc_loss": 0.051025390625, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2846264839172363, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8506924510002136, "num_tokens": 91165273.0, "step": 2399 }, { "epoch": 0.3053046686172243, "ewc_loss": 0.047607421875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.2496533393859863, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8653916120529175, "num_tokens": 91198034.0, "step": 2400 }, { "epoch": 0.3054318788958148, "ewc_loss": 0.048828125, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.240391492843628, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8541818261146545, "num_tokens": 91241415.0, "step": 2401 }, { "epoch": 0.3055590891744053, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.065462589263916, "learning_rate": 1e-05, "loss": 0.5042, "mean_token_accuracy": 0.8461121916770935, "num_tokens": 91284713.0, "step": 2402 }, { "epoch": 0.3056862994529958, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.2393176555633545, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8467801809310913, "num_tokens": 91322143.0, "step": 2403 }, { "epoch": 0.30581350973158633, "ewc_loss": 0.043212890625, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.952969551086426, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8459008932113647, "num_tokens": 91366825.0, "step": 2404 }, { "epoch": 0.3059407200101768, "ewc_loss": 0.047607421875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.281620740890503, "learning_rate": 1e-05, "loss": 0.5447, "mean_token_accuracy": 0.8438019752502441, "num_tokens": 91404469.0, "step": 2405 }, { "epoch": 0.30606793028876733, "ewc_loss": 0.042724609375, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.8597609996795654, "learning_rate": 1e-05, "loss": 0.5226, "mean_token_accuracy": 0.841626763343811, "num_tokens": 91444799.0, "step": 2406 }, { "epoch": 0.30619514056735786, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.3787596225738525, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.8483163118362427, "num_tokens": 91483843.0, "step": 2407 }, { "epoch": 0.30632235084594833, "ewc_loss": 0.041748046875, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 3.552436828613281e-05, "grad_norm": 2.8042125701904297, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8509040474891663, "num_tokens": 91523922.0, "step": 2408 }, { "epoch": 0.30644956112453886, "ewc_loss": 0.0537109375, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.5461502075195312, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8546718955039978, "num_tokens": 91561155.0, "step": 2409 }, { "epoch": 0.3065767714031294, "ewc_loss": 0.042724609375, "ewc_loss_diag": 6.031990051269531e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.832228183746338, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.852499783039093, "num_tokens": 91596072.0, "step": 2410 }, { "epoch": 0.30670398168171986, "ewc_loss": 0.058349609375, "ewc_loss_diag": 6.0558319091796875e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.71474027633667, "learning_rate": 1e-05, "loss": 0.509, "mean_token_accuracy": 0.849134087562561, "num_tokens": 91640369.0, "step": 2411 }, { "epoch": 0.3068311919603104, "ewc_loss": 0.047607421875, "ewc_loss_diag": 6.0558319091796875e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.9847025871276855, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8523397445678711, "num_tokens": 91674941.0, "step": 2412 }, { "epoch": 0.3069584022389009, "ewc_loss": 0.060546875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.77115797996521, "learning_rate": 1e-05, "loss": 0.5095, "mean_token_accuracy": 0.8524770140647888, "num_tokens": 91709023.0, "step": 2413 }, { "epoch": 0.3070856125174914, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1717662811279297, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8527268171310425, "num_tokens": 91750684.0, "step": 2414 }, { "epoch": 0.3072128227960819, "ewc_loss": 0.057861328125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6615071296691895, "learning_rate": 1e-05, "loss": 0.5211, "mean_token_accuracy": 0.8492349982261658, "num_tokens": 91789369.0, "step": 2415 }, { "epoch": 0.30734003307467245, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.258636236190796, "learning_rate": 1e-05, "loss": 0.546, "mean_token_accuracy": 0.8402203917503357, "num_tokens": 91828228.0, "step": 2416 }, { "epoch": 0.3074672433532629, "ewc_loss": 0.053466796875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4699318408966064, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8595521450042725, "num_tokens": 91866191.0, "step": 2417 }, { "epoch": 0.30759445363185345, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2261321544647217, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8524482250213623, "num_tokens": 91907945.0, "step": 2418 }, { "epoch": 0.307721663910444, "ewc_loss": 0.05029296875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.321727991104126, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8585504293441772, "num_tokens": 91938651.0, "step": 2419 }, { "epoch": 0.30784887418903445, "ewc_loss": 0.048583984375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1758837699890137, "learning_rate": 1e-05, "loss": 0.5227, "mean_token_accuracy": 0.8508124351501465, "num_tokens": 91977557.0, "step": 2420 }, { "epoch": 0.307976084467625, "ewc_loss": 0.0478515625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.2147674560546875, "learning_rate": 1e-05, "loss": 0.5197, "mean_token_accuracy": 0.842812180519104, "num_tokens": 92012316.0, "step": 2421 }, { "epoch": 0.3081032947462155, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.1100566387176514, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8495343923568726, "num_tokens": 92049262.0, "step": 2422 }, { "epoch": 0.308230505024806, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.1515681743621826, "learning_rate": 1e-05, "loss": 0.5259, "mean_token_accuracy": 0.8465269207954407, "num_tokens": 92086048.0, "step": 2423 }, { "epoch": 0.3083577153033965, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0514814853668213, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8597666025161743, "num_tokens": 92123540.0, "step": 2424 }, { "epoch": 0.30848492558198704, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.139615058898926, "learning_rate": 1e-05, "loss": 0.5565, "mean_token_accuracy": 0.8304232954978943, "num_tokens": 92158275.0, "step": 2425 }, { "epoch": 0.3086121358605775, "ewc_loss": 0.043701171875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.985089063644409, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8684391975402832, "num_tokens": 92196444.0, "step": 2426 }, { "epoch": 0.30873934613916804, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.1475908756256104, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8582183122634888, "num_tokens": 92233988.0, "step": 2427 }, { "epoch": 0.30886655641775856, "ewc_loss": 0.04296875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.930694341659546, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8606114983558655, "num_tokens": 92271102.0, "step": 2428 }, { "epoch": 0.3089937666963491, "ewc_loss": 0.04638671875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.2042317390441895, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8501474857330322, "num_tokens": 92306112.0, "step": 2429 }, { "epoch": 0.30912097697493957, "ewc_loss": 0.042724609375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.8682923316955566, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8578543663024902, "num_tokens": 92347796.0, "step": 2430 }, { "epoch": 0.3092481872535301, "ewc_loss": 0.049072265625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.31400990486145, "learning_rate": 1e-05, "loss": 0.5219, "mean_token_accuracy": 0.8452258110046387, "num_tokens": 92382018.0, "step": 2431 }, { "epoch": 0.3093753975321206, "ewc_loss": 0.042236328125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.80045747756958, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8479584455490112, "num_tokens": 92428352.0, "step": 2432 }, { "epoch": 0.3095026078107111, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.477118968963623, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.8480404615402222, "num_tokens": 92462826.0, "step": 2433 }, { "epoch": 0.3096298180893016, "ewc_loss": 0.04296875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.793172836303711, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8578906059265137, "num_tokens": 92498371.0, "step": 2434 }, { "epoch": 0.30975702836789215, "ewc_loss": 0.057373046875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.6786770820617676, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8511480093002319, "num_tokens": 92536921.0, "step": 2435 }, { "epoch": 0.3098842386464826, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9144721031188965, "learning_rate": 1e-05, "loss": 0.5298, "mean_token_accuracy": 0.8409482836723328, "num_tokens": 92570627.0, "step": 2436 }, { "epoch": 0.31001144892507315, "ewc_loss": 0.060791015625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.790287494659424, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8563511371612549, "num_tokens": 92608683.0, "step": 2437 }, { "epoch": 0.3101386592036637, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.0970263481140137, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8586480617523193, "num_tokens": 92644689.0, "step": 2438 }, { "epoch": 0.31026586948225415, "ewc_loss": 0.059814453125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.7495501041412354, "learning_rate": 1e-05, "loss": 0.5158, "mean_token_accuracy": 0.8510882258415222, "num_tokens": 92684471.0, "step": 2439 }, { "epoch": 0.3103930797608447, "ewc_loss": 0.0537109375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2585296630859375, "learning_rate": 1e-05, "loss": 0.5423, "mean_token_accuracy": 0.8430252075195312, "num_tokens": 92724991.0, "step": 2440 }, { "epoch": 0.3105202900394352, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.578676223754883, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.843928873538971, "num_tokens": 92759302.0, "step": 2441 }, { "epoch": 0.3106475003180257, "ewc_loss": 0.052978515625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.265761375427246, "learning_rate": 1e-05, "loss": 0.5237, "mean_token_accuracy": 0.8469052314758301, "num_tokens": 92794585.0, "step": 2442 }, { "epoch": 0.3107747105966162, "ewc_loss": 0.051513671875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3884170055389404, "learning_rate": 1e-05, "loss": 0.5266, "mean_token_accuracy": 0.8443297743797302, "num_tokens": 92832405.0, "step": 2443 }, { "epoch": 0.31090192087520674, "ewc_loss": 0.051513671875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.202199697494507, "learning_rate": 1e-05, "loss": 0.5261, "mean_token_accuracy": 0.8419001698493958, "num_tokens": 92877960.0, "step": 2444 }, { "epoch": 0.3110291311537972, "ewc_loss": 0.049072265625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.2423479557037354, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8629132509231567, "num_tokens": 92914421.0, "step": 2445 }, { "epoch": 0.31115634143238774, "ewc_loss": 0.0478515625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1518714427948, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8488176465034485, "num_tokens": 92954844.0, "step": 2446 }, { "epoch": 0.31128355171097827, "ewc_loss": 0.04638671875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.16611647605896, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8644918203353882, "num_tokens": 92987719.0, "step": 2447 }, { "epoch": 0.31141076198956874, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.089038133621216, "learning_rate": 1e-05, "loss": 0.5058, "mean_token_accuracy": 0.8494837880134583, "num_tokens": 93027335.0, "step": 2448 }, { "epoch": 0.31153797226815927, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.115891695022583, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.8519023656845093, "num_tokens": 93063325.0, "step": 2449 }, { "epoch": 0.3116651825467498, "ewc_loss": 0.044677734375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.0343639850616455, "learning_rate": 1e-05, "loss": 0.512, "mean_token_accuracy": 0.8414095044136047, "num_tokens": 93104057.0, "step": 2450 }, { "epoch": 0.31179239282534027, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.122823715209961, "learning_rate": 1e-05, "loss": 0.5172, "mean_token_accuracy": 0.8448760509490967, "num_tokens": 93138912.0, "step": 2451 }, { "epoch": 0.3119196031039308, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.9749979972839355, "learning_rate": 1e-05, "loss": 0.5498, "mean_token_accuracy": 0.8395427465438843, "num_tokens": 93178658.0, "step": 2452 }, { "epoch": 0.3120468133825213, "ewc_loss": 0.045654296875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.1381945610046387, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8523135781288147, "num_tokens": 93217723.0, "step": 2453 }, { "epoch": 0.3121740236611118, "ewc_loss": 0.043212890625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.9243063926696777, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.857947051525116, "num_tokens": 93253917.0, "step": 2454 }, { "epoch": 0.3123012339397023, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.196850299835205, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8584763407707214, "num_tokens": 93289735.0, "step": 2455 }, { "epoch": 0.31242844421829286, "ewc_loss": 0.042724609375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.857619524002075, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8508689403533936, "num_tokens": 93331549.0, "step": 2456 }, { "epoch": 0.31255565449688333, "ewc_loss": 0.048828125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.304696559906006, "learning_rate": 1e-05, "loss": 0.5433, "mean_token_accuracy": 0.8374853730201721, "num_tokens": 93372977.0, "step": 2457 }, { "epoch": 0.31268286477547386, "ewc_loss": 0.042724609375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.7995216846466064, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8633668422698975, "num_tokens": 93406842.0, "step": 2458 }, { "epoch": 0.3128100750540644, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.471376657485962, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8523976802825928, "num_tokens": 93444113.0, "step": 2459 }, { "epoch": 0.31293728533265486, "ewc_loss": 0.042724609375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.7508533000946045, "learning_rate": 1e-05, "loss": 0.5284, "mean_token_accuracy": 0.8402695655822754, "num_tokens": 93490618.0, "step": 2460 }, { "epoch": 0.3130644956112454, "ewc_loss": 0.05859375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.725177049636841, "learning_rate": 1e-05, "loss": 0.5257, "mean_token_accuracy": 0.8434481024742126, "num_tokens": 93524711.0, "step": 2461 }, { "epoch": 0.3131917058898359, "ewc_loss": 0.045654296875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.884723424911499, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8503416180610657, "num_tokens": 93559169.0, "step": 2462 }, { "epoch": 0.3133189161684264, "ewc_loss": 0.0625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8879127502441406, "learning_rate": 1e-05, "loss": 0.5141, "mean_token_accuracy": 0.8530315160751343, "num_tokens": 93598941.0, "step": 2463 }, { "epoch": 0.3134461264470169, "ewc_loss": 0.054443359375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.151454448699951, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8555032014846802, "num_tokens": 93634979.0, "step": 2464 }, { "epoch": 0.31357333672560744, "ewc_loss": 0.06201171875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.83599591255188, "learning_rate": 1e-05, "loss": 0.5668, "mean_token_accuracy": 0.8297783136367798, "num_tokens": 93676736.0, "step": 2465 }, { "epoch": 0.3137005470041979, "ewc_loss": 0.056884765625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3183066844940186, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8621954321861267, "num_tokens": 93720118.0, "step": 2466 }, { "epoch": 0.31382775728278844, "ewc_loss": 0.0576171875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6407101154327393, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8580346703529358, "num_tokens": 93760814.0, "step": 2467 }, { "epoch": 0.313954967561379, "ewc_loss": 0.0546875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.322348117828369, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8686498403549194, "num_tokens": 93795326.0, "step": 2468 }, { "epoch": 0.31408217783996945, "ewc_loss": 0.05224609375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4610962867736816, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8519238829612732, "num_tokens": 93832993.0, "step": 2469 }, { "epoch": 0.31420938811856, "ewc_loss": 0.05224609375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2498323917388916, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8649344444274902, "num_tokens": 93871282.0, "step": 2470 }, { "epoch": 0.3143365983971505, "ewc_loss": 0.05029296875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.3162319660186768, "learning_rate": 1e-05, "loss": 0.5173, "mean_token_accuracy": 0.8482725620269775, "num_tokens": 93912698.0, "step": 2471 }, { "epoch": 0.314463808675741, "ewc_loss": 0.04833984375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.15384840965271, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8545494675636292, "num_tokens": 93952499.0, "step": 2472 }, { "epoch": 0.3145910189543315, "ewc_loss": 0.04833984375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.228548049926758, "learning_rate": 1e-05, "loss": 0.5378, "mean_token_accuracy": 0.8386595249176025, "num_tokens": 93992277.0, "step": 2473 }, { "epoch": 0.31471822923292203, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.081921339035034, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8545347452163696, "num_tokens": 94032012.0, "step": 2474 }, { "epoch": 0.3148454395115125, "ewc_loss": 0.046630859375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.18108868598938, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8598339557647705, "num_tokens": 94072210.0, "step": 2475 }, { "epoch": 0.31497264979010303, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.0140540599823, "learning_rate": 1e-05, "loss": 0.5109, "mean_token_accuracy": 0.8481876254081726, "num_tokens": 94112269.0, "step": 2476 }, { "epoch": 0.31509986006869356, "ewc_loss": 0.046875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.196416139602661, "learning_rate": 1e-05, "loss": 0.5805, "mean_token_accuracy": 0.8266838788986206, "num_tokens": 94153596.0, "step": 2477 }, { "epoch": 0.31522707034728403, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.945786476135254, "learning_rate": 1e-05, "loss": 0.5215, "mean_token_accuracy": 0.8459001183509827, "num_tokens": 94195146.0, "step": 2478 }, { "epoch": 0.31535428062587456, "ewc_loss": 0.047607421875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.235599994659424, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8603783845901489, "num_tokens": 94233196.0, "step": 2479 }, { "epoch": 0.3154814909044651, "ewc_loss": 0.043212890625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.867272138595581, "learning_rate": 1e-05, "loss": 0.516, "mean_token_accuracy": 0.8480519652366638, "num_tokens": 94275314.0, "step": 2480 }, { "epoch": 0.3156087011830556, "ewc_loss": 0.050048828125, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.342308282852173, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8541805148124695, "num_tokens": 94316673.0, "step": 2481 }, { "epoch": 0.3157359114616461, "ewc_loss": 0.04248046875, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 2.811480760574341, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8559259176254272, "num_tokens": 94354118.0, "step": 2482 }, { "epoch": 0.3158631217402366, "ewc_loss": 0.052978515625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.5233328342437744, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8452332019805908, "num_tokens": 94386476.0, "step": 2483 }, { "epoch": 0.31599033201882715, "ewc_loss": 0.043212890625, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.8205103874206543, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8502413034439087, "num_tokens": 94420146.0, "step": 2484 }, { "epoch": 0.3161175422974176, "ewc_loss": 0.058349609375, "ewc_loss_diag": 6.103515625e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.7076194286346436, "learning_rate": 1e-05, "loss": 0.5193, "mean_token_accuracy": 0.8482844829559326, "num_tokens": 94455057.0, "step": 2485 }, { "epoch": 0.31624475257600815, "ewc_loss": 0.046142578125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9367353916168213, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8589076995849609, "num_tokens": 94500666.0, "step": 2486 }, { "epoch": 0.3163719628545987, "ewc_loss": 0.061279296875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.81209135055542, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8652236461639404, "num_tokens": 94537623.0, "step": 2487 }, { "epoch": 0.31649917313318915, "ewc_loss": 0.05322265625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1790549755096436, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8640932440757751, "num_tokens": 94568563.0, "step": 2488 }, { "epoch": 0.3166263834117797, "ewc_loss": 0.05908203125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.701820135116577, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8652392625808716, "num_tokens": 94600931.0, "step": 2489 }, { "epoch": 0.3167535936903702, "ewc_loss": 0.054931640625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.30041241645813, "learning_rate": 1e-05, "loss": 0.5655, "mean_token_accuracy": 0.8338921070098877, "num_tokens": 94634345.0, "step": 2490 }, { "epoch": 0.3168808039689607, "ewc_loss": 0.0537109375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.480539321899414, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8641089200973511, "num_tokens": 94673157.0, "step": 2491 }, { "epoch": 0.3170080142475512, "ewc_loss": 0.052490234375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.268671989440918, "learning_rate": 1e-05, "loss": 0.5402, "mean_token_accuracy": 0.8365609645843506, "num_tokens": 94712727.0, "step": 2492 }, { "epoch": 0.31713522452614173, "ewc_loss": 0.05029296875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.30961012840271, "learning_rate": 1e-05, "loss": 0.568, "mean_token_accuracy": 0.8328666090965271, "num_tokens": 94752341.0, "step": 2493 }, { "epoch": 0.3172624348047322, "ewc_loss": 0.05029296875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.21846604347229, "learning_rate": 1e-05, "loss": 0.5265, "mean_token_accuracy": 0.8437488675117493, "num_tokens": 94791287.0, "step": 2494 }, { "epoch": 0.31738964508332274, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.149190664291382, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8458935618400574, "num_tokens": 94833029.0, "step": 2495 }, { "epoch": 0.31751685536191326, "ewc_loss": 0.0478515625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1741528511047363, "learning_rate": 1e-05, "loss": 0.4996, "mean_token_accuracy": 0.8502912521362305, "num_tokens": 94873400.0, "step": 2496 }, { "epoch": 0.31764406564050374, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.0560762882232666, "learning_rate": 1e-05, "loss": 0.5426, "mean_token_accuracy": 0.8328689932823181, "num_tokens": 94914603.0, "step": 2497 }, { "epoch": 0.31777127591909426, "ewc_loss": 0.046875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1551411151885986, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8720685243606567, "num_tokens": 94955925.0, "step": 2498 }, { "epoch": 0.3178984861976848, "ewc_loss": 0.044677734375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.996168375015259, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8696638345718384, "num_tokens": 94989420.0, "step": 2499 }, { "epoch": 0.31802569647627527, "ewc_loss": 0.04638671875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.1618740558624268, "learning_rate": 1e-05, "loss": 0.5205, "mean_token_accuracy": 0.8451129198074341, "num_tokens": 95028835.0, "step": 2500 }, { "epoch": 0.3181529067548658, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.9469056129455566, "learning_rate": 1e-05, "loss": 0.5625, "mean_token_accuracy": 0.8381880521774292, "num_tokens": 95069246.0, "step": 2501 }, { "epoch": 0.3182801170334563, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.199155569076538, "learning_rate": 1e-05, "loss": 0.6068, "mean_token_accuracy": 0.8190374374389648, "num_tokens": 95110184.0, "step": 2502 }, { "epoch": 0.3184073273120468, "ewc_loss": 0.043212890625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.8808271884918213, "learning_rate": 1e-05, "loss": 0.49, "mean_token_accuracy": 0.855557918548584, "num_tokens": 95149118.0, "step": 2503 }, { "epoch": 0.3185345375906373, "ewc_loss": 0.04833984375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.263376235961914, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.8529351353645325, "num_tokens": 95195605.0, "step": 2504 }, { "epoch": 0.31866174786922785, "ewc_loss": 0.04248046875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.647804260253906e-05, "grad_norm": 2.8205747604370117, "learning_rate": 1e-05, "loss": 0.5092, "mean_token_accuracy": 0.8444878458976746, "num_tokens": 95234602.0, "step": 2505 }, { "epoch": 0.3187889581478183, "ewc_loss": 0.05126953125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3982958793640137, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8600172996520996, "num_tokens": 95275646.0, "step": 2506 }, { "epoch": 0.31891616842640885, "ewc_loss": 0.042724609375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.783682346343994, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8486985564231873, "num_tokens": 95309948.0, "step": 2507 }, { "epoch": 0.3190433787049994, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5893163681030273, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8510861396789551, "num_tokens": 95347506.0, "step": 2508 }, { "epoch": 0.31917058898358985, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.8025293350219727, "learning_rate": 1e-05, "loss": 0.4961, "mean_token_accuracy": 0.8505803942680359, "num_tokens": 95386809.0, "step": 2509 }, { "epoch": 0.3192977992621804, "ewc_loss": 0.060791015625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.804023504257202, "learning_rate": 1e-05, "loss": 0.57, "mean_token_accuracy": 0.8433089852333069, "num_tokens": 95429287.0, "step": 2510 }, { "epoch": 0.3194250095407709, "ewc_loss": 0.05078125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.9761762619018555, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8732532262802124, "num_tokens": 95466443.0, "step": 2511 }, { "epoch": 0.3195522198193614, "ewc_loss": 0.0634765625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.9044387340545654, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8576076626777649, "num_tokens": 95511593.0, "step": 2512 }, { "epoch": 0.3196794300979519, "ewc_loss": 0.056396484375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.251495122909546, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8527927398681641, "num_tokens": 95546756.0, "step": 2513 }, { "epoch": 0.31980664037654244, "ewc_loss": 0.060791015625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.767385244369507, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8515486121177673, "num_tokens": 95586434.0, "step": 2514 }, { "epoch": 0.3199338506551329, "ewc_loss": 0.056884765625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.350790500640869, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8699871897697449, "num_tokens": 95616980.0, "step": 2515 }, { "epoch": 0.32006106093372344, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5681028366088867, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8524633646011353, "num_tokens": 95656919.0, "step": 2516 }, { "epoch": 0.32018827121231397, "ewc_loss": 0.05419921875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3031091690063477, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8535760641098022, "num_tokens": 95691785.0, "step": 2517 }, { "epoch": 0.32031548149090444, "ewc_loss": 0.05126953125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3925235271453857, "learning_rate": 1e-05, "loss": 0.5039, "mean_token_accuracy": 0.8492379188537598, "num_tokens": 95731055.0, "step": 2518 }, { "epoch": 0.32044269176949497, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.20458722114563, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8713514804840088, "num_tokens": 95767609.0, "step": 2519 }, { "epoch": 0.3205699020480855, "ewc_loss": 0.049560546875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.283795118331909, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8616742491722107, "num_tokens": 95805156.0, "step": 2520 }, { "epoch": 0.32069711232667597, "ewc_loss": 0.047607421875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1217291355133057, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8618432879447937, "num_tokens": 95848809.0, "step": 2521 }, { "epoch": 0.3208243226052665, "ewc_loss": 0.0478515625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.229304313659668, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8465966582298279, "num_tokens": 95882961.0, "step": 2522 }, { "epoch": 0.320951532883857, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0595052242279053, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8642477989196777, "num_tokens": 95916360.0, "step": 2523 }, { "epoch": 0.3210787431624475, "ewc_loss": 0.046875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.2063496112823486, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.853155255317688, "num_tokens": 95952064.0, "step": 2524 }, { "epoch": 0.32120595344103803, "ewc_loss": 0.044189453125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.9691405296325684, "learning_rate": 1e-05, "loss": 0.49, "mean_token_accuracy": 0.8523813486099243, "num_tokens": 95993073.0, "step": 2525 }, { "epoch": 0.32133316371962856, "ewc_loss": 0.048095703125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.250070333480835, "learning_rate": 1e-05, "loss": 0.5236, "mean_token_accuracy": 0.8406164646148682, "num_tokens": 96028659.0, "step": 2526 }, { "epoch": 0.32146037399821903, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.9134371280670166, "learning_rate": 1e-05, "loss": 0.5371, "mean_token_accuracy": 0.8462097644805908, "num_tokens": 96064802.0, "step": 2527 }, { "epoch": 0.32158758427680956, "ewc_loss": 0.05029296875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.312342405319214, "learning_rate": 1e-05, "loss": 0.5253, "mean_token_accuracy": 0.8436222076416016, "num_tokens": 96112980.0, "step": 2528 }, { "epoch": 0.3217147945554001, "ewc_loss": 0.043212890625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.8647959232330322, "learning_rate": 1e-05, "loss": 0.5201, "mean_token_accuracy": 0.8415787816047668, "num_tokens": 96144114.0, "step": 2529 }, { "epoch": 0.3218420048339906, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.4390532970428467, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8565130829811096, "num_tokens": 96177383.0, "step": 2530 }, { "epoch": 0.3219692151125811, "ewc_loss": 0.043212890625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.824321746826172, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8564544916152954, "num_tokens": 96216513.0, "step": 2531 }, { "epoch": 0.3220964253911716, "ewc_loss": 0.05615234375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.611405849456787, "learning_rate": 1e-05, "loss": 0.5002, "mean_token_accuracy": 0.8582064509391785, "num_tokens": 96256588.0, "step": 2532 }, { "epoch": 0.32222363566976214, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9002468585968018, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8505282402038574, "num_tokens": 96295365.0, "step": 2533 }, { "epoch": 0.3223508459483526, "ewc_loss": 0.060302734375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.745325803756714, "learning_rate": 1e-05, "loss": 0.538, "mean_token_accuracy": 0.8473581075668335, "num_tokens": 96328048.0, "step": 2534 }, { "epoch": 0.32247805622694314, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.069502830505371, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.852759838104248, "num_tokens": 96362088.0, "step": 2535 }, { "epoch": 0.32260526650553367, "ewc_loss": 0.06005859375, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.7280797958374023, "learning_rate": 1e-05, "loss": 0.5339, "mean_token_accuracy": 0.848209023475647, "num_tokens": 96396161.0, "step": 2536 }, { "epoch": 0.32273247678412414, "ewc_loss": 0.05419921875, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2175137996673584, "learning_rate": 1e-05, "loss": 0.4966, "mean_token_accuracy": 0.8527277708053589, "num_tokens": 96433402.0, "step": 2537 }, { "epoch": 0.3228596870627147, "ewc_loss": 0.055908203125, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5550684928894043, "learning_rate": 1e-05, "loss": 0.513, "mean_token_accuracy": 0.8447109460830688, "num_tokens": 96473011.0, "step": 2538 }, { "epoch": 0.3229868973413052, "ewc_loss": 0.05322265625, "ewc_loss_diag": 6.151199340820312e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2497384548187256, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.859093427658081, "num_tokens": 96511595.0, "step": 2539 }, { "epoch": 0.3231141076198957, "ewc_loss": 0.051025390625, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.374206066131592, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.8476088047027588, "num_tokens": 96546192.0, "step": 2540 }, { "epoch": 0.3232413178984862, "ewc_loss": 0.05224609375, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2224018573760986, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8778730630874634, "num_tokens": 96582613.0, "step": 2541 }, { "epoch": 0.32336852817707673, "ewc_loss": 0.048583984375, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.206725835800171, "learning_rate": 1e-05, "loss": 0.5128, "mean_token_accuracy": 0.8471264839172363, "num_tokens": 96620219.0, "step": 2542 }, { "epoch": 0.3234957384556672, "ewc_loss": 0.04931640625, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.19368052482605, "learning_rate": 1e-05, "loss": 0.5173, "mean_token_accuracy": 0.8478332757949829, "num_tokens": 96661126.0, "step": 2543 }, { "epoch": 0.32362294873425773, "ewc_loss": 0.046142578125, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0883336067199707, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.844139814376831, "num_tokens": 96697995.0, "step": 2544 }, { "epoch": 0.32375015901284826, "ewc_loss": 0.046875, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.183567523956299, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8552982211112976, "num_tokens": 96732542.0, "step": 2545 }, { "epoch": 0.32387736929143873, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.0209431648254395, "learning_rate": 1e-05, "loss": 0.526, "mean_token_accuracy": 0.8411173820495605, "num_tokens": 96761787.0, "step": 2546 }, { "epoch": 0.32400457957002926, "ewc_loss": 0.046875, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1645278930664062, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8475905060768127, "num_tokens": 96804918.0, "step": 2547 }, { "epoch": 0.3241317898486198, "ewc_loss": 0.043701171875, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.9489519596099854, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8533118963241577, "num_tokens": 96843646.0, "step": 2548 }, { "epoch": 0.32425900012721026, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.2244873046875, "learning_rate": 1e-05, "loss": 0.5534, "mean_token_accuracy": 0.8326804637908936, "num_tokens": 96874796.0, "step": 2549 }, { "epoch": 0.3243862104058008, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.878236770629883, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8578386306762695, "num_tokens": 96912777.0, "step": 2550 }, { "epoch": 0.3245134206843913, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.198883056640625e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.3007919788360596, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8462040424346924, "num_tokens": 96952045.0, "step": 2551 }, { "epoch": 0.3246406309629818, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.29425048828125e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.8222532272338867, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8568093776702881, "num_tokens": 96984625.0, "step": 2552 }, { "epoch": 0.3247678412415723, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.29425048828125e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.4353933334350586, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.858893632888794, "num_tokens": 97022409.0, "step": 2553 }, { "epoch": 0.32489505152016285, "ewc_loss": 0.043701171875, "ewc_loss_diag": 6.437301635742188e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.7868785858154297, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8461403846740723, "num_tokens": 97062461.0, "step": 2554 }, { "epoch": 0.3250222617987533, "ewc_loss": 0.05712890625, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.617682456970215, "learning_rate": 1e-05, "loss": 0.5818, "mean_token_accuracy": 0.8281362652778625, "num_tokens": 97102010.0, "step": 2555 }, { "epoch": 0.32514947207734385, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.8459484577178955, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8553361296653748, "num_tokens": 97140304.0, "step": 2556 }, { "epoch": 0.3252766823559344, "ewc_loss": 0.0615234375, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.787217617034912, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8726909160614014, "num_tokens": 97180460.0, "step": 2557 }, { "epoch": 0.32540389263452485, "ewc_loss": 0.052978515625, "ewc_loss_diag": 6.437301635742188e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.031114339828491, "learning_rate": 1e-05, "loss": 0.5336, "mean_token_accuracy": 0.8436088562011719, "num_tokens": 97223439.0, "step": 2558 }, { "epoch": 0.3255311029131154, "ewc_loss": 0.062255859375, "ewc_loss_diag": 6.437301635742188e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8431050777435303, "learning_rate": 1e-05, "loss": 0.5643, "mean_token_accuracy": 0.8384060263633728, "num_tokens": 97257050.0, "step": 2559 }, { "epoch": 0.3256583131917059, "ewc_loss": 0.056884765625, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2488315105438232, "learning_rate": 1e-05, "loss": 0.5229, "mean_token_accuracy": 0.8458918929100037, "num_tokens": 97296790.0, "step": 2560 }, { "epoch": 0.3257855234702964, "ewc_loss": 0.059326171875, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6778481006622314, "learning_rate": 1e-05, "loss": 0.5043, "mean_token_accuracy": 0.8493073582649231, "num_tokens": 97336511.0, "step": 2561 }, { "epoch": 0.3259127337488869, "ewc_loss": 0.056396484375, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.299980640411377, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8508773446083069, "num_tokens": 97373589.0, "step": 2562 }, { "epoch": 0.32603994402747744, "ewc_loss": 0.053955078125, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.483768939971924, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8680185079574585, "num_tokens": 97413986.0, "step": 2563 }, { "epoch": 0.3261671543060679, "ewc_loss": 0.05322265625, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.249462366104126, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8470172882080078, "num_tokens": 97457878.0, "step": 2564 }, { "epoch": 0.32629436458465844, "ewc_loss": 0.051025390625, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.3215835094451904, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.856702446937561, "num_tokens": 97495463.0, "step": 2565 }, { "epoch": 0.32642157486324896, "ewc_loss": 0.05078125, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.197227716445923, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8453631401062012, "num_tokens": 97533748.0, "step": 2566 }, { "epoch": 0.32654878514183944, "ewc_loss": 0.048583984375, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.207252025604248, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8536778688430786, "num_tokens": 97572476.0, "step": 2567 }, { "epoch": 0.32667599542042997, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.389617919921875e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.144639730453491, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8579227924346924, "num_tokens": 97606658.0, "step": 2568 }, { "epoch": 0.3268032056990205, "ewc_loss": 0.046630859375, "ewc_loss_diag": 6.437301635742188e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.122138023376465, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.858313798904419, "num_tokens": 97645083.0, "step": 2569 }, { "epoch": 0.32693041597761097, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.437301635742188e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.103250503540039, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8500995635986328, "num_tokens": 97679440.0, "step": 2570 }, { "epoch": 0.3270576262562015, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.437301635742188e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0887198448181152, "learning_rate": 1e-05, "loss": 0.5653, "mean_token_accuracy": 0.8319118022918701, "num_tokens": 97719295.0, "step": 2571 }, { "epoch": 0.327184836534792, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.437301635742188e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0501606464385986, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8600525856018066, "num_tokens": 97759417.0, "step": 2572 }, { "epoch": 0.3273120468133825, "ewc_loss": 0.044677734375, "ewc_loss_diag": 6.437301635742188e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.0622096061706543, "learning_rate": 1e-05, "loss": 0.5256, "mean_token_accuracy": 0.8397624492645264, "num_tokens": 97804004.0, "step": 2573 }, { "epoch": 0.327439257091973, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.4849853515625e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 3.0316388607025146, "learning_rate": 1e-05, "loss": 0.554, "mean_token_accuracy": 0.836579442024231, "num_tokens": 97845496.0, "step": 2574 }, { "epoch": 0.32756646737056355, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.060227870941162, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8505021333694458, "num_tokens": 97887184.0, "step": 2575 }, { "epoch": 0.327693677649154, "ewc_loss": 0.044189453125, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 3.017732620239258, "learning_rate": 1e-05, "loss": 0.5259, "mean_token_accuracy": 0.8401947021484375, "num_tokens": 97924498.0, "step": 2576 }, { "epoch": 0.32782088792774455, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.058928966522217, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.857568621635437, "num_tokens": 97960225.0, "step": 2577 }, { "epoch": 0.3279480982063351, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.993375062942505, "learning_rate": 1e-05, "loss": 0.5298, "mean_token_accuracy": 0.835957407951355, "num_tokens": 98001746.0, "step": 2578 }, { "epoch": 0.3280753084849256, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0745131969451904, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8575787544250488, "num_tokens": 98038812.0, "step": 2579 }, { "epoch": 0.3282025187635161, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.9571056365966797, "learning_rate": 1e-05, "loss": 0.5117, "mean_token_accuracy": 0.8486289978027344, "num_tokens": 98073782.0, "step": 2580 }, { "epoch": 0.3283297290421066, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.1101856231689453, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8632555603981018, "num_tokens": 98109379.0, "step": 2581 }, { "epoch": 0.32845693932069714, "ewc_loss": 0.043701171875, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.916256904602051, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8633160591125488, "num_tokens": 98141661.0, "step": 2582 }, { "epoch": 0.3285841495992876, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.161867618560791, "learning_rate": 1e-05, "loss": 0.527, "mean_token_accuracy": 0.8398528099060059, "num_tokens": 98182634.0, "step": 2583 }, { "epoch": 0.32871135987787814, "ewc_loss": 0.043212890625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.846998691558838, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8545719981193542, "num_tokens": 98225356.0, "step": 2584 }, { "epoch": 0.32883857015646867, "ewc_loss": 0.04931640625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.288079023361206, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8531545400619507, "num_tokens": 98263575.0, "step": 2585 }, { "epoch": 0.32896578043505914, "ewc_loss": 0.042724609375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.62396240234375e-05, "grad_norm": 2.7548303604125977, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.848301112651825, "num_tokens": 98304993.0, "step": 2586 }, { "epoch": 0.32909299071364967, "ewc_loss": 0.053466796875, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.5153698921203613, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8561649322509766, "num_tokens": 98339963.0, "step": 2587 }, { "epoch": 0.3292202009922402, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.69916033744812, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8576170206069946, "num_tokens": 98375507.0, "step": 2588 }, { "epoch": 0.32934741127083067, "ewc_loss": 0.062255859375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.840346336364746, "learning_rate": 1e-05, "loss": 0.5036, "mean_token_accuracy": 0.8535518646240234, "num_tokens": 98413577.0, "step": 2589 }, { "epoch": 0.3294746215494212, "ewc_loss": 0.048583984375, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.8514890670776367, "learning_rate": 1e-05, "loss": 0.5277, "mean_token_accuracy": 0.8432273864746094, "num_tokens": 98451615.0, "step": 2590 }, { "epoch": 0.3296018318280117, "ewc_loss": 0.068359375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.147037506103516, "learning_rate": 1e-05, "loss": 0.6043, "mean_token_accuracy": 0.8280526995658875, "num_tokens": 98490728.0, "step": 2591 }, { "epoch": 0.3297290421066022, "ewc_loss": 0.06005859375, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.319547653198242, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8659170866012573, "num_tokens": 98526645.0, "step": 2592 }, { "epoch": 0.3298562523851927, "ewc_loss": 0.0654296875, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 4.0252299308776855, "learning_rate": 1e-05, "loss": 0.528, "mean_token_accuracy": 0.844258725643158, "num_tokens": 98568447.0, "step": 2593 }, { "epoch": 0.32998346266378326, "ewc_loss": 0.0634765625, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5496251583099365, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.851972222328186, "num_tokens": 98608952.0, "step": 2594 }, { "epoch": 0.33011067294237373, "ewc_loss": 0.060302734375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.7677934169769287, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8706026077270508, "num_tokens": 98649180.0, "step": 2595 }, { "epoch": 0.33023788322096426, "ewc_loss": 0.059326171875, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4723784923553467, "learning_rate": 1e-05, "loss": 0.5161, "mean_token_accuracy": 0.8505430817604065, "num_tokens": 98692506.0, "step": 2596 }, { "epoch": 0.3303650934995548, "ewc_loss": 0.055419921875, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5869643688201904, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8477107882499695, "num_tokens": 98727817.0, "step": 2597 }, { "epoch": 0.33049230377814526, "ewc_loss": 0.05419921875, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3274707794189453, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8568996787071228, "num_tokens": 98769714.0, "step": 2598 }, { "epoch": 0.3306195140567358, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4556334018707275, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8617386817932129, "num_tokens": 98811894.0, "step": 2599 }, { "epoch": 0.3307467243353263, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.212096691131592, "learning_rate": 1e-05, "loss": 0.5206, "mean_token_accuracy": 0.845069408416748, "num_tokens": 98847117.0, "step": 2600 }, { "epoch": 0.3308739346139168, "ewc_loss": 0.051025390625, "ewc_loss_diag": 6.532669067382812e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.3934178352355957, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8544341921806335, "num_tokens": 98882745.0, "step": 2601 }, { "epoch": 0.3310011448925073, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0841569900512695, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8627095222473145, "num_tokens": 98918039.0, "step": 2602 }, { "epoch": 0.33112835517109784, "ewc_loss": 0.05126953125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.390254497528076, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8532991409301758, "num_tokens": 98958571.0, "step": 2603 }, { "epoch": 0.3312555654496883, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9704999923706055, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8545531034469604, "num_tokens": 99004970.0, "step": 2604 }, { "epoch": 0.33138277572827884, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4920923709869385, "learning_rate": 1e-05, "loss": 0.5476, "mean_token_accuracy": 0.8415537476539612, "num_tokens": 99041112.0, "step": 2605 }, { "epoch": 0.3315099860068694, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.93232798576355, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8547878265380859, "num_tokens": 99076771.0, "step": 2606 }, { "epoch": 0.33163719628545985, "ewc_loss": 0.055908203125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.629133462905884, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.8496091961860657, "num_tokens": 99112183.0, "step": 2607 }, { "epoch": 0.3317644065640504, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9952893257141113, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.8535646796226501, "num_tokens": 99148789.0, "step": 2608 }, { "epoch": 0.3318916168426409, "ewc_loss": 0.05859375, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.699152708053589, "learning_rate": 1e-05, "loss": 0.5777, "mean_token_accuracy": 0.8289921283721924, "num_tokens": 99189697.0, "step": 2609 }, { "epoch": 0.3320188271212314, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1086511611938477, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8482798337936401, "num_tokens": 99228269.0, "step": 2610 }, { "epoch": 0.3321460373998219, "ewc_loss": 0.05810546875, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6734116077423096, "learning_rate": 1e-05, "loss": 0.5046, "mean_token_accuracy": 0.8512033224105835, "num_tokens": 99264275.0, "step": 2611 }, { "epoch": 0.33227324767841243, "ewc_loss": 0.051513671875, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2460014820098877, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8551709651947021, "num_tokens": 99294598.0, "step": 2612 }, { "epoch": 0.3324004579570029, "ewc_loss": 0.0537109375, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.488736391067505, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8572855591773987, "num_tokens": 99336307.0, "step": 2613 }, { "epoch": 0.33252766823559343, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.255063056945801, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.863823652267456, "num_tokens": 99373983.0, "step": 2614 }, { "epoch": 0.33265487851418396, "ewc_loss": 0.05029296875, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2958755493164062, "learning_rate": 1e-05, "loss": 0.5195, "mean_token_accuracy": 0.8461993932723999, "num_tokens": 99415794.0, "step": 2615 }, { "epoch": 0.33278208879277443, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.2436413764953613, "learning_rate": 1e-05, "loss": 0.5093, "mean_token_accuracy": 0.8458042144775391, "num_tokens": 99457572.0, "step": 2616 }, { "epoch": 0.33290929907136496, "ewc_loss": 0.046142578125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.1192171573638916, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8527752161026001, "num_tokens": 99496074.0, "step": 2617 }, { "epoch": 0.3330365093499555, "ewc_loss": 0.048828125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.235952854156494, "learning_rate": 1e-05, "loss": 0.5131, "mean_token_accuracy": 0.8435022830963135, "num_tokens": 99539655.0, "step": 2618 }, { "epoch": 0.33316371962854596, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 3.0058131217956543, "learning_rate": 1e-05, "loss": 0.5212, "mean_token_accuracy": 0.8450846672058105, "num_tokens": 99574696.0, "step": 2619 }, { "epoch": 0.3332909299071365, "ewc_loss": 0.048583984375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.2513678073883057, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8555458188056946, "num_tokens": 99612960.0, "step": 2620 }, { "epoch": 0.333418140185727, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.937645196914673, "learning_rate": 1e-05, "loss": 0.5674, "mean_token_accuracy": 0.8274009227752686, "num_tokens": 99644107.0, "step": 2621 }, { "epoch": 0.3335453504643175, "ewc_loss": 0.049560546875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.2876834869384766, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.8509823083877563, "num_tokens": 99682014.0, "step": 2622 }, { "epoch": 0.333672560742908, "ewc_loss": 0.043701171875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.856448173522949, "learning_rate": 1e-05, "loss": 0.5106, "mean_token_accuracy": 0.8452776670455933, "num_tokens": 99718969.0, "step": 2623 }, { "epoch": 0.33379977102149855, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.580352783203125e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.4051053524017334, "learning_rate": 1e-05, "loss": 0.5479, "mean_token_accuracy": 0.8375584483146667, "num_tokens": 99756955.0, "step": 2624 }, { "epoch": 0.333926981300089, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.8162295818328857, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8590080738067627, "num_tokens": 99792216.0, "step": 2625 }, { "epoch": 0.33405419157867955, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5789408683776855, "learning_rate": 1e-05, "loss": 0.5121, "mean_token_accuracy": 0.8429595232009888, "num_tokens": 99823973.0, "step": 2626 }, { "epoch": 0.3341814018572701, "ewc_loss": 0.044677734375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.8580358028411865, "learning_rate": 1e-05, "loss": 0.5087, "mean_token_accuracy": 0.8468616008758545, "num_tokens": 99857206.0, "step": 2627 }, { "epoch": 0.33430861213586055, "ewc_loss": 0.060546875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.7417078018188477, "learning_rate": 1e-05, "loss": 0.5738, "mean_token_accuracy": 0.8348997831344604, "num_tokens": 99891146.0, "step": 2628 }, { "epoch": 0.3344358224144511, "ewc_loss": 0.05078125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.996865749359131, "learning_rate": 1e-05, "loss": 0.5607, "mean_token_accuracy": 0.8335666656494141, "num_tokens": 99931456.0, "step": 2629 }, { "epoch": 0.3345630326930416, "ewc_loss": 0.062255859375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.797600030899048, "learning_rate": 1e-05, "loss": 0.5189, "mean_token_accuracy": 0.8452364802360535, "num_tokens": 99965968.0, "step": 2630 }, { "epoch": 0.33469024297163213, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.179643392562866, "learning_rate": 1e-05, "loss": 0.541, "mean_token_accuracy": 0.839989185333252, "num_tokens": 100010874.0, "step": 2631 }, { "epoch": 0.3348174532502226, "ewc_loss": 0.06005859375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6888253688812256, "learning_rate": 1e-05, "loss": 0.4943, "mean_token_accuracy": 0.8569915294647217, "num_tokens": 100052166.0, "step": 2632 }, { "epoch": 0.33494466352881314, "ewc_loss": 0.055908203125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.276672601699829, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.856600821018219, "num_tokens": 100093927.0, "step": 2633 }, { "epoch": 0.33507187380740366, "ewc_loss": 0.05517578125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.517104387283325, "learning_rate": 1e-05, "loss": 0.5218, "mean_token_accuracy": 0.8454810380935669, "num_tokens": 100128414.0, "step": 2634 }, { "epoch": 0.33519908408599414, "ewc_loss": 0.053466796875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2638018131256104, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8588598966598511, "num_tokens": 100160334.0, "step": 2635 }, { "epoch": 0.33532629436458466, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3386166095733643, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8516943454742432, "num_tokens": 100195188.0, "step": 2636 }, { "epoch": 0.3354535046431752, "ewc_loss": 0.051025390625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.20271372795105, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8556802868843079, "num_tokens": 100235338.0, "step": 2637 }, { "epoch": 0.33558071492176567, "ewc_loss": 0.04931640625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.2070889472961426, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8549737930297852, "num_tokens": 100276267.0, "step": 2638 }, { "epoch": 0.3357079252003562, "ewc_loss": 0.04833984375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.155820846557617, "learning_rate": 1e-05, "loss": 0.5525, "mean_token_accuracy": 0.8318221569061279, "num_tokens": 100316201.0, "step": 2639 }, { "epoch": 0.3358351354789467, "ewc_loss": 0.04736328125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.130173683166504, "learning_rate": 1e-05, "loss": 0.5496, "mean_token_accuracy": 0.8385639786720276, "num_tokens": 100353278.0, "step": 2640 }, { "epoch": 0.3359623457575372, "ewc_loss": 0.04638671875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.1296019554138184, "learning_rate": 1e-05, "loss": 0.5151, "mean_token_accuracy": 0.8470582365989685, "num_tokens": 100384185.0, "step": 2641 }, { "epoch": 0.3360895560361277, "ewc_loss": 0.045654296875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.06807804107666, "learning_rate": 1e-05, "loss": 0.5395, "mean_token_accuracy": 0.8345918655395508, "num_tokens": 100423232.0, "step": 2642 }, { "epoch": 0.33621676631471825, "ewc_loss": 0.046142578125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.101186513900757, "learning_rate": 1e-05, "loss": 0.5608, "mean_token_accuracy": 0.8282871246337891, "num_tokens": 100457912.0, "step": 2643 }, { "epoch": 0.3363439765933087, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0157229900360107, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8545846939086914, "num_tokens": 100499894.0, "step": 2644 }, { "epoch": 0.33647118687189925, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0868704319000244, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8653309345245361, "num_tokens": 100540961.0, "step": 2645 }, { "epoch": 0.3365983971504898, "ewc_loss": 0.044677734375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.9906485080718994, "learning_rate": 1e-05, "loss": 0.5022, "mean_token_accuracy": 0.8460718393325806, "num_tokens": 100581302.0, "step": 2646 }, { "epoch": 0.33672560742908025, "ewc_loss": 0.046142578125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.1091880798339844, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8558851480484009, "num_tokens": 100615289.0, "step": 2647 }, { "epoch": 0.3368528177076708, "ewc_loss": 0.044189453125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.935173511505127, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8614749908447266, "num_tokens": 100659911.0, "step": 2648 }, { "epoch": 0.3369800279862613, "ewc_loss": 0.046875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.1548402309417725, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8540142774581909, "num_tokens": 100700861.0, "step": 2649 }, { "epoch": 0.3371072382648518, "ewc_loss": 0.043701171875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.896955728530884, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8510137796401978, "num_tokens": 100737329.0, "step": 2650 }, { "epoch": 0.3372344485434423, "ewc_loss": 0.049072265625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.2385482788085938, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8677304983139038, "num_tokens": 100774278.0, "step": 2651 }, { "epoch": 0.33736165882203284, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.816443920135498, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8482364416122437, "num_tokens": 100819083.0, "step": 2652 }, { "epoch": 0.3374888691006233, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3931267261505127, "learning_rate": 1e-05, "loss": 0.5615, "mean_token_accuracy": 0.8371391296386719, "num_tokens": 100859711.0, "step": 2653 }, { "epoch": 0.33761607937921384, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.766557216644287, "learning_rate": 1e-05, "loss": 0.558, "mean_token_accuracy": 0.8278281688690186, "num_tokens": 100893915.0, "step": 2654 }, { "epoch": 0.33774328965780437, "ewc_loss": 0.056396484375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.6127443313598633, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.8590324521064758, "num_tokens": 100932609.0, "step": 2655 }, { "epoch": 0.33787049993639484, "ewc_loss": 0.044677734375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.7649941444396973, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8535503149032593, "num_tokens": 100966539.0, "step": 2656 }, { "epoch": 0.33799771021498537, "ewc_loss": 0.06396484375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.913693904876709, "learning_rate": 1e-05, "loss": 0.5652, "mean_token_accuracy": 0.83903968334198, "num_tokens": 100999979.0, "step": 2657 }, { "epoch": 0.3381249204935759, "ewc_loss": 0.052978515625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.007963180541992, "learning_rate": 1e-05, "loss": 0.5285, "mean_token_accuracy": 0.8427282571792603, "num_tokens": 101036546.0, "step": 2658 }, { "epoch": 0.33825213077216637, "ewc_loss": 0.06640625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 4.047512054443359, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8503251075744629, "num_tokens": 101067861.0, "step": 2659 }, { "epoch": 0.3383793410507569, "ewc_loss": 0.06005859375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3529608249664307, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8709525465965271, "num_tokens": 101104268.0, "step": 2660 }, { "epoch": 0.3385065513293474, "ewc_loss": 0.06298828125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.885143280029297, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.868880033493042, "num_tokens": 101143364.0, "step": 2661 }, { "epoch": 0.3386337616079379, "ewc_loss": 0.060546875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.4773170948028564, "learning_rate": 1e-05, "loss": 0.5462, "mean_token_accuracy": 0.840179443359375, "num_tokens": 101178359.0, "step": 2662 }, { "epoch": 0.33876097188652843, "ewc_loss": 0.057861328125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.662782669067383, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8633416891098022, "num_tokens": 101215993.0, "step": 2663 }, { "epoch": 0.33888818216511896, "ewc_loss": 0.05712890625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.380173921585083, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8638335466384888, "num_tokens": 101254924.0, "step": 2664 }, { "epoch": 0.33901539244370943, "ewc_loss": 0.053955078125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.488872766494751, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8613384962081909, "num_tokens": 101295632.0, "step": 2665 }, { "epoch": 0.33914260272229996, "ewc_loss": 0.052978515625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2710695266723633, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8582585453987122, "num_tokens": 101333295.0, "step": 2666 }, { "epoch": 0.3392698130008905, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3774144649505615, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8558049201965332, "num_tokens": 101372319.0, "step": 2667 }, { "epoch": 0.33939702327948096, "ewc_loss": 0.048828125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.172088384628296, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8530771732330322, "num_tokens": 101405300.0, "step": 2668 }, { "epoch": 0.3395242335580715, "ewc_loss": 0.05078125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.3358466625213623, "learning_rate": 1e-05, "loss": 0.5524, "mean_token_accuracy": 0.8387740254402161, "num_tokens": 101445576.0, "step": 2669 }, { "epoch": 0.339651443836662, "ewc_loss": 0.046630859375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0755293369293213, "learning_rate": 1e-05, "loss": 0.5642, "mean_token_accuracy": 0.8340667486190796, "num_tokens": 101481840.0, "step": 2670 }, { "epoch": 0.3397786541152525, "ewc_loss": 0.05078125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.3379712104797363, "learning_rate": 1e-05, "loss": 0.5335, "mean_token_accuracy": 0.8407402038574219, "num_tokens": 101511968.0, "step": 2671 }, { "epoch": 0.339905864393843, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9632582664489746, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8543661832809448, "num_tokens": 101549909.0, "step": 2672 }, { "epoch": 0.34003307467243354, "ewc_loss": 0.052001953125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.4212613105773926, "learning_rate": 1e-05, "loss": 0.5277, "mean_token_accuracy": 0.8439558744430542, "num_tokens": 101580892.0, "step": 2673 }, { "epoch": 0.340160284951024, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9007599353790283, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8567551970481873, "num_tokens": 101614589.0, "step": 2674 }, { "epoch": 0.34028749522961454, "ewc_loss": 0.054931640625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5337717533111572, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8558052778244019, "num_tokens": 101658988.0, "step": 2675 }, { "epoch": 0.3404147055082051, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9053525924682617, "learning_rate": 1e-05, "loss": 0.5625, "mean_token_accuracy": 0.8358380794525146, "num_tokens": 101697838.0, "step": 2676 }, { "epoch": 0.34054191578679555, "ewc_loss": 0.058837890625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6818108558654785, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8609106540679932, "num_tokens": 101733831.0, "step": 2677 }, { "epoch": 0.3406691260653861, "ewc_loss": 0.048583984375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.010120391845703, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8535367846488953, "num_tokens": 101772697.0, "step": 2678 }, { "epoch": 0.3407963363439766, "ewc_loss": 0.060302734375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.745965003967285, "learning_rate": 1e-05, "loss": 0.5771, "mean_token_accuracy": 0.8335946798324585, "num_tokens": 101810107.0, "step": 2679 }, { "epoch": 0.34092354662256713, "ewc_loss": 0.052978515625, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1923224925994873, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8553561568260193, "num_tokens": 101847239.0, "step": 2680 }, { "epoch": 0.3410507569011576, "ewc_loss": 0.0576171875, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.621795415878296, "learning_rate": 1e-05, "loss": 0.5365, "mean_token_accuracy": 0.8413124680519104, "num_tokens": 101884508.0, "step": 2681 }, { "epoch": 0.34117796717974813, "ewc_loss": 0.0537109375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.264045238494873, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8758693933486938, "num_tokens": 101925351.0, "step": 2682 }, { "epoch": 0.34130517745833866, "ewc_loss": 0.052490234375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.407057523727417, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.85498046875, "num_tokens": 101966086.0, "step": 2683 }, { "epoch": 0.34143238773692913, "ewc_loss": 0.052490234375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2684714794158936, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8573493957519531, "num_tokens": 102004935.0, "step": 2684 }, { "epoch": 0.34155959801551966, "ewc_loss": 0.048828125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.2208094596862793, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.8500707745552063, "num_tokens": 102039026.0, "step": 2685 }, { "epoch": 0.3416868082941102, "ewc_loss": 0.050537109375, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2456135749816895, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8495558500289917, "num_tokens": 102078990.0, "step": 2686 }, { "epoch": 0.34181401857270066, "ewc_loss": 0.046142578125, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.076611042022705, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8561952114105225, "num_tokens": 102114631.0, "step": 2687 }, { "epoch": 0.3419412288512912, "ewc_loss": 0.048828125, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.228314161300659, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8530885577201843, "num_tokens": 102161583.0, "step": 2688 }, { "epoch": 0.3420684391298817, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.968920946121216, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8579641580581665, "num_tokens": 102202183.0, "step": 2689 }, { "epoch": 0.3421956494084722, "ewc_loss": 0.049560546875, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.260496139526367, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.85924232006073, "num_tokens": 102235288.0, "step": 2690 }, { "epoch": 0.3423228596870627, "ewc_loss": 0.044189453125, "ewc_loss_diag": 6.628036499023438e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.9170966148376465, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8469120264053345, "num_tokens": 102272814.0, "step": 2691 }, { "epoch": 0.34245006996565325, "ewc_loss": 0.051025390625, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.325392961502075, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8504283428192139, "num_tokens": 102312763.0, "step": 2692 }, { "epoch": 0.3425772802442437, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.8682918548583984, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8552674055099487, "num_tokens": 102351490.0, "step": 2693 }, { "epoch": 0.34270449052283425, "ewc_loss": 0.052490234375, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.4290826320648193, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8497545719146729, "num_tokens": 102394495.0, "step": 2694 }, { "epoch": 0.3428317008014248, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.836247682571411, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8454767465591431, "num_tokens": 102431689.0, "step": 2695 }, { "epoch": 0.34295891108001525, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5760250091552734, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8605228662490845, "num_tokens": 102468653.0, "step": 2696 }, { "epoch": 0.3430861213586058, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.67572021484375e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.861023187637329, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8661942481994629, "num_tokens": 102504442.0, "step": 2697 }, { "epoch": 0.3432133316371963, "ewc_loss": 0.060546875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.7207860946655273, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.859306812286377, "num_tokens": 102539401.0, "step": 2698 }, { "epoch": 0.3433405419157868, "ewc_loss": 0.05078125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0024542808532715, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8584935665130615, "num_tokens": 102579874.0, "step": 2699 }, { "epoch": 0.3434677521943773, "ewc_loss": 0.0615234375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.773571729660034, "learning_rate": 1e-05, "loss": 0.5309, "mean_token_accuracy": 0.847737193107605, "num_tokens": 102617906.0, "step": 2700 }, { "epoch": 0.34359496247296784, "ewc_loss": 0.05517578125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.174670934677124, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8567997217178345, "num_tokens": 102657744.0, "step": 2701 }, { "epoch": 0.3437221727515583, "ewc_loss": 0.0595703125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.668046712875366, "learning_rate": 1e-05, "loss": 0.5227, "mean_token_accuracy": 0.8474221229553223, "num_tokens": 102696650.0, "step": 2702 }, { "epoch": 0.34384938303014884, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2660434246063232, "learning_rate": 1e-05, "loss": 0.4982, "mean_token_accuracy": 0.8505290150642395, "num_tokens": 102729114.0, "step": 2703 }, { "epoch": 0.34397659330873936, "ewc_loss": 0.05419921875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4961624145507812, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8523116707801819, "num_tokens": 102763562.0, "step": 2704 }, { "epoch": 0.34410380358732984, "ewc_loss": 0.05322265625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2413759231567383, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.8436942100524902, "num_tokens": 102803680.0, "step": 2705 }, { "epoch": 0.34423101386592037, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.316986083984375, "learning_rate": 1e-05, "loss": 0.5433, "mean_token_accuracy": 0.8408068418502808, "num_tokens": 102847932.0, "step": 2706 }, { "epoch": 0.3443582241445109, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1945343017578125, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8578554391860962, "num_tokens": 102888371.0, "step": 2707 }, { "epoch": 0.34448543442310137, "ewc_loss": 0.04931640625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1989247798919678, "learning_rate": 1e-05, "loss": 0.5466, "mean_token_accuracy": 0.8345177173614502, "num_tokens": 102928943.0, "step": 2708 }, { "epoch": 0.3446126447016919, "ewc_loss": 0.04833984375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1592607498168945, "learning_rate": 1e-05, "loss": 0.5011, "mean_token_accuracy": 0.8508604764938354, "num_tokens": 102969773.0, "step": 2709 }, { "epoch": 0.3447398549802824, "ewc_loss": 0.046875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.1122453212738037, "learning_rate": 1e-05, "loss": 0.5329, "mean_token_accuracy": 0.8399969339370728, "num_tokens": 103012496.0, "step": 2710 }, { "epoch": 0.3448670652588729, "ewc_loss": 0.04638671875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.1353049278259277, "learning_rate": 1e-05, "loss": 0.5151, "mean_token_accuracy": 0.8476452827453613, "num_tokens": 103047259.0, "step": 2711 }, { "epoch": 0.3449942755374634, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.049811601638794, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8610461950302124, "num_tokens": 103087534.0, "step": 2712 }, { "epoch": 0.34512148581605395, "ewc_loss": 0.046875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0990755558013916, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8611499071121216, "num_tokens": 103128782.0, "step": 2713 }, { "epoch": 0.3452486960946444, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.027846336364746, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8553003072738647, "num_tokens": 103161252.0, "step": 2714 }, { "epoch": 0.34537590637323495, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.0855462551116943, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8592662811279297, "num_tokens": 103201524.0, "step": 2715 }, { "epoch": 0.3455031166518255, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.006328582763672, "learning_rate": 1e-05, "loss": 0.509, "mean_token_accuracy": 0.8432478308677673, "num_tokens": 103235444.0, "step": 2716 }, { "epoch": 0.34563032693041595, "ewc_loss": 0.04638671875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.0921390056610107, "learning_rate": 1e-05, "loss": 0.5321, "mean_token_accuracy": 0.8381925225257874, "num_tokens": 103270927.0, "step": 2717 }, { "epoch": 0.3457575372090065, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.9706547260284424, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8616149425506592, "num_tokens": 103310518.0, "step": 2718 }, { "epoch": 0.345884747487597, "ewc_loss": 0.04638671875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.1092967987060547, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8718152046203613, "num_tokens": 103347940.0, "step": 2719 }, { "epoch": 0.3460119577661875, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.9310054779052734, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8634116649627686, "num_tokens": 103383995.0, "step": 2720 }, { "epoch": 0.346139168044778, "ewc_loss": 0.04736328125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.1738333702087402, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8706774711608887, "num_tokens": 103416411.0, "step": 2721 }, { "epoch": 0.34626637832336854, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.8831422328948975, "learning_rate": 1e-05, "loss": 0.5195, "mean_token_accuracy": 0.8450778126716614, "num_tokens": 103454199.0, "step": 2722 }, { "epoch": 0.346393588601959, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.2777764797210693, "learning_rate": 1e-05, "loss": 0.5259, "mean_token_accuracy": 0.8426609039306641, "num_tokens": 103490996.0, "step": 2723 }, { "epoch": 0.34652079888054954, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.6716461181640625e-05, "grad_norm": 2.795558214187622, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8581988215446472, "num_tokens": 103525987.0, "step": 2724 }, { "epoch": 0.34664800915914007, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.4528582096099854, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.8409807682037354, "num_tokens": 103565702.0, "step": 2725 }, { "epoch": 0.34677521943773054, "ewc_loss": 0.04345703125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.695487976074219e-05, "grad_norm": 2.731093406677246, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8532800674438477, "num_tokens": 103601764.0, "step": 2726 }, { "epoch": 0.34690242971632107, "ewc_loss": 0.06005859375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.7262144088745117, "learning_rate": 1e-05, "loss": 0.5141, "mean_token_accuracy": 0.8483966588973999, "num_tokens": 103640293.0, "step": 2727 }, { "epoch": 0.3470296399949116, "ewc_loss": 0.04638671875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.799740791320801, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8483190536499023, "num_tokens": 103680464.0, "step": 2728 }, { "epoch": 0.3471568502735021, "ewc_loss": 0.0673828125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.033100605010986, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8536372780799866, "num_tokens": 103717434.0, "step": 2729 }, { "epoch": 0.3472840605520926, "ewc_loss": 0.0576171875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.162482976913452, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8549507856369019, "num_tokens": 103755186.0, "step": 2730 }, { "epoch": 0.3474112708306831, "ewc_loss": 0.0673828125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.0594682693481445, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8663008809089661, "num_tokens": 103789516.0, "step": 2731 }, { "epoch": 0.34753848110927366, "ewc_loss": 0.0634765625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.480776071548462, "learning_rate": 1e-05, "loss": 0.5258, "mean_token_accuracy": 0.8456122279167175, "num_tokens": 103826103.0, "step": 2732 }, { "epoch": 0.34766569138786413, "ewc_loss": 0.06201171875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.821401596069336, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8528109192848206, "num_tokens": 103863059.0, "step": 2733 }, { "epoch": 0.34779290166645466, "ewc_loss": 0.060546875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.4836223125457764, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8578033447265625, "num_tokens": 103897976.0, "step": 2734 }, { "epoch": 0.3479201119450452, "ewc_loss": 0.05712890625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.6093764305114746, "learning_rate": 1e-05, "loss": 0.5548, "mean_token_accuracy": 0.837348222732544, "num_tokens": 103935485.0, "step": 2735 }, { "epoch": 0.34804732222363566, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3657102584838867, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8551719188690186, "num_tokens": 103971917.0, "step": 2736 }, { "epoch": 0.3481745325022262, "ewc_loss": 0.0546875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.47639799118042, "learning_rate": 1e-05, "loss": 0.5421, "mean_token_accuracy": 0.8454504609107971, "num_tokens": 104008797.0, "step": 2737 }, { "epoch": 0.3483017427808167, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.232541561126709, "learning_rate": 1e-05, "loss": 0.5177, "mean_token_accuracy": 0.8447424173355103, "num_tokens": 104049551.0, "step": 2738 }, { "epoch": 0.3484289530594072, "ewc_loss": 0.05224609375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.384384870529175, "learning_rate": 1e-05, "loss": 0.5467, "mean_token_accuracy": 0.8347330093383789, "num_tokens": 104089341.0, "step": 2739 }, { "epoch": 0.3485561633379977, "ewc_loss": 0.0478515625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.113284111022949, "learning_rate": 1e-05, "loss": 0.5113, "mean_token_accuracy": 0.8477091789245605, "num_tokens": 104133445.0, "step": 2740 }, { "epoch": 0.34868337361658824, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.3611624240875244, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8537086248397827, "num_tokens": 104170435.0, "step": 2741 }, { "epoch": 0.3488105838951787, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0099661350250244, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8624373078346252, "num_tokens": 104205660.0, "step": 2742 }, { "epoch": 0.34893779417376924, "ewc_loss": 0.05224609375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.4056715965270996, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.86246657371521, "num_tokens": 104243749.0, "step": 2743 }, { "epoch": 0.3490650044523598, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9364349842071533, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8522656559944153, "num_tokens": 104283411.0, "step": 2744 }, { "epoch": 0.34919221473095025, "ewc_loss": 0.0546875, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.5353050231933594, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8558510541915894, "num_tokens": 104324570.0, "step": 2745 }, { "epoch": 0.3493194250095408, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.9155097007751465, "learning_rate": 1e-05, "loss": 0.5156, "mean_token_accuracy": 0.8469974994659424, "num_tokens": 104363619.0, "step": 2746 }, { "epoch": 0.3494466352881313, "ewc_loss": 0.05908203125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6700761318206787, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8575125932693481, "num_tokens": 104405572.0, "step": 2747 }, { "epoch": 0.3495738455667218, "ewc_loss": 0.04736328125, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.9944865703582764, "learning_rate": 1e-05, "loss": 0.5026, "mean_token_accuracy": 0.8467658162117004, "num_tokens": 104446772.0, "step": 2748 }, { "epoch": 0.3497010558453123, "ewc_loss": 0.06103515625, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7424042224884033, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8573406934738159, "num_tokens": 104490912.0, "step": 2749 }, { "epoch": 0.34982826612390283, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1741836071014404, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8600906133651733, "num_tokens": 104525307.0, "step": 2750 }, { "epoch": 0.3499554764024933, "ewc_loss": 0.05859375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6445672512054443, "learning_rate": 1e-05, "loss": 0.516, "mean_token_accuracy": 0.8483103513717651, "num_tokens": 104562431.0, "step": 2751 }, { "epoch": 0.35008268668108383, "ewc_loss": 0.0537109375, "ewc_loss_diag": 6.723403930664062e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2769086360931396, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8698115944862366, "num_tokens": 104596440.0, "step": 2752 }, { "epoch": 0.35020989695967436, "ewc_loss": 0.05322265625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.432774066925049, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.864152193069458, "num_tokens": 104630804.0, "step": 2753 }, { "epoch": 0.35033710723826483, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.282662868499756, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8609721660614014, "num_tokens": 104667997.0, "step": 2754 }, { "epoch": 0.35046431751685536, "ewc_loss": 0.05029296875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2179718017578125, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8590098023414612, "num_tokens": 104711605.0, "step": 2755 }, { "epoch": 0.3505915277954459, "ewc_loss": 0.050048828125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.274808883666992, "learning_rate": 1e-05, "loss": 0.5108, "mean_token_accuracy": 0.8485870361328125, "num_tokens": 104748427.0, "step": 2756 }, { "epoch": 0.35071873807403636, "ewc_loss": 0.045654296875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.059013605117798, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8549370765686035, "num_tokens": 104784536.0, "step": 2757 }, { "epoch": 0.3508459483526269, "ewc_loss": 0.050048828125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.2913594245910645, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.857332170009613, "num_tokens": 104820096.0, "step": 2758 }, { "epoch": 0.3509731586312174, "ewc_loss": 0.044677734375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.941439390182495, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8587786555290222, "num_tokens": 104857016.0, "step": 2759 }, { "epoch": 0.3511003689098079, "ewc_loss": 0.051025390625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.335988759994507, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8536545038223267, "num_tokens": 104903999.0, "step": 2760 }, { "epoch": 0.3512275791883984, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.8735861778259277, "learning_rate": 1e-05, "loss": 0.5583, "mean_token_accuracy": 0.8341178894042969, "num_tokens": 104943954.0, "step": 2761 }, { "epoch": 0.35135478946698895, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.4508981704711914, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.8434505462646484, "num_tokens": 104980028.0, "step": 2762 }, { "epoch": 0.3514819997455794, "ewc_loss": 0.044189453125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.8413736820220947, "learning_rate": 1e-05, "loss": 0.4751, "mean_token_accuracy": 0.857113242149353, "num_tokens": 105020471.0, "step": 2763 }, { "epoch": 0.35160921002416995, "ewc_loss": 0.056884765625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.6188595294952393, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8541434407234192, "num_tokens": 105050419.0, "step": 2764 }, { "epoch": 0.3517364203027605, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.906615972518921, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.860990583896637, "num_tokens": 105079656.0, "step": 2765 }, { "epoch": 0.35186363058135095, "ewc_loss": 0.061279296875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.733865976333618, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8559969663619995, "num_tokens": 105119715.0, "step": 2766 }, { "epoch": 0.3519908408599415, "ewc_loss": 0.051025390625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0504939556121826, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8469403386116028, "num_tokens": 105158281.0, "step": 2767 }, { "epoch": 0.352118051138532, "ewc_loss": 0.061279296875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7365520000457764, "learning_rate": 1e-05, "loss": 0.5238, "mean_token_accuracy": 0.8462530970573425, "num_tokens": 105201372.0, "step": 2768 }, { "epoch": 0.3522452614171225, "ewc_loss": 0.05419921875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1908798217773438, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8472165465354919, "num_tokens": 105240557.0, "step": 2769 }, { "epoch": 0.352372471695713, "ewc_loss": 0.05859375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.634272336959839, "learning_rate": 1e-05, "loss": 0.5211, "mean_token_accuracy": 0.851375937461853, "num_tokens": 105276701.0, "step": 2770 }, { "epoch": 0.35249968197430354, "ewc_loss": 0.053955078125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2493138313293457, "learning_rate": 1e-05, "loss": 0.5561, "mean_token_accuracy": 0.8415521383285522, "num_tokens": 105319185.0, "step": 2771 }, { "epoch": 0.352626892252894, "ewc_loss": 0.054443359375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.452723741531372, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8704941272735596, "num_tokens": 105362800.0, "step": 2772 }, { "epoch": 0.35275410253148454, "ewc_loss": 0.05322265625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2234158515930176, "learning_rate": 1e-05, "loss": 0.5152, "mean_token_accuracy": 0.8487706184387207, "num_tokens": 105400731.0, "step": 2773 }, { "epoch": 0.35288131281007507, "ewc_loss": 0.05126953125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.3201067447662354, "learning_rate": 1e-05, "loss": 0.5358, "mean_token_accuracy": 0.8465008735656738, "num_tokens": 105439712.0, "step": 2774 }, { "epoch": 0.35300852308866554, "ewc_loss": 0.050537109375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.178640365600586, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8566049933433533, "num_tokens": 105475159.0, "step": 2775 }, { "epoch": 0.35313573336725607, "ewc_loss": 0.04931640625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.224356174468994, "learning_rate": 1e-05, "loss": 0.5208, "mean_token_accuracy": 0.8419482707977295, "num_tokens": 105511323.0, "step": 2776 }, { "epoch": 0.3532629436458466, "ewc_loss": 0.047607421875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1255362033843994, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8534559607505798, "num_tokens": 105544056.0, "step": 2777 }, { "epoch": 0.35339015392443707, "ewc_loss": 0.048095703125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.1530091762542725, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8674376010894775, "num_tokens": 105582335.0, "step": 2778 }, { "epoch": 0.3535173642030276, "ewc_loss": 0.046630859375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.085397720336914, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8592768311500549, "num_tokens": 105614046.0, "step": 2779 }, { "epoch": 0.3536445744816181, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.1397945880889893, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8607417941093445, "num_tokens": 105648294.0, "step": 2780 }, { "epoch": 0.35377178476020865, "ewc_loss": 0.045654296875, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.033174753189087, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8491762280464172, "num_tokens": 105685707.0, "step": 2781 }, { "epoch": 0.3538989950387991, "ewc_loss": 0.04736328125, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.149351119995117, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8538497686386108, "num_tokens": 105724662.0, "step": 2782 }, { "epoch": 0.35402620531738965, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 3.006732940673828, "learning_rate": 1e-05, "loss": 0.543, "mean_token_accuracy": 0.8407353758811951, "num_tokens": 105756781.0, "step": 2783 }, { "epoch": 0.3541534155959802, "ewc_loss": 0.047119140625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.1506683826446533, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8613616824150085, "num_tokens": 105798466.0, "step": 2784 }, { "epoch": 0.35428062587457065, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.955644369125366, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8564364910125732, "num_tokens": 105835085.0, "step": 2785 }, { "epoch": 0.3544078361531612, "ewc_loss": 0.04833984375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1894352436065674, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.862452507019043, "num_tokens": 105872268.0, "step": 2786 }, { "epoch": 0.3545350464317517, "ewc_loss": 0.044677734375, "ewc_loss_diag": 6.771087646484375e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.916510581970215, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8522078990936279, "num_tokens": 105905721.0, "step": 2787 }, { "epoch": 0.3546622567103422, "ewc_loss": 0.050048828125, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.2666382789611816, "learning_rate": 1e-05, "loss": 0.5445, "mean_token_accuracy": 0.8382574319839478, "num_tokens": 105945573.0, "step": 2788 }, { "epoch": 0.3547894669889327, "ewc_loss": 0.044189453125, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.8575124740600586, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8506816029548645, "num_tokens": 105978800.0, "step": 2789 }, { "epoch": 0.35491667726752324, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3895726203918457, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8524748682975769, "num_tokens": 106014035.0, "step": 2790 }, { "epoch": 0.3550438875461137, "ewc_loss": 0.0439453125, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 3.719329833984375e-05, "grad_norm": 2.8065505027770996, "learning_rate": 1e-05, "loss": 0.5864, "mean_token_accuracy": 0.827462911605835, "num_tokens": 106050721.0, "step": 2791 }, { "epoch": 0.35517109782470424, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5712859630584717, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.85672926902771, "num_tokens": 106086802.0, "step": 2792 }, { "epoch": 0.35529830810329477, "ewc_loss": 0.045166015625, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.805882692337036, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8635128736495972, "num_tokens": 106124174.0, "step": 2793 }, { "epoch": 0.35542551838188524, "ewc_loss": 0.06201171875, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.781456708908081, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.8571321964263916, "num_tokens": 106166147.0, "step": 2794 }, { "epoch": 0.35555272866047577, "ewc_loss": 0.05126953125, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 2.964202642440796, "learning_rate": 1e-05, "loss": 0.5215, "mean_token_accuracy": 0.8445211052894592, "num_tokens": 106206758.0, "step": 2795 }, { "epoch": 0.3556799389390663, "ewc_loss": 0.06494140625, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9085607528686523, "learning_rate": 1e-05, "loss": 0.5179, "mean_token_accuracy": 0.8501900434494019, "num_tokens": 106241823.0, "step": 2796 }, { "epoch": 0.35580714921765677, "ewc_loss": 0.057373046875, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2278268337249756, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8646717071533203, "num_tokens": 106275716.0, "step": 2797 }, { "epoch": 0.3559343594962473, "ewc_loss": 0.0625, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8125948905944824, "learning_rate": 1e-05, "loss": 0.5346, "mean_token_accuracy": 0.8437857031822205, "num_tokens": 106312595.0, "step": 2798 }, { "epoch": 0.3560615697748378, "ewc_loss": 0.058837890625, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.357646942138672, "learning_rate": 1e-05, "loss": 0.5335, "mean_token_accuracy": 0.8445980548858643, "num_tokens": 106346009.0, "step": 2799 }, { "epoch": 0.3561887800534283, "ewc_loss": 0.057861328125, "ewc_loss_diag": 6.818771362304688e-05, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.601571559906006, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8636149764060974, "num_tokens": 106384501.0, "step": 2800 }, { "epoch": 0.35631599033201883, "ewc_loss": 0.055908203125, "ewc_loss_diag": 6.866455078125e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3121745586395264, "learning_rate": 1e-05, "loss": 0.5032, "mean_token_accuracy": 0.8543177843093872, "num_tokens": 106420099.0, "step": 2801 }, { "epoch": 0.35644320061060936, "ewc_loss": 0.0537109375, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4201009273529053, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8666627407073975, "num_tokens": 106459697.0, "step": 2802 }, { "epoch": 0.35657041088919983, "ewc_loss": 0.052978515625, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2632884979248047, "learning_rate": 1e-05, "loss": 0.6004, "mean_token_accuracy": 0.8276598453521729, "num_tokens": 106499054.0, "step": 2803 }, { "epoch": 0.35669762116779036, "ewc_loss": 0.05126953125, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.273481607437134, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8561455011367798, "num_tokens": 106538704.0, "step": 2804 }, { "epoch": 0.3568248314463809, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.169233560562134, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8438506722450256, "num_tokens": 106580012.0, "step": 2805 }, { "epoch": 0.35695204172497136, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1991324424743652, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8543978333473206, "num_tokens": 106618160.0, "step": 2806 }, { "epoch": 0.3570792520035619, "ewc_loss": 0.047607421875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.112769842147827, "learning_rate": 1e-05, "loss": 0.5684, "mean_token_accuracy": 0.82846599817276, "num_tokens": 106654387.0, "step": 2807 }, { "epoch": 0.3572064622821524, "ewc_loss": 0.047607421875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1516342163085938, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.863456130027771, "num_tokens": 106693318.0, "step": 2808 }, { "epoch": 0.3573336725607429, "ewc_loss": 0.046142578125, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0641469955444336, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8456908464431763, "num_tokens": 106728779.0, "step": 2809 }, { "epoch": 0.3574608828393334, "ewc_loss": 0.04736328125, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.131664276123047, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.853638231754303, "num_tokens": 106771149.0, "step": 2810 }, { "epoch": 0.35758809311792394, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 3.01806640625, "learning_rate": 1e-05, "loss": 0.5314, "mean_token_accuracy": 0.8404490351676941, "num_tokens": 106809132.0, "step": 2811 }, { "epoch": 0.3577153033965144, "ewc_loss": 0.046875, "ewc_loss_diag": 6.914138793945312e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.239292621612549, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8699636459350586, "num_tokens": 106842747.0, "step": 2812 }, { "epoch": 0.35784251367510495, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.919912815093994, "learning_rate": 1e-05, "loss": 0.5307, "mean_token_accuracy": 0.8406334519386292, "num_tokens": 106883519.0, "step": 2813 }, { "epoch": 0.3579697239536955, "ewc_loss": 0.05126953125, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.294755220413208, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8407087326049805, "num_tokens": 106914818.0, "step": 2814 }, { "epoch": 0.35809693423228595, "ewc_loss": 0.04443359375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.8271608352661133, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8625661134719849, "num_tokens": 106954387.0, "step": 2815 }, { "epoch": 0.3582241445108765, "ewc_loss": 0.0537109375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.463197708129883, "learning_rate": 1e-05, "loss": 0.5259, "mean_token_accuracy": 0.8448143005371094, "num_tokens": 106994264.0, "step": 2816 }, { "epoch": 0.358351354789467, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.7801949977874756, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8590139150619507, "num_tokens": 107035575.0, "step": 2817 }, { "epoch": 0.3584785650680575, "ewc_loss": 0.06005859375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6969106197357178, "learning_rate": 1e-05, "loss": 0.5372, "mean_token_accuracy": 0.8428908586502075, "num_tokens": 107074644.0, "step": 2818 }, { "epoch": 0.358605775346648, "ewc_loss": 0.04833984375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.871154308319092, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8564236164093018, "num_tokens": 107108758.0, "step": 2819 }, { "epoch": 0.35873298562523853, "ewc_loss": 0.0654296875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8970134258270264, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8698229193687439, "num_tokens": 107147572.0, "step": 2820 }, { "epoch": 0.358860195903829, "ewc_loss": 0.05615234375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.15889310836792, "learning_rate": 1e-05, "loss": 0.5258, "mean_token_accuracy": 0.8457626104354858, "num_tokens": 107184928.0, "step": 2821 }, { "epoch": 0.35898740618241953, "ewc_loss": 0.064453125, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8571999073028564, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8651069402694702, "num_tokens": 107223462.0, "step": 2822 }, { "epoch": 0.35911461646101006, "ewc_loss": 0.05908203125, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3016276359558105, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8569169640541077, "num_tokens": 107263114.0, "step": 2823 }, { "epoch": 0.35924182673960053, "ewc_loss": 0.0615234375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.705488920211792, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8726950883865356, "num_tokens": 107301234.0, "step": 2824 }, { "epoch": 0.35936903701819106, "ewc_loss": 0.0576171875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3436408042907715, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.8570897579193115, "num_tokens": 107334903.0, "step": 2825 }, { "epoch": 0.3594962472967816, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5105092525482178, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8724753856658936, "num_tokens": 107374989.0, "step": 2826 }, { "epoch": 0.35962345757537206, "ewc_loss": 0.0546875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.269638776779175, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8617209792137146, "num_tokens": 107412545.0, "step": 2827 }, { "epoch": 0.3597506678539626, "ewc_loss": 0.052734375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.389496326446533, "learning_rate": 1e-05, "loss": 0.5542, "mean_token_accuracy": 0.8349837064743042, "num_tokens": 107447642.0, "step": 2828 }, { "epoch": 0.3598778781325531, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1937754154205322, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8557661771774292, "num_tokens": 107486514.0, "step": 2829 }, { "epoch": 0.36000508841114365, "ewc_loss": 0.0517578125, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2843401432037354, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.8464368581771851, "num_tokens": 107527351.0, "step": 2830 }, { "epoch": 0.3601322986897341, "ewc_loss": 0.0478515625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.1278927326202393, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8568609952926636, "num_tokens": 107558973.0, "step": 2831 }, { "epoch": 0.36025950896832465, "ewc_loss": 0.05029296875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.237469434738159, "learning_rate": 1e-05, "loss": 0.5065, "mean_token_accuracy": 0.8489636778831482, "num_tokens": 107597659.0, "step": 2832 }, { "epoch": 0.3603867192469152, "ewc_loss": 0.046875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0641696453094482, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8498064279556274, "num_tokens": 107635833.0, "step": 2833 }, { "epoch": 0.36051392952550565, "ewc_loss": 0.04931640625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.2108569145202637, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8620954751968384, "num_tokens": 107673347.0, "step": 2834 }, { "epoch": 0.3606411398040962, "ewc_loss": 0.0458984375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.98836350440979, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.851158618927002, "num_tokens": 107713820.0, "step": 2835 }, { "epoch": 0.3607683500826867, "ewc_loss": 0.0498046875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.252232313156128, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8587086200714111, "num_tokens": 107745586.0, "step": 2836 }, { "epoch": 0.3608955603612772, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9133810997009277, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8658228516578674, "num_tokens": 107787649.0, "step": 2837 }, { "epoch": 0.3610227706398677, "ewc_loss": 0.05224609375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3450675010681152, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8492898344993591, "num_tokens": 107831545.0, "step": 2838 }, { "epoch": 0.36114998091845824, "ewc_loss": 0.044921875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.84199857711792, "learning_rate": 1e-05, "loss": 0.5381, "mean_token_accuracy": 0.8389666080474854, "num_tokens": 107875252.0, "step": 2839 }, { "epoch": 0.3612771911970487, "ewc_loss": 0.0546875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.523226737976074, "learning_rate": 1e-05, "loss": 0.5678, "mean_token_accuracy": 0.8318960070610046, "num_tokens": 107912060.0, "step": 2840 }, { "epoch": 0.36140440147563924, "ewc_loss": 0.04541015625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.8362467288970947, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8485344648361206, "num_tokens": 107949778.0, "step": 2841 }, { "epoch": 0.36153161175422976, "ewc_loss": 0.060546875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.706951379776001, "learning_rate": 1e-05, "loss": 0.5052, "mean_token_accuracy": 0.8535169363021851, "num_tokens": 107991160.0, "step": 2842 }, { "epoch": 0.36165882203282024, "ewc_loss": 0.04931640625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.9566404819488525, "learning_rate": 1e-05, "loss": 0.5253, "mean_token_accuracy": 0.8401139974594116, "num_tokens": 108027059.0, "step": 2843 }, { "epoch": 0.36178603231141077, "ewc_loss": 0.06298828125, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.825732707977295, "learning_rate": 1e-05, "loss": 0.521, "mean_token_accuracy": 0.850176990032196, "num_tokens": 108066738.0, "step": 2844 }, { "epoch": 0.3619132425900013, "ewc_loss": 0.05517578125, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.173304557800293, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.841768741607666, "num_tokens": 108106498.0, "step": 2845 }, { "epoch": 0.36204045286859177, "ewc_loss": 0.0615234375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.730984926223755, "learning_rate": 1e-05, "loss": 0.5064, "mean_token_accuracy": 0.8513458371162415, "num_tokens": 108139146.0, "step": 2846 }, { "epoch": 0.3621676631471823, "ewc_loss": 0.05615234375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3086721897125244, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8675387501716614, "num_tokens": 108176070.0, "step": 2847 }, { "epoch": 0.3622948734257728, "ewc_loss": 0.0556640625, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.514315366744995, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8607190251350403, "num_tokens": 108214859.0, "step": 2848 }, { "epoch": 0.3624220837043633, "ewc_loss": 0.0546875, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.294264554977417, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8521698713302612, "num_tokens": 108250201.0, "step": 2849 }, { "epoch": 0.3625492939829538, "ewc_loss": 0.05224609375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3353612422943115, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.858361005783081, "num_tokens": 108282111.0, "step": 2850 }, { "epoch": 0.36267650426154435, "ewc_loss": 0.05224609375, "ewc_loss_diag": 6.961822509765625e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2496063709259033, "learning_rate": 1e-05, "loss": 0.5315, "mean_token_accuracy": 0.8379965424537659, "num_tokens": 108317873.0, "step": 2851 }, { "epoch": 0.3628037145401348, "ewc_loss": 0.049560546875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.170692205429077, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8663687705993652, "num_tokens": 108352874.0, "step": 2852 }, { "epoch": 0.36293092481872535, "ewc_loss": 0.05029296875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.223515033721924, "learning_rate": 1e-05, "loss": 0.5197, "mean_token_accuracy": 0.844147801399231, "num_tokens": 108387620.0, "step": 2853 }, { "epoch": 0.3630581350973159, "ewc_loss": 0.046875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0605242252349854, "learning_rate": 1e-05, "loss": 0.5331, "mean_token_accuracy": 0.838638186454773, "num_tokens": 108428961.0, "step": 2854 }, { "epoch": 0.36318534537590635, "ewc_loss": 0.049072265625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.210172414779663, "learning_rate": 1e-05, "loss": 0.5754, "mean_token_accuracy": 0.8307597637176514, "num_tokens": 108471505.0, "step": 2855 }, { "epoch": 0.3633125556544969, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.985523223876953, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.850794792175293, "num_tokens": 108515874.0, "step": 2856 }, { "epoch": 0.3634397659330874, "ewc_loss": 0.04931640625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.2311453819274902, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.854297935962677, "num_tokens": 108558302.0, "step": 2857 }, { "epoch": 0.3635669762116779, "ewc_loss": 0.045166015625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.926811456680298, "learning_rate": 1e-05, "loss": 0.5094, "mean_token_accuracy": 0.8446884155273438, "num_tokens": 108597668.0, "step": 2858 }, { "epoch": 0.3636941864902684, "ewc_loss": 0.051025390625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2931389808654785, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8514461517333984, "num_tokens": 108636623.0, "step": 2859 }, { "epoch": 0.36382139676885894, "ewc_loss": 0.044677734375, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.860865354537964, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8730053901672363, "num_tokens": 108673660.0, "step": 2860 }, { "epoch": 0.3639486070474494, "ewc_loss": 0.052978515625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.4081027507781982, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8646686673164368, "num_tokens": 108710350.0, "step": 2861 }, { "epoch": 0.36407581732603994, "ewc_loss": 0.044921875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.830226182937622, "learning_rate": 1e-05, "loss": 0.5044, "mean_token_accuracy": 0.8472352027893066, "num_tokens": 108747028.0, "step": 2862 }, { "epoch": 0.36420302760463047, "ewc_loss": 0.056640625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5894935131073, "learning_rate": 1e-05, "loss": 0.5413, "mean_token_accuracy": 0.8422338366508484, "num_tokens": 108780315.0, "step": 2863 }, { "epoch": 0.36433023788322094, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.8564491271972656, "learning_rate": 1e-05, "loss": 0.5028, "mean_token_accuracy": 0.8427266478538513, "num_tokens": 108811706.0, "step": 2864 }, { "epoch": 0.36445744816181147, "ewc_loss": 0.061767578125, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7508721351623535, "learning_rate": 1e-05, "loss": 0.5453, "mean_token_accuracy": 0.8441423773765564, "num_tokens": 108851545.0, "step": 2865 }, { "epoch": 0.364584658440402, "ewc_loss": 0.0517578125, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.9724228382110596, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.866144597530365, "num_tokens": 108887013.0, "step": 2866 }, { "epoch": 0.36471186871899247, "ewc_loss": 0.0634765625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8332061767578125, "learning_rate": 1e-05, "loss": 0.5375, "mean_token_accuracy": 0.8426709175109863, "num_tokens": 108934289.0, "step": 2867 }, { "epoch": 0.364839078997583, "ewc_loss": 0.056884765625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.180567979812622, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8613810539245605, "num_tokens": 108973823.0, "step": 2868 }, { "epoch": 0.36496628927617353, "ewc_loss": 0.0625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.765376329421997, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8608728647232056, "num_tokens": 109006947.0, "step": 2869 }, { "epoch": 0.365093499554764, "ewc_loss": 0.05810546875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.319044828414917, "learning_rate": 1e-05, "loss": 0.5374, "mean_token_accuracy": 0.840679943561554, "num_tokens": 109039839.0, "step": 2870 }, { "epoch": 0.36522070983335453, "ewc_loss": 0.057373046875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5729875564575195, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8626679182052612, "num_tokens": 109076348.0, "step": 2871 }, { "epoch": 0.36534792011194506, "ewc_loss": 0.0556640625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2847180366516113, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8665426969528198, "num_tokens": 109117050.0, "step": 2872 }, { "epoch": 0.36547513039053553, "ewc_loss": 0.05322265625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4049367904663086, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.8511064052581787, "num_tokens": 109151131.0, "step": 2873 }, { "epoch": 0.36560234066912606, "ewc_loss": 0.05322265625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.229139804840088, "learning_rate": 1e-05, "loss": 0.5131, "mean_token_accuracy": 0.8491911888122559, "num_tokens": 109194542.0, "step": 2874 }, { "epoch": 0.3657295509477166, "ewc_loss": 0.051025390625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2796876430511475, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8508106470108032, "num_tokens": 109229636.0, "step": 2875 }, { "epoch": 0.36585676122630706, "ewc_loss": 0.05029296875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.180176258087158, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.8459294438362122, "num_tokens": 109264705.0, "step": 2876 }, { "epoch": 0.3659839715048976, "ewc_loss": 0.04931640625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1842947006225586, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8609811663627625, "num_tokens": 109304220.0, "step": 2877 }, { "epoch": 0.3661111817834881, "ewc_loss": 0.0478515625, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1212873458862305, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8508324027061462, "num_tokens": 109347315.0, "step": 2878 }, { "epoch": 0.3662383920620786, "ewc_loss": 0.048095703125, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.13584303855896, "learning_rate": 1e-05, "loss": 0.5285, "mean_token_accuracy": 0.8414181470870972, "num_tokens": 109382679.0, "step": 2879 }, { "epoch": 0.3663656023406691, "ewc_loss": 0.046875, "ewc_loss_diag": 7.009506225585938e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0794811248779297, "learning_rate": 1e-05, "loss": 0.5628, "mean_token_accuracy": 0.8357074856758118, "num_tokens": 109428491.0, "step": 2880 }, { "epoch": 0.36649281261925964, "ewc_loss": 0.047119140625, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.1163525581359863, "learning_rate": 1e-05, "loss": 0.5169, "mean_token_accuracy": 0.8425562381744385, "num_tokens": 109469700.0, "step": 2881 }, { "epoch": 0.3666200228978502, "ewc_loss": 0.0458984375, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.022852897644043, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8606793880462646, "num_tokens": 109510245.0, "step": 2882 }, { "epoch": 0.36674723317644065, "ewc_loss": 0.04736328125, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.1399941444396973, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8431410193443298, "num_tokens": 109544932.0, "step": 2883 }, { "epoch": 0.3668744434550312, "ewc_loss": 0.0458984375, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9743356704711914, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.859520673751831, "num_tokens": 109579519.0, "step": 2884 }, { "epoch": 0.3670016537336217, "ewc_loss": 0.048095703125, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.1732325553894043, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8611887097358704, "num_tokens": 109620313.0, "step": 2885 }, { "epoch": 0.3671288640122122, "ewc_loss": 0.044677734375, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.9282124042510986, "learning_rate": 1e-05, "loss": 0.5349, "mean_token_accuracy": 0.8386439681053162, "num_tokens": 109663800.0, "step": 2886 }, { "epoch": 0.3672560742908027, "ewc_loss": 0.0498046875, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.2408969402313232, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8623198866844177, "num_tokens": 109701118.0, "step": 2887 }, { "epoch": 0.36738328456939323, "ewc_loss": 0.04443359375, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.86257004737854, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8695497512817383, "num_tokens": 109739486.0, "step": 2888 }, { "epoch": 0.3675104948479837, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3660855293273926, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8663530945777893, "num_tokens": 109781097.0, "step": 2889 }, { "epoch": 0.36763770512657423, "ewc_loss": 0.04443359375, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.8158817291259766, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8585211634635925, "num_tokens": 109813562.0, "step": 2890 }, { "epoch": 0.36776491540516476, "ewc_loss": 0.054931640625, "ewc_loss_diag": 7.05718994140625e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.5552823543548584, "learning_rate": 1e-05, "loss": 0.556, "mean_token_accuracy": 0.835207462310791, "num_tokens": 109855326.0, "step": 2891 }, { "epoch": 0.36789212568375523, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.104873657226562e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.7947776317596436, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.8503683805465698, "num_tokens": 109890787.0, "step": 2892 }, { "epoch": 0.36801933596234576, "ewc_loss": 0.062255859375, "ewc_loss_diag": 7.152557373046875e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7886886596679688, "learning_rate": 1e-05, "loss": 0.5206, "mean_token_accuracy": 0.8494020104408264, "num_tokens": 109932199.0, "step": 2893 }, { "epoch": 0.3681465462409363, "ewc_loss": 0.051025390625, "ewc_loss_diag": 7.152557373046875e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 2.933377981185913, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.85723876953125, "num_tokens": 109967388.0, "step": 2894 }, { "epoch": 0.36827375651952676, "ewc_loss": 0.06689453125, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.96238374710083, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8644837737083435, "num_tokens": 110003437.0, "step": 2895 }, { "epoch": 0.3684009667981173, "ewc_loss": 0.05810546875, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2241158485412598, "learning_rate": 1e-05, "loss": 0.5086, "mean_token_accuracy": 0.8486361503601074, "num_tokens": 110040822.0, "step": 2896 }, { "epoch": 0.3685281770767078, "ewc_loss": 0.06494140625, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8889784812927246, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8579440712928772, "num_tokens": 110079628.0, "step": 2897 }, { "epoch": 0.3686553873552983, "ewc_loss": 0.060546875, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.461512565612793, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8620714545249939, "num_tokens": 110115794.0, "step": 2898 }, { "epoch": 0.3687825976338888, "ewc_loss": 0.05908203125, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.647596597671509, "learning_rate": 1e-05, "loss": 0.5228, "mean_token_accuracy": 0.848189115524292, "num_tokens": 110153227.0, "step": 2899 }, { "epoch": 0.36890980791247935, "ewc_loss": 0.0576171875, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.345606565475464, "learning_rate": 1e-05, "loss": 0.5547, "mean_token_accuracy": 0.8352850675582886, "num_tokens": 110193186.0, "step": 2900 }, { "epoch": 0.3690370181910698, "ewc_loss": 0.0546875, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.466048240661621, "learning_rate": 1e-05, "loss": 0.5368, "mean_token_accuracy": 0.8439913392066956, "num_tokens": 110232130.0, "step": 2901 }, { "epoch": 0.36916422846966035, "ewc_loss": 0.05419921875, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2491016387939453, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8656051158905029, "num_tokens": 110271376.0, "step": 2902 }, { "epoch": 0.3692914387482509, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3443214893341064, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.8467825651168823, "num_tokens": 110314220.0, "step": 2903 }, { "epoch": 0.36941864902684135, "ewc_loss": 0.051513671875, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.163376569747925, "learning_rate": 1e-05, "loss": 0.5108, "mean_token_accuracy": 0.8476936221122742, "num_tokens": 110357292.0, "step": 2904 }, { "epoch": 0.3695458593054319, "ewc_loss": 0.05126953125, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.284463882446289, "learning_rate": 1e-05, "loss": 0.5519, "mean_token_accuracy": 0.8357783555984497, "num_tokens": 110397813.0, "step": 2905 }, { "epoch": 0.3696730695840224, "ewc_loss": 0.0478515625, "ewc_loss_diag": 7.200241088867188e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0719056129455566, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8639211654663086, "num_tokens": 110435773.0, "step": 2906 }, { "epoch": 0.3698002798626129, "ewc_loss": 0.05078125, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2880218029022217, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8759510517120361, "num_tokens": 110469518.0, "step": 2907 }, { "epoch": 0.3699274901412034, "ewc_loss": 0.046142578125, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0080997943878174, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8512654900550842, "num_tokens": 110503156.0, "step": 2908 }, { "epoch": 0.37005470041979394, "ewc_loss": 0.051513671875, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.3132517337799072, "learning_rate": 1e-05, "loss": 0.5205, "mean_token_accuracy": 0.8436874151229858, "num_tokens": 110543960.0, "step": 2909 }, { "epoch": 0.3701819106983844, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9261274337768555, "learning_rate": 1e-05, "loss": 0.5321, "mean_token_accuracy": 0.840313196182251, "num_tokens": 110584811.0, "step": 2910 }, { "epoch": 0.37030912097697494, "ewc_loss": 0.052978515625, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.411123752593994, "learning_rate": 1e-05, "loss": 0.5324, "mean_token_accuracy": 0.8432410955429077, "num_tokens": 110622433.0, "step": 2911 }, { "epoch": 0.37043633125556547, "ewc_loss": 0.04541015625, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.8728115558624268, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8484927415847778, "num_tokens": 110660994.0, "step": 2912 }, { "epoch": 0.37056354153415594, "ewc_loss": 0.05615234375, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5610873699188232, "learning_rate": 1e-05, "loss": 0.5816, "mean_token_accuracy": 0.8313490748405457, "num_tokens": 110703022.0, "step": 2913 }, { "epoch": 0.37069075181274647, "ewc_loss": 0.046142578125, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.871271848678589, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8637007474899292, "num_tokens": 110743864.0, "step": 2914 }, { "epoch": 0.370817962091337, "ewc_loss": 0.061279296875, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.7397284507751465, "learning_rate": 1e-05, "loss": 0.5115, "mean_token_accuracy": 0.8480115532875061, "num_tokens": 110778591.0, "step": 2915 }, { "epoch": 0.37094517236992747, "ewc_loss": 0.051513671875, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0156426429748535, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8608791828155518, "num_tokens": 110818402.0, "step": 2916 }, { "epoch": 0.371072382648518, "ewc_loss": 0.0625, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7941794395446777, "learning_rate": 1e-05, "loss": 0.5051, "mean_token_accuracy": 0.8541646003723145, "num_tokens": 110855129.0, "step": 2917 }, { "epoch": 0.3711995929271085, "ewc_loss": 0.055908203125, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2157673835754395, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8557175397872925, "num_tokens": 110896307.0, "step": 2918 }, { "epoch": 0.371326803205699, "ewc_loss": 0.059814453125, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.652235269546509, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8572176098823547, "num_tokens": 110940639.0, "step": 2919 }, { "epoch": 0.3714540134842895, "ewc_loss": 0.056396484375, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3081305027008057, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8536309003829956, "num_tokens": 110979134.0, "step": 2920 }, { "epoch": 0.37158122376288005, "ewc_loss": 0.054443359375, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.445136785507202, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8645186424255371, "num_tokens": 111013799.0, "step": 2921 }, { "epoch": 0.3717084340414705, "ewc_loss": 0.0546875, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2921628952026367, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8538353443145752, "num_tokens": 111055486.0, "step": 2922 }, { "epoch": 0.37183564432006105, "ewc_loss": 0.05078125, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2330081462860107, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8566909432411194, "num_tokens": 111101332.0, "step": 2923 }, { "epoch": 0.3719628545986516, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.283299446105957, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.8553661704063416, "num_tokens": 111137285.0, "step": 2924 }, { "epoch": 0.37209006487724205, "ewc_loss": 0.047607421875, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0820038318634033, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8461611270904541, "num_tokens": 111177254.0, "step": 2925 }, { "epoch": 0.3722172751558326, "ewc_loss": 0.051025390625, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.29339337348938, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8533422350883484, "num_tokens": 111217633.0, "step": 2926 }, { "epoch": 0.3723444854344231, "ewc_loss": 0.0458984375, "ewc_loss_diag": 7.2479248046875e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.986975908279419, "learning_rate": 1e-05, "loss": 0.5833, "mean_token_accuracy": 0.8231135010719299, "num_tokens": 111259273.0, "step": 2927 }, { "epoch": 0.3724716957130136, "ewc_loss": 0.05126953125, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.3258440494537354, "learning_rate": 1e-05, "loss": 0.5332, "mean_token_accuracy": 0.8383387327194214, "num_tokens": 111295068.0, "step": 2928 }, { "epoch": 0.3725989059916041, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9115512371063232, "learning_rate": 1e-05, "loss": 0.5248, "mean_token_accuracy": 0.8402096033096313, "num_tokens": 111336863.0, "step": 2929 }, { "epoch": 0.37272611627019464, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.411241054534912, "learning_rate": 1e-05, "loss": 0.545, "mean_token_accuracy": 0.8398727178573608, "num_tokens": 111378432.0, "step": 2930 }, { "epoch": 0.37285332654878517, "ewc_loss": 0.045166015625, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.8734359741210938, "learning_rate": 1e-05, "loss": 0.5086, "mean_token_accuracy": 0.8461138010025024, "num_tokens": 111412481.0, "step": 2931 }, { "epoch": 0.37298053682737564, "ewc_loss": 0.0556640625, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5373337268829346, "learning_rate": 1e-05, "loss": 0.5476, "mean_token_accuracy": 0.8364088535308838, "num_tokens": 111450534.0, "step": 2932 }, { "epoch": 0.37310774710596617, "ewc_loss": 0.0458984375, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.8786826133728027, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8609040975570679, "num_tokens": 111484278.0, "step": 2933 }, { "epoch": 0.3732349573845567, "ewc_loss": 0.060302734375, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6608517169952393, "learning_rate": 1e-05, "loss": 0.5099, "mean_token_accuracy": 0.848283052444458, "num_tokens": 111526797.0, "step": 2934 }, { "epoch": 0.37336216766314717, "ewc_loss": 0.05029296875, "ewc_loss_diag": 7.295608520507812e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.96461820602417, "learning_rate": 1e-05, "loss": 0.5311, "mean_token_accuracy": 0.8412533402442932, "num_tokens": 111568139.0, "step": 2935 }, { "epoch": 0.3734893779417377, "ewc_loss": 0.06201171875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7437729835510254, "learning_rate": 1e-05, "loss": 0.575, "mean_token_accuracy": 0.8292648792266846, "num_tokens": 111614519.0, "step": 2936 }, { "epoch": 0.3736165882203282, "ewc_loss": 0.053955078125, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1313788890838623, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8584595322608948, "num_tokens": 111647965.0, "step": 2937 }, { "epoch": 0.3737437984989187, "ewc_loss": 0.061279296875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6862471103668213, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.8526434898376465, "num_tokens": 111682492.0, "step": 2938 }, { "epoch": 0.37387100877750923, "ewc_loss": 0.0546875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2208995819091797, "learning_rate": 1e-05, "loss": 0.5334, "mean_token_accuracy": 0.8423312902450562, "num_tokens": 111717550.0, "step": 2939 }, { "epoch": 0.37399821905609976, "ewc_loss": 0.056640625, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.544729471206665, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8628547191619873, "num_tokens": 111752173.0, "step": 2940 }, { "epoch": 0.37412542933469023, "ewc_loss": 0.054443359375, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2220799922943115, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8523092269897461, "num_tokens": 111790845.0, "step": 2941 }, { "epoch": 0.37425263961328076, "ewc_loss": 0.053466796875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3968095779418945, "learning_rate": 1e-05, "loss": 0.5032, "mean_token_accuracy": 0.8483177423477173, "num_tokens": 111827837.0, "step": 2942 }, { "epoch": 0.3743798498918713, "ewc_loss": 0.052490234375, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.202178716659546, "learning_rate": 1e-05, "loss": 0.5243, "mean_token_accuracy": 0.842826247215271, "num_tokens": 111868785.0, "step": 2943 }, { "epoch": 0.37450706017046176, "ewc_loss": 0.05126953125, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.263728380203247, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.848060667514801, "num_tokens": 111907296.0, "step": 2944 }, { "epoch": 0.3746342704490523, "ewc_loss": 0.0498046875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.149263381958008, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8667596578598022, "num_tokens": 111947619.0, "step": 2945 }, { "epoch": 0.3747614807276428, "ewc_loss": 0.0498046875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1956114768981934, "learning_rate": 1e-05, "loss": 0.5504, "mean_token_accuracy": 0.8352890014648438, "num_tokens": 111987255.0, "step": 2946 }, { "epoch": 0.3748886910062333, "ewc_loss": 0.0478515625, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.1111247539520264, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8491859436035156, "num_tokens": 112023003.0, "step": 2947 }, { "epoch": 0.3750159012848238, "ewc_loss": 0.04833984375, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.150761127471924, "learning_rate": 1e-05, "loss": 0.5151, "mean_token_accuracy": 0.8433759212493896, "num_tokens": 112056292.0, "step": 2948 }, { "epoch": 0.37514311156341434, "ewc_loss": 0.046875, "ewc_loss_diag": 7.343292236328125e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.0776355266571045, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8496429920196533, "num_tokens": 112089947.0, "step": 2949 }, { "epoch": 0.3752703218420048, "ewc_loss": 0.04736328125, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.1398251056671143, "learning_rate": 1e-05, "loss": 0.5754, "mean_token_accuracy": 0.827910304069519, "num_tokens": 112125415.0, "step": 2950 }, { "epoch": 0.37539753212059535, "ewc_loss": 0.046142578125, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0244786739349365, "learning_rate": 1e-05, "loss": 0.5129, "mean_token_accuracy": 0.8460027575492859, "num_tokens": 112164364.0, "step": 2951 }, { "epoch": 0.3755247423991859, "ewc_loss": 0.04736328125, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.124835729598999, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8606784343719482, "num_tokens": 112208994.0, "step": 2952 }, { "epoch": 0.37565195267777635, "ewc_loss": 0.046142578125, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.988525629043579, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8540029525756836, "num_tokens": 112249902.0, "step": 2953 }, { "epoch": 0.3757791629563669, "ewc_loss": 0.048095703125, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.163144111633301, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8610185980796814, "num_tokens": 112290446.0, "step": 2954 }, { "epoch": 0.3759063732349574, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.962578296661377, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8501427173614502, "num_tokens": 112323217.0, "step": 2955 }, { "epoch": 0.3760335835135479, "ewc_loss": 0.04931640625, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.2071545124053955, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.854874849319458, "num_tokens": 112365863.0, "step": 2956 }, { "epoch": 0.3761607937921384, "ewc_loss": 0.045166015625, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.9034218788146973, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8520383834838867, "num_tokens": 112400708.0, "step": 2957 }, { "epoch": 0.37628800407072893, "ewc_loss": 0.051513671875, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.284651279449463, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.865810751914978, "num_tokens": 112440063.0, "step": 2958 }, { "epoch": 0.3764152143493194, "ewc_loss": 0.044677734375, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.8250932693481445, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.8775423765182495, "num_tokens": 112478856.0, "step": 2959 }, { "epoch": 0.37654242462790993, "ewc_loss": 0.054443359375, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.461822271347046, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.8526229858398438, "num_tokens": 112515707.0, "step": 2960 }, { "epoch": 0.37666963490650046, "ewc_loss": 0.045166015625, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.7932984828948975, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.844754695892334, "num_tokens": 112551302.0, "step": 2961 }, { "epoch": 0.37679684518509093, "ewc_loss": 0.06005859375, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6798360347747803, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8639897108078003, "num_tokens": 112585821.0, "step": 2962 }, { "epoch": 0.37692405546368146, "ewc_loss": 0.048095703125, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.843242645263672, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8532852530479431, "num_tokens": 112624261.0, "step": 2963 }, { "epoch": 0.377051265742272, "ewc_loss": 0.06591796875, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9052536487579346, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8681859970092773, "num_tokens": 112656944.0, "step": 2964 }, { "epoch": 0.37717847602086246, "ewc_loss": 0.05615234375, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.085675001144409, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8596290946006775, "num_tokens": 112694386.0, "step": 2965 }, { "epoch": 0.377305686299453, "ewc_loss": 0.0673828125, "ewc_loss_diag": 7.390975952148438e-05, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.95509934425354, "learning_rate": 1e-05, "loss": 0.5204, "mean_token_accuracy": 0.8530114889144897, "num_tokens": 112731597.0, "step": 2966 }, { "epoch": 0.3774328965780435, "ewc_loss": 0.06005859375, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.336148262023926, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8786778450012207, "num_tokens": 112768012.0, "step": 2967 }, { "epoch": 0.377560106856634, "ewc_loss": 0.0634765625, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8098621368408203, "learning_rate": 1e-05, "loss": 0.5426, "mean_token_accuracy": 0.839706540107727, "num_tokens": 112798089.0, "step": 2968 }, { "epoch": 0.3776873171352245, "ewc_loss": 0.060546875, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.39850115776062, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8565863370895386, "num_tokens": 112842332.0, "step": 2969 }, { "epoch": 0.37781452741381505, "ewc_loss": 0.05810546875, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.600019931793213, "learning_rate": 1e-05, "loss": 0.5162, "mean_token_accuracy": 0.8473597764968872, "num_tokens": 112882382.0, "step": 2970 }, { "epoch": 0.3779417376924055, "ewc_loss": 0.056640625, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3316519260406494, "learning_rate": 1e-05, "loss": 0.5712, "mean_token_accuracy": 0.8299020528793335, "num_tokens": 112922627.0, "step": 2971 }, { "epoch": 0.37806894797099605, "ewc_loss": 0.0546875, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.456735134124756, "learning_rate": 1e-05, "loss": 0.5605, "mean_token_accuracy": 0.8314846754074097, "num_tokens": 112957571.0, "step": 2972 }, { "epoch": 0.3781961582495866, "ewc_loss": 0.05322265625, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2225112915039062, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8514275550842285, "num_tokens": 113001128.0, "step": 2973 }, { "epoch": 0.37832336852817705, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3564043045043945, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8581525087356567, "num_tokens": 113034515.0, "step": 2974 }, { "epoch": 0.3784505788067676, "ewc_loss": 0.0498046875, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.12613844871521, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8685258030891418, "num_tokens": 113071221.0, "step": 2975 }, { "epoch": 0.3785777890853581, "ewc_loss": 0.0517578125, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.3306450843811035, "learning_rate": 1e-05, "loss": 0.5626, "mean_token_accuracy": 0.8345544338226318, "num_tokens": 113105804.0, "step": 2976 }, { "epoch": 0.3787049993639486, "ewc_loss": 0.04736328125, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.033202886581421, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8605004549026489, "num_tokens": 113146051.0, "step": 2977 }, { "epoch": 0.3788322096425391, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.350991725921631, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8664035201072693, "num_tokens": 113186696.0, "step": 2978 }, { "epoch": 0.37895941992112964, "ewc_loss": 0.046875, "ewc_loss_diag": 7.43865966796875e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.957878828048706, "learning_rate": 1e-05, "loss": 0.5138, "mean_token_accuracy": 0.8446575403213501, "num_tokens": 113228402.0, "step": 2979 }, { "epoch": 0.37908663019972016, "ewc_loss": 0.053955078125, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4224963188171387, "learning_rate": 1e-05, "loss": 0.5304, "mean_token_accuracy": 0.8415431976318359, "num_tokens": 113270407.0, "step": 2980 }, { "epoch": 0.37921384047831064, "ewc_loss": 0.04638671875, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9277679920196533, "learning_rate": 1e-05, "loss": 0.5082, "mean_token_accuracy": 0.8460270166397095, "num_tokens": 113304523.0, "step": 2981 }, { "epoch": 0.37934105075690117, "ewc_loss": 0.056396484375, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.526268243789673, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8556305766105652, "num_tokens": 113345224.0, "step": 2982 }, { "epoch": 0.3794682610354917, "ewc_loss": 0.046875, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.923131227493286, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8633100986480713, "num_tokens": 113382287.0, "step": 2983 }, { "epoch": 0.37959547131408217, "ewc_loss": 0.06005859375, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6453025341033936, "learning_rate": 1e-05, "loss": 0.5389, "mean_token_accuracy": 0.8413358926773071, "num_tokens": 113427308.0, "step": 2984 }, { "epoch": 0.3797226815926727, "ewc_loss": 0.050048828125, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0252254009246826, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8504145741462708, "num_tokens": 113461272.0, "step": 2985 }, { "epoch": 0.3798498918712632, "ewc_loss": 0.06103515625, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.676368236541748, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8719884157180786, "num_tokens": 113499942.0, "step": 2986 }, { "epoch": 0.3799771021498537, "ewc_loss": 0.053466796875, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1423792839050293, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8713152408599854, "num_tokens": 113531801.0, "step": 2987 }, { "epoch": 0.3801043124284442, "ewc_loss": 0.05908203125, "ewc_loss_diag": 7.486343383789062e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.578601360321045, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8691458702087402, "num_tokens": 113567475.0, "step": 2988 }, { "epoch": 0.38023152270703475, "ewc_loss": 0.0537109375, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2167978286743164, "learning_rate": 1e-05, "loss": 0.5314, "mean_token_accuracy": 0.8407242298126221, "num_tokens": 113603767.0, "step": 2989 }, { "epoch": 0.3803587329856252, "ewc_loss": 0.053955078125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4140138626098633, "learning_rate": 1e-05, "loss": 0.5099, "mean_token_accuracy": 0.8490151762962341, "num_tokens": 113641213.0, "step": 2990 }, { "epoch": 0.38048594326421575, "ewc_loss": 0.053466796875, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2315948009490967, "learning_rate": 1e-05, "loss": 0.574, "mean_token_accuracy": 0.8284905552864075, "num_tokens": 113683364.0, "step": 2991 }, { "epoch": 0.3806131535428063, "ewc_loss": 0.05078125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.2397050857543945, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8494428396224976, "num_tokens": 113718488.0, "step": 2992 }, { "epoch": 0.38074036382139675, "ewc_loss": 0.05224609375, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2196922302246094, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8652612566947937, "num_tokens": 113756664.0, "step": 2993 }, { "epoch": 0.3808675740999873, "ewc_loss": 0.048095703125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0890746116638184, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8585760593414307, "num_tokens": 113790711.0, "step": 2994 }, { "epoch": 0.3809947843785778, "ewc_loss": 0.050537109375, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.2305140495300293, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8631515502929688, "num_tokens": 113827091.0, "step": 2995 }, { "epoch": 0.3811219946571683, "ewc_loss": 0.046630859375, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.0116312503814697, "learning_rate": 1e-05, "loss": 0.5295, "mean_token_accuracy": 0.8401697278022766, "num_tokens": 113862971.0, "step": 2996 }, { "epoch": 0.3812492049357588, "ewc_loss": 0.05126953125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.253955125808716, "learning_rate": 1e-05, "loss": 0.5152, "mean_token_accuracy": 0.8476899862289429, "num_tokens": 113897430.0, "step": 2997 }, { "epoch": 0.38137641521434934, "ewc_loss": 0.046142578125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.9420900344848633, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.8506007194519043, "num_tokens": 113926626.0, "step": 2998 }, { "epoch": 0.3815036254929398, "ewc_loss": 0.052001953125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.290208578109741, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8607586026191711, "num_tokens": 113969656.0, "step": 2999 }, { "epoch": 0.38163083577153034, "ewc_loss": 0.04541015625, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.872884511947632, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8677406907081604, "num_tokens": 114006140.0, "step": 3000 }, { "epoch": 0.38175804605012087, "ewc_loss": 0.053955078125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4055774211883545, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.858220100402832, "num_tokens": 114047300.0, "step": 3001 }, { "epoch": 0.38188525632871134, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.8344218730926514, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8477993011474609, "num_tokens": 114086129.0, "step": 3002 }, { "epoch": 0.38201246660730187, "ewc_loss": 0.0576171875, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.562180280685425, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8532766699790955, "num_tokens": 114123345.0, "step": 3003 }, { "epoch": 0.3821396768858924, "ewc_loss": 0.046630859375, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.85137939453125, "learning_rate": 1e-05, "loss": 0.4878, "mean_token_accuracy": 0.8507203459739685, "num_tokens": 114158983.0, "step": 3004 }, { "epoch": 0.38226688716448287, "ewc_loss": 0.0625, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7509677410125732, "learning_rate": 1e-05, "loss": 0.5735, "mean_token_accuracy": 0.8297961354255676, "num_tokens": 114199055.0, "step": 3005 }, { "epoch": 0.3823940974430734, "ewc_loss": 0.0517578125, "ewc_loss_diag": 7.534027099609375e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.969468593597412, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8587128520011902, "num_tokens": 114236355.0, "step": 3006 }, { "epoch": 0.38252130772166393, "ewc_loss": 0.064453125, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8364388942718506, "learning_rate": 1e-05, "loss": 0.529, "mean_token_accuracy": 0.8465863466262817, "num_tokens": 114285111.0, "step": 3007 }, { "epoch": 0.3826485180002544, "ewc_loss": 0.057861328125, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.1966137886047363, "learning_rate": 1e-05, "loss": 0.5594, "mean_token_accuracy": 0.8297747373580933, "num_tokens": 114319306.0, "step": 3008 }, { "epoch": 0.38277572827884493, "ewc_loss": 0.0634765625, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.746722459793091, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8651366233825684, "num_tokens": 114354117.0, "step": 3009 }, { "epoch": 0.38290293855743546, "ewc_loss": 0.05859375, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2860193252563477, "learning_rate": 1e-05, "loss": 0.5366, "mean_token_accuracy": 0.842972993850708, "num_tokens": 114395865.0, "step": 3010 }, { "epoch": 0.38303014883602593, "ewc_loss": 0.059326171875, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.587146282196045, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.859420657157898, "num_tokens": 114427862.0, "step": 3011 }, { "epoch": 0.38315735911461646, "ewc_loss": 0.056884765625, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2850160598754883, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8622833490371704, "num_tokens": 114462892.0, "step": 3012 }, { "epoch": 0.383284569393207, "ewc_loss": 0.054931640625, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4108903408050537, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.863968551158905, "num_tokens": 114500477.0, "step": 3013 }, { "epoch": 0.38341177967179746, "ewc_loss": 0.05419921875, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2345409393310547, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8507217764854431, "num_tokens": 114534695.0, "step": 3014 }, { "epoch": 0.383538989950388, "ewc_loss": 0.052001953125, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.277249813079834, "learning_rate": 1e-05, "loss": 0.5435, "mean_token_accuracy": 0.8363634943962097, "num_tokens": 114579969.0, "step": 3015 }, { "epoch": 0.3836662002289785, "ewc_loss": 0.05126953125, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1916110515594482, "learning_rate": 1e-05, "loss": 0.5547, "mean_token_accuracy": 0.8357431292533875, "num_tokens": 114614904.0, "step": 3016 }, { "epoch": 0.383793410507569, "ewc_loss": 0.050048828125, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.196580648422241, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8637840747833252, "num_tokens": 114647126.0, "step": 3017 }, { "epoch": 0.3839206207861595, "ewc_loss": 0.048828125, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.126725912094116, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8523932695388794, "num_tokens": 114683529.0, "step": 3018 }, { "epoch": 0.38404783106475004, "ewc_loss": 0.048828125, "ewc_loss_diag": 7.581710815429688e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.1422362327575684, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8602594137191772, "num_tokens": 114722471.0, "step": 3019 }, { "epoch": 0.3841750413433405, "ewc_loss": 0.047607421875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.076802968978882, "learning_rate": 1e-05, "loss": 0.515, "mean_token_accuracy": 0.8469551801681519, "num_tokens": 114765243.0, "step": 3020 }, { "epoch": 0.38430225162193105, "ewc_loss": 0.048095703125, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.1240530014038086, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8558223247528076, "num_tokens": 114806341.0, "step": 3021 }, { "epoch": 0.3844294619005216, "ewc_loss": 0.046875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0545361042022705, "learning_rate": 1e-05, "loss": 0.5854, "mean_token_accuracy": 0.8280714154243469, "num_tokens": 114850014.0, "step": 3022 }, { "epoch": 0.38455667217911205, "ewc_loss": 0.048095703125, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.1338093280792236, "learning_rate": 1e-05, "loss": 0.5176, "mean_token_accuracy": 0.8477051854133606, "num_tokens": 114884771.0, "step": 3023 }, { "epoch": 0.3846838824577026, "ewc_loss": 0.04638671875, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0026278495788574, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8598778247833252, "num_tokens": 114926319.0, "step": 3024 }, { "epoch": 0.3848110927362931, "ewc_loss": 0.048095703125, "ewc_loss_diag": 7.62939453125e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.145766258239746, "learning_rate": 1e-05, "loss": 0.5213, "mean_token_accuracy": 0.8448160290718079, "num_tokens": 114971008.0, "step": 3025 }, { "epoch": 0.3849383030148836, "ewc_loss": 0.04638671875, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.992476463317871, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8490734100341797, "num_tokens": 115007294.0, "step": 3026 }, { "epoch": 0.3850655132934741, "ewc_loss": 0.048828125, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.169360399246216, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8571637868881226, "num_tokens": 115043066.0, "step": 3027 }, { "epoch": 0.38519272357206463, "ewc_loss": 0.0458984375, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.9353957176208496, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8568551540374756, "num_tokens": 115077404.0, "step": 3028 }, { "epoch": 0.3853199338506551, "ewc_loss": 0.05078125, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.225513458251953, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8636913299560547, "num_tokens": 115114092.0, "step": 3029 }, { "epoch": 0.38544714412924563, "ewc_loss": 0.04541015625, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.86069655418396, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8652232885360718, "num_tokens": 115157095.0, "step": 3030 }, { "epoch": 0.38557435440783616, "ewc_loss": 0.05322265625, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3641295433044434, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8671910762786865, "num_tokens": 115189752.0, "step": 3031 }, { "epoch": 0.3857015646864267, "ewc_loss": 0.044921875, "ewc_loss_diag": 7.677078247070312e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.785205125808716, "learning_rate": 1e-05, "loss": 0.5607, "mean_token_accuracy": 0.8317900896072388, "num_tokens": 115232650.0, "step": 3032 }, { "epoch": 0.38582877496501716, "ewc_loss": 0.05712890625, "ewc_loss_diag": 7.724761962890625e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5846219062805176, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8661050200462341, "num_tokens": 115267812.0, "step": 3033 }, { "epoch": 0.3859559852436077, "ewc_loss": 0.046142578125, "ewc_loss_diag": 7.724761962890625e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.782299518585205, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8601605892181396, "num_tokens": 115300905.0, "step": 3034 }, { "epoch": 0.3860831955221982, "ewc_loss": 0.064453125, "ewc_loss_diag": 7.724761962890625e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8502235412597656, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8635261058807373, "num_tokens": 115341869.0, "step": 3035 }, { "epoch": 0.3862104058007887, "ewc_loss": 0.052978515625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 2.9743752479553223, "learning_rate": 1e-05, "loss": 0.5668, "mean_token_accuracy": 0.8323136568069458, "num_tokens": 115380499.0, "step": 3036 }, { "epoch": 0.3863376160793792, "ewc_loss": 0.068359375, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.0170416831970215, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8624081611633301, "num_tokens": 115413573.0, "step": 3037 }, { "epoch": 0.38646482635796975, "ewc_loss": 0.060302734375, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.269345998764038, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8625144362449646, "num_tokens": 115452634.0, "step": 3038 }, { "epoch": 0.3865920366365602, "ewc_loss": 0.06689453125, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.945599317550659, "learning_rate": 1e-05, "loss": 0.598, "mean_token_accuracy": 0.8293643593788147, "num_tokens": 115492509.0, "step": 3039 }, { "epoch": 0.38671924691515075, "ewc_loss": 0.06201171875, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.4548215866088867, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8647876381874084, "num_tokens": 115531256.0, "step": 3040 }, { "epoch": 0.3868464571937413, "ewc_loss": 0.061279296875, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.7232882976531982, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.8449540138244629, "num_tokens": 115564173.0, "step": 3041 }, { "epoch": 0.38697366747233175, "ewc_loss": 0.0595703125, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4281182289123535, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8592770099639893, "num_tokens": 115598359.0, "step": 3042 }, { "epoch": 0.3871008777509223, "ewc_loss": 0.056884765625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5258524417877197, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8510007262229919, "num_tokens": 115637782.0, "step": 3043 }, { "epoch": 0.3872280880295128, "ewc_loss": 0.0556640625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.299370527267456, "learning_rate": 1e-05, "loss": 0.5536, "mean_token_accuracy": 0.8388925194740295, "num_tokens": 115680809.0, "step": 3044 }, { "epoch": 0.3873552983081033, "ewc_loss": 0.05419921875, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4147579669952393, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8510653972625732, "num_tokens": 115712092.0, "step": 3045 }, { "epoch": 0.3874825085866938, "ewc_loss": 0.052978515625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1973109245300293, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8509780764579773, "num_tokens": 115749400.0, "step": 3046 }, { "epoch": 0.38760971886528434, "ewc_loss": 0.052978515625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.347883939743042, "learning_rate": 1e-05, "loss": 0.5115, "mean_token_accuracy": 0.8480691313743591, "num_tokens": 115783932.0, "step": 3047 }, { "epoch": 0.3877369291438748, "ewc_loss": 0.04931640625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0973117351531982, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8530672788619995, "num_tokens": 115819358.0, "step": 3048 }, { "epoch": 0.38786413942246534, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.3316028118133545, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8644758462905884, "num_tokens": 115855370.0, "step": 3049 }, { "epoch": 0.38799134970105587, "ewc_loss": 0.047607421875, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.010861396789551, "learning_rate": 1e-05, "loss": 0.5429, "mean_token_accuracy": 0.8361215591430664, "num_tokens": 115894722.0, "step": 3050 }, { "epoch": 0.38811855997964634, "ewc_loss": 0.053466796875, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3775217533111572, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8599413633346558, "num_tokens": 115935276.0, "step": 3051 }, { "epoch": 0.38824577025823687, "ewc_loss": 0.046630859375, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.939349412918091, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.852559506893158, "num_tokens": 115971255.0, "step": 3052 }, { "epoch": 0.3883729805368274, "ewc_loss": 0.055908203125, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.491792678833008, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8643772006034851, "num_tokens": 116003818.0, "step": 3053 }, { "epoch": 0.38850019081541787, "ewc_loss": 0.046630859375, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.908672571182251, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8580941557884216, "num_tokens": 116040720.0, "step": 3054 }, { "epoch": 0.3886274010940084, "ewc_loss": 0.058837890625, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.6294403076171875, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8673481345176697, "num_tokens": 116076485.0, "step": 3055 }, { "epoch": 0.3887546113725989, "ewc_loss": 0.048583984375, "ewc_loss_diag": 7.772445678710938e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.9648470878601074, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8653044700622559, "num_tokens": 116109802.0, "step": 3056 }, { "epoch": 0.3888818216511894, "ewc_loss": 0.0625, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7399983406066895, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.8529824018478394, "num_tokens": 116149865.0, "step": 3057 }, { "epoch": 0.3890090319297799, "ewc_loss": 0.0537109375, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1242759227752686, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.8564577102661133, "num_tokens": 116185887.0, "step": 3058 }, { "epoch": 0.38913624220837045, "ewc_loss": 0.061767578125, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.692645311355591, "learning_rate": 1e-05, "loss": 0.5438, "mean_token_accuracy": 0.8455384969711304, "num_tokens": 116233090.0, "step": 3059 }, { "epoch": 0.3892634524869609, "ewc_loss": 0.054931640625, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2368125915527344, "learning_rate": 1e-05, "loss": 0.5104, "mean_token_accuracy": 0.8505120277404785, "num_tokens": 116274433.0, "step": 3060 }, { "epoch": 0.38939066276555145, "ewc_loss": 0.057373046875, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5536575317382812, "learning_rate": 1e-05, "loss": 0.5717, "mean_token_accuracy": 0.8351787328720093, "num_tokens": 116309104.0, "step": 3061 }, { "epoch": 0.389517873044142, "ewc_loss": 0.054931640625, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2701501846313477, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.859790027141571, "num_tokens": 116350248.0, "step": 3062 }, { "epoch": 0.38964508332273246, "ewc_loss": 0.05322265625, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3463642597198486, "learning_rate": 1e-05, "loss": 0.5263, "mean_token_accuracy": 0.8424586057662964, "num_tokens": 116391078.0, "step": 3063 }, { "epoch": 0.389772293601323, "ewc_loss": 0.053955078125, "ewc_loss_diag": 7.82012939453125e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2541966438293457, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8501807451248169, "num_tokens": 116430652.0, "step": 3064 }, { "epoch": 0.3898995038799135, "ewc_loss": 0.050048828125, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.186708450317383, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8652939796447754, "num_tokens": 116467024.0, "step": 3065 }, { "epoch": 0.390026714158504, "ewc_loss": 0.0517578125, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.242929220199585, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8618793487548828, "num_tokens": 116504304.0, "step": 3066 }, { "epoch": 0.3901539244370945, "ewc_loss": 0.047607421875, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.062166690826416, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8477644920349121, "num_tokens": 116541079.0, "step": 3067 }, { "epoch": 0.39028113471568504, "ewc_loss": 0.051513671875, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.257988452911377, "learning_rate": 1e-05, "loss": 0.5359, "mean_token_accuracy": 0.8442727327346802, "num_tokens": 116578690.0, "step": 3068 }, { "epoch": 0.3904083449942755, "ewc_loss": 0.04638671875, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.9703781604766846, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.851946234703064, "num_tokens": 116614945.0, "step": 3069 }, { "epoch": 0.39053555527286604, "ewc_loss": 0.05224609375, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.3043203353881836, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8676496744155884, "num_tokens": 116646795.0, "step": 3070 }, { "epoch": 0.39066276555145657, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.9323203563690186, "learning_rate": 1e-05, "loss": 0.5082, "mean_token_accuracy": 0.8461316823959351, "num_tokens": 116685731.0, "step": 3071 }, { "epoch": 0.39078997583004704, "ewc_loss": 0.053466796875, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.354579210281372, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8736788034439087, "num_tokens": 116722559.0, "step": 3072 }, { "epoch": 0.39091718610863757, "ewc_loss": 0.045654296875, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.869309902191162, "learning_rate": 1e-05, "loss": 0.5442, "mean_token_accuracy": 0.8377694487571716, "num_tokens": 116762954.0, "step": 3073 }, { "epoch": 0.3910443963872281, "ewc_loss": 0.0546875, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.469139337539673, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8637853860855103, "num_tokens": 116801745.0, "step": 3074 }, { "epoch": 0.39117160666581857, "ewc_loss": 0.046630859375, "ewc_loss_diag": 7.867813110351562e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.8390848636627197, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8668175935745239, "num_tokens": 116837650.0, "step": 3075 }, { "epoch": 0.3912988169444091, "ewc_loss": 0.06005859375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6535654067993164, "learning_rate": 1e-05, "loss": 0.5321, "mean_token_accuracy": 0.8440742492675781, "num_tokens": 116872722.0, "step": 3076 }, { "epoch": 0.39142602722299963, "ewc_loss": 0.0498046875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.913587808609009, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8570972681045532, "num_tokens": 116911175.0, "step": 3077 }, { "epoch": 0.3915532375015901, "ewc_loss": 0.0634765625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7820169925689697, "learning_rate": 1e-05, "loss": 0.4993, "mean_token_accuracy": 0.8514253497123718, "num_tokens": 116946317.0, "step": 3078 }, { "epoch": 0.39168044778018063, "ewc_loss": 0.05517578125, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.0727787017822266, "learning_rate": 1e-05, "loss": 0.5549, "mean_token_accuracy": 0.836840808391571, "num_tokens": 116988661.0, "step": 3079 }, { "epoch": 0.39180765805877116, "ewc_loss": 0.06396484375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7877917289733887, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8600174784660339, "num_tokens": 117027309.0, "step": 3080 }, { "epoch": 0.3919348683373617, "ewc_loss": 0.05810546875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2288527488708496, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8573293685913086, "num_tokens": 117065608.0, "step": 3081 }, { "epoch": 0.39206207861595216, "ewc_loss": 0.06103515625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.66300892829895, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8498532772064209, "num_tokens": 117101824.0, "step": 3082 }, { "epoch": 0.3921892888945427, "ewc_loss": 0.0576171875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2832632064819336, "learning_rate": 1e-05, "loss": 0.5192, "mean_token_accuracy": 0.8487012386322021, "num_tokens": 117141985.0, "step": 3083 }, { "epoch": 0.3923164991731332, "ewc_loss": 0.05615234375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4809975624084473, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.860439121723175, "num_tokens": 117177606.0, "step": 3084 }, { "epoch": 0.3924437094517237, "ewc_loss": 0.05517578125, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2606146335601807, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8540236949920654, "num_tokens": 117220214.0, "step": 3085 }, { "epoch": 0.3925709197303142, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.324558973312378, "learning_rate": 1e-05, "loss": 0.5113, "mean_token_accuracy": 0.8486006259918213, "num_tokens": 117259188.0, "step": 3086 }, { "epoch": 0.39269813000890474, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2192726135253906, "learning_rate": 1e-05, "loss": 0.5042, "mean_token_accuracy": 0.8521384596824646, "num_tokens": 117299146.0, "step": 3087 }, { "epoch": 0.3928253402874952, "ewc_loss": 0.05126953125, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.2069084644317627, "learning_rate": 1e-05, "loss": 0.4966, "mean_token_accuracy": 0.8516018390655518, "num_tokens": 117337560.0, "step": 3088 }, { "epoch": 0.39295255056608575, "ewc_loss": 0.05029296875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1680681705474854, "learning_rate": 1e-05, "loss": 0.5399, "mean_token_accuracy": 0.8407667875289917, "num_tokens": 117376935.0, "step": 3089 }, { "epoch": 0.3930797608446763, "ewc_loss": 0.0498046875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1394879817962646, "learning_rate": 1e-05, "loss": 0.5017, "mean_token_accuracy": 0.8549425005912781, "num_tokens": 117420273.0, "step": 3090 }, { "epoch": 0.39320697112326675, "ewc_loss": 0.048828125, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.1267690658569336, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8676917552947998, "num_tokens": 117457338.0, "step": 3091 }, { "epoch": 0.3933341814018573, "ewc_loss": 0.04833984375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.108973979949951, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8639477491378784, "num_tokens": 117491082.0, "step": 3092 }, { "epoch": 0.3934613916804478, "ewc_loss": 0.0478515625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0943081378936768, "learning_rate": 1e-05, "loss": 0.3888, "mean_token_accuracy": 0.8831760287284851, "num_tokens": 117520425.0, "step": 3093 }, { "epoch": 0.3935886019590383, "ewc_loss": 0.04736328125, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0770273208618164, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8645436763763428, "num_tokens": 117559021.0, "step": 3094 }, { "epoch": 0.3937158122376288, "ewc_loss": 0.046875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.1563668251037598, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.862708330154419, "num_tokens": 117598705.0, "step": 3095 }, { "epoch": 0.39384302251621933, "ewc_loss": 0.046875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.014608144760132, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.866276741027832, "num_tokens": 117636185.0, "step": 3096 }, { "epoch": 0.3939702327948098, "ewc_loss": 0.048828125, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1434695720672607, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.86626136302948, "num_tokens": 117674290.0, "step": 3097 }, { "epoch": 0.39409744307340033, "ewc_loss": 0.04638671875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9574265480041504, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.857836127281189, "num_tokens": 117708708.0, "step": 3098 }, { "epoch": 0.39422465335199086, "ewc_loss": 0.05029296875, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.2056729793548584, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8532960414886475, "num_tokens": 117742528.0, "step": 3099 }, { "epoch": 0.39435186363058133, "ewc_loss": 0.0458984375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.907930850982666, "learning_rate": 1e-05, "loss": 0.5188, "mean_token_accuracy": 0.8427777290344238, "num_tokens": 117779050.0, "step": 3100 }, { "epoch": 0.39447907390917186, "ewc_loss": 0.052734375, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2833664417266846, "learning_rate": 1e-05, "loss": 0.5271, "mean_token_accuracy": 0.8427811861038208, "num_tokens": 117824004.0, "step": 3101 }, { "epoch": 0.3946062841877624, "ewc_loss": 0.04541015625, "ewc_loss_diag": 7.915496826171875e-05, "ewc_loss_parallel": 3.7670135498046875e-05, "grad_norm": 2.8180978298187256, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8651153445243835, "num_tokens": 117864837.0, "step": 3102 }, { "epoch": 0.39473349446635286, "ewc_loss": 0.0546875, "ewc_loss_diag": 7.963180541992188e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4661972522735596, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8578125834465027, "num_tokens": 117903675.0, "step": 3103 }, { "epoch": 0.3948607047449434, "ewc_loss": 0.0458984375, "ewc_loss_diag": 7.963180541992188e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.764446258544922, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8493392467498779, "num_tokens": 117942561.0, "step": 3104 }, { "epoch": 0.3949879150235339, "ewc_loss": 0.0625, "ewc_loss_diag": 7.963180541992188e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.727675199508667, "learning_rate": 1e-05, "loss": 0.53, "mean_token_accuracy": 0.8489761352539062, "num_tokens": 117982439.0, "step": 3105 }, { "epoch": 0.3951151253021244, "ewc_loss": 0.04931640625, "ewc_loss_diag": 7.963180541992188e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.812053680419922, "learning_rate": 1e-05, "loss": 0.5028, "mean_token_accuracy": 0.8511879444122314, "num_tokens": 118025153.0, "step": 3106 }, { "epoch": 0.3952423355807149, "ewc_loss": 0.06884765625, "ewc_loss_diag": 7.963180541992188e-05, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.021262168884277, "learning_rate": 1e-05, "loss": 0.5233, "mean_token_accuracy": 0.8488117456436157, "num_tokens": 118064508.0, "step": 3107 }, { "epoch": 0.39536954585930545, "ewc_loss": 0.0595703125, "ewc_loss_diag": 7.963180541992188e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.162885904312134, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8594179153442383, "num_tokens": 118093368.0, "step": 3108 }, { "epoch": 0.3954967561378959, "ewc_loss": 0.06884765625, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.052834987640381, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8674928545951843, "num_tokens": 118129164.0, "step": 3109 }, { "epoch": 0.39562396641648645, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.452388048171997, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8616456389427185, "num_tokens": 118169092.0, "step": 3110 }, { "epoch": 0.395751176695077, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.82904314994812, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8566473722457886, "num_tokens": 118211616.0, "step": 3111 }, { "epoch": 0.39587838697366745, "ewc_loss": 0.062255859375, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.4793362617492676, "learning_rate": 1e-05, "loss": 0.5114, "mean_token_accuracy": 0.8484095335006714, "num_tokens": 118248483.0, "step": 3112 }, { "epoch": 0.396005597252258, "ewc_loss": 0.0595703125, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6208834648132324, "learning_rate": 1e-05, "loss": 0.5039, "mean_token_accuracy": 0.8547478914260864, "num_tokens": 118288540.0, "step": 3113 }, { "epoch": 0.3961328075308485, "ewc_loss": 0.058349609375, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3758926391601562, "learning_rate": 1e-05, "loss": 0.5006, "mean_token_accuracy": 0.8550320267677307, "num_tokens": 118326988.0, "step": 3114 }, { "epoch": 0.396260017809439, "ewc_loss": 0.05615234375, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4604549407958984, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8559862375259399, "num_tokens": 118366100.0, "step": 3115 }, { "epoch": 0.3963872280880295, "ewc_loss": 0.053955078125, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2710659503936768, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8530553579330444, "num_tokens": 118397243.0, "step": 3116 }, { "epoch": 0.39651443836662004, "ewc_loss": 0.053466796875, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3619110584259033, "learning_rate": 1e-05, "loss": 0.5528, "mean_token_accuracy": 0.8386805653572083, "num_tokens": 118438198.0, "step": 3117 }, { "epoch": 0.3966416486452105, "ewc_loss": 0.05078125, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.165106773376465, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8515681028366089, "num_tokens": 118473496.0, "step": 3118 }, { "epoch": 0.39676885892380104, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.316112518310547, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8654172420501709, "num_tokens": 118511988.0, "step": 3119 }, { "epoch": 0.39689606920239157, "ewc_loss": 0.04833984375, "ewc_loss_diag": 8.0108642578125e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0566537380218506, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8634684085845947, "num_tokens": 118548287.0, "step": 3120 }, { "epoch": 0.39702327948098204, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.35062313079834, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8621840476989746, "num_tokens": 118580637.0, "step": 3121 }, { "epoch": 0.39715048975957257, "ewc_loss": 0.047119140625, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.964317798614502, "learning_rate": 1e-05, "loss": 0.5091, "mean_token_accuracy": 0.8466871976852417, "num_tokens": 118622316.0, "step": 3122 }, { "epoch": 0.3972777000381631, "ewc_loss": 0.054443359375, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.443763494491577, "learning_rate": 1e-05, "loss": 0.5249, "mean_token_accuracy": 0.8432461619377136, "num_tokens": 118660497.0, "step": 3123 }, { "epoch": 0.39740491031675357, "ewc_loss": 0.046630859375, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.912006378173828, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8485913276672363, "num_tokens": 118691733.0, "step": 3124 }, { "epoch": 0.3975321205953441, "ewc_loss": 0.05810546875, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.595118761062622, "learning_rate": 1e-05, "loss": 0.5169, "mean_token_accuracy": 0.8422605395317078, "num_tokens": 118728576.0, "step": 3125 }, { "epoch": 0.3976593308739346, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9114794731140137, "learning_rate": 1e-05, "loss": 0.5289, "mean_token_accuracy": 0.8434316515922546, "num_tokens": 118765394.0, "step": 3126 }, { "epoch": 0.3977865411525251, "ewc_loss": 0.06298828125, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7598636150360107, "learning_rate": 1e-05, "loss": 0.5275, "mean_token_accuracy": 0.8414909839630127, "num_tokens": 118804856.0, "step": 3127 }, { "epoch": 0.3979137514311156, "ewc_loss": 0.053466796875, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0706653594970703, "learning_rate": 1e-05, "loss": 0.492, "mean_token_accuracy": 0.8542224168777466, "num_tokens": 118839274.0, "step": 3128 }, { "epoch": 0.39804096170970615, "ewc_loss": 0.06396484375, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.791572332382202, "learning_rate": 1e-05, "loss": 0.5323, "mean_token_accuracy": 0.8438698053359985, "num_tokens": 118877866.0, "step": 3129 }, { "epoch": 0.3981681719882967, "ewc_loss": 0.05712890625, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.255072832107544, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8682496547698975, "num_tokens": 118908942.0, "step": 3130 }, { "epoch": 0.39829538226688715, "ewc_loss": 0.06005859375, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.636704444885254, "learning_rate": 1e-05, "loss": 0.5442, "mean_token_accuracy": 0.8392491340637207, "num_tokens": 118948957.0, "step": 3131 }, { "epoch": 0.3984225925454777, "ewc_loss": 0.0576171875, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3116073608398438, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8599698543548584, "num_tokens": 118991246.0, "step": 3132 }, { "epoch": 0.3985498028240682, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.440852165222168, "learning_rate": 1e-05, "loss": 0.5356, "mean_token_accuracy": 0.8384428024291992, "num_tokens": 119025661.0, "step": 3133 }, { "epoch": 0.3986770131026587, "ewc_loss": 0.054931640625, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.307239055633545, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8507193922996521, "num_tokens": 119059524.0, "step": 3134 }, { "epoch": 0.3988042233812492, "ewc_loss": 0.051513671875, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2404637336730957, "learning_rate": 1e-05, "loss": 0.5272, "mean_token_accuracy": 0.8448047041893005, "num_tokens": 119101636.0, "step": 3135 }, { "epoch": 0.39893143365983974, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2819297313690186, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.8478801250457764, "num_tokens": 119138448.0, "step": 3136 }, { "epoch": 0.3990586439384302, "ewc_loss": 0.048583984375, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0883400440216064, "learning_rate": 1e-05, "loss": 0.5146, "mean_token_accuracy": 0.8419978022575378, "num_tokens": 119183337.0, "step": 3137 }, { "epoch": 0.39918585421702074, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.270451307296753, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8800222873687744, "num_tokens": 119217503.0, "step": 3138 }, { "epoch": 0.39931306449561127, "ewc_loss": 0.046875, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0033204555511475, "learning_rate": 1e-05, "loss": 0.541, "mean_token_accuracy": 0.8405189514160156, "num_tokens": 119255824.0, "step": 3139 }, { "epoch": 0.39944027477420174, "ewc_loss": 0.052490234375, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.3087921142578125, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8497647047042847, "num_tokens": 119292677.0, "step": 3140 }, { "epoch": 0.39956748505279227, "ewc_loss": 0.046630859375, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.925487756729126, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8566749691963196, "num_tokens": 119330961.0, "step": 3141 }, { "epoch": 0.3996946953313828, "ewc_loss": 0.053955078125, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.40551495552063, "learning_rate": 1e-05, "loss": 0.5322, "mean_token_accuracy": 0.8423798680305481, "num_tokens": 119362122.0, "step": 3142 }, { "epoch": 0.39982190560997327, "ewc_loss": 0.04638671875, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.8652231693267822, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.856331467628479, "num_tokens": 119398018.0, "step": 3143 }, { "epoch": 0.3999491158885638, "ewc_loss": 0.05712890625, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5361881256103516, "learning_rate": 1e-05, "loss": 0.5296, "mean_token_accuracy": 0.8460051417350769, "num_tokens": 119438552.0, "step": 3144 }, { "epoch": 0.40007632616715433, "ewc_loss": 0.046875, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.8690991401672363, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.851389467716217, "num_tokens": 119470777.0, "step": 3145 }, { "epoch": 0.4002035364457448, "ewc_loss": 0.061767578125, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.709465265274048, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8664970397949219, "num_tokens": 119505062.0, "step": 3146 }, { "epoch": 0.40033074672433533, "ewc_loss": 0.0517578125, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.9668750762939453, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8488806486129761, "num_tokens": 119540394.0, "step": 3147 }, { "epoch": 0.40045795700292586, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.058547973632812e-05, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.794154167175293, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8750609159469604, "num_tokens": 119581657.0, "step": 3148 }, { "epoch": 0.40058516728151633, "ewc_loss": 0.056884765625, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.1585066318511963, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8607938289642334, "num_tokens": 119617140.0, "step": 3149 }, { "epoch": 0.40071237756010686, "ewc_loss": 0.0634765625, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.737802028656006, "learning_rate": 1e-05, "loss": 0.5578, "mean_token_accuracy": 0.8387669920921326, "num_tokens": 119660155.0, "step": 3150 }, { "epoch": 0.4008395878386974, "ewc_loss": 0.05810546875, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2460885047912598, "learning_rate": 1e-05, "loss": 0.5356, "mean_token_accuracy": 0.8419482707977295, "num_tokens": 119706258.0, "step": 3151 }, { "epoch": 0.40096679811728786, "ewc_loss": 0.06005859375, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5995821952819824, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8564701080322266, "num_tokens": 119742564.0, "step": 3152 }, { "epoch": 0.4010940083958784, "ewc_loss": 0.056884765625, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.267880439758301, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8544389009475708, "num_tokens": 119781301.0, "step": 3153 }, { "epoch": 0.4012212186744689, "ewc_loss": 0.055908203125, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4460291862487793, "learning_rate": 1e-05, "loss": 0.5092, "mean_token_accuracy": 0.8478810787200928, "num_tokens": 119816513.0, "step": 3154 }, { "epoch": 0.4013484289530594, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2259528636932373, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8587661385536194, "num_tokens": 119858510.0, "step": 3155 }, { "epoch": 0.4014756392316499, "ewc_loss": 0.0537109375, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.310455799102783, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8569481372833252, "num_tokens": 119902243.0, "step": 3156 }, { "epoch": 0.40160284951024044, "ewc_loss": 0.0517578125, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.178520441055298, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.862204909324646, "num_tokens": 119938956.0, "step": 3157 }, { "epoch": 0.4017300597888309, "ewc_loss": 0.0517578125, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.221403121948242, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.864510178565979, "num_tokens": 119976070.0, "step": 3158 }, { "epoch": 0.40185727006742145, "ewc_loss": 0.050048828125, "ewc_loss_diag": 8.106231689453125e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.1294848918914795, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8598192930221558, "num_tokens": 120011811.0, "step": 3159 }, { "epoch": 0.401984480346012, "ewc_loss": 0.05078125, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.198310613632202, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8484675884246826, "num_tokens": 120042278.0, "step": 3160 }, { "epoch": 0.40211169062460245, "ewc_loss": 0.0478515625, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0686864852905273, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8537640571594238, "num_tokens": 120078379.0, "step": 3161 }, { "epoch": 0.402238900903193, "ewc_loss": 0.05078125, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.178980827331543, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.865838348865509, "num_tokens": 120118852.0, "step": 3162 }, { "epoch": 0.4023661111817835, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.0279364585876465, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8654491901397705, "num_tokens": 120153825.0, "step": 3163 }, { "epoch": 0.402493321460374, "ewc_loss": 0.05029296875, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.203859806060791, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8510582447052002, "num_tokens": 120190945.0, "step": 3164 }, { "epoch": 0.4026205317389645, "ewc_loss": 0.046875, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.9923629760742188, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8613545298576355, "num_tokens": 120221435.0, "step": 3165 }, { "epoch": 0.40274774201755503, "ewc_loss": 0.05126953125, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.246765375137329, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8615180253982544, "num_tokens": 120253553.0, "step": 3166 }, { "epoch": 0.4028749522961455, "ewc_loss": 0.046875, "ewc_loss_diag": 8.153915405273438e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.9268150329589844, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8499775528907776, "num_tokens": 120291438.0, "step": 3167 }, { "epoch": 0.40300216257473603, "ewc_loss": 0.053466796875, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3366448879241943, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.8493117094039917, "num_tokens": 120332005.0, "step": 3168 }, { "epoch": 0.40312937285332656, "ewc_loss": 0.046630859375, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.867485523223877, "learning_rate": 1e-05, "loss": 0.5251, "mean_token_accuracy": 0.8377667665481567, "num_tokens": 120369280.0, "step": 3169 }, { "epoch": 0.40325658313191703, "ewc_loss": 0.054931640625, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.480543613433838, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.855743408203125, "num_tokens": 120402350.0, "step": 3170 }, { "epoch": 0.40338379341050756, "ewc_loss": 0.046875, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.8223440647125244, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8610788583755493, "num_tokens": 120441791.0, "step": 3171 }, { "epoch": 0.4035110036890981, "ewc_loss": 0.06201171875, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.681945323944092, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8686197400093079, "num_tokens": 120476895.0, "step": 3172 }, { "epoch": 0.40363821396768856, "ewc_loss": 0.050537109375, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.883983612060547, "learning_rate": 1e-05, "loss": 0.5073, "mean_token_accuracy": 0.8487625122070312, "num_tokens": 120518016.0, "step": 3173 }, { "epoch": 0.4037654242462791, "ewc_loss": 0.0654296875, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8656110763549805, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8673878908157349, "num_tokens": 120551500.0, "step": 3174 }, { "epoch": 0.4038926345248696, "ewc_loss": 0.05712890625, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.11184024810791, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8627064228057861, "num_tokens": 120588543.0, "step": 3175 }, { "epoch": 0.4040198448034601, "ewc_loss": 0.06689453125, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.865867853164673, "learning_rate": 1e-05, "loss": 0.5354, "mean_token_accuracy": 0.8525627851486206, "num_tokens": 120634394.0, "step": 3176 }, { "epoch": 0.4041470550820506, "ewc_loss": 0.060302734375, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.313929319381714, "learning_rate": 1e-05, "loss": 0.5173, "mean_token_accuracy": 0.8503392934799194, "num_tokens": 120672982.0, "step": 3177 }, { "epoch": 0.40427426536064115, "ewc_loss": 0.06201171875, "ewc_loss_diag": 8.249282836914062e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.7036612033843994, "learning_rate": 1e-05, "loss": 0.5671, "mean_token_accuracy": 0.8321852087974548, "num_tokens": 120712091.0, "step": 3178 }, { "epoch": 0.4044014756392316, "ewc_loss": 0.059814453125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3528451919555664, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8657599687576294, "num_tokens": 120743530.0, "step": 3179 }, { "epoch": 0.40452868591782215, "ewc_loss": 0.056396484375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4958176612854004, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8662457466125488, "num_tokens": 120778884.0, "step": 3180 }, { "epoch": 0.4046558961964127, "ewc_loss": 0.05615234375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.294509172439575, "learning_rate": 1e-05, "loss": 0.5022, "mean_token_accuracy": 0.8499215841293335, "num_tokens": 120815951.0, "step": 3181 }, { "epoch": 0.4047831064750032, "ewc_loss": 0.0537109375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3365554809570312, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.854421854019165, "num_tokens": 120853111.0, "step": 3182 }, { "epoch": 0.4049103167535937, "ewc_loss": 0.053466796875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2262423038482666, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8497111797332764, "num_tokens": 120894041.0, "step": 3183 }, { "epoch": 0.4050375270321842, "ewc_loss": 0.05224609375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2348992824554443, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8594637513160706, "num_tokens": 120930730.0, "step": 3184 }, { "epoch": 0.40516473731077474, "ewc_loss": 0.05029296875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.1755242347717285, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8506420254707336, "num_tokens": 120965044.0, "step": 3185 }, { "epoch": 0.4052919475893652, "ewc_loss": 0.050048828125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.165315628051758, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.8521902561187744, "num_tokens": 120999777.0, "step": 3186 }, { "epoch": 0.40541915786795574, "ewc_loss": 0.049072265625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.139028787612915, "learning_rate": 1e-05, "loss": 0.5394, "mean_token_accuracy": 0.8376448154449463, "num_tokens": 121031910.0, "step": 3187 }, { "epoch": 0.40554636814654627, "ewc_loss": 0.04931640625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.124328136444092, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8593122959136963, "num_tokens": 121069600.0, "step": 3188 }, { "epoch": 0.40567357842513674, "ewc_loss": 0.048095703125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0915615558624268, "learning_rate": 1e-05, "loss": 0.5268, "mean_token_accuracy": 0.8409370183944702, "num_tokens": 121103790.0, "step": 3189 }, { "epoch": 0.40580078870372727, "ewc_loss": 0.048095703125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.1155877113342285, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8481287956237793, "num_tokens": 121135199.0, "step": 3190 }, { "epoch": 0.4059279989823178, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.0468530654907227, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8537706136703491, "num_tokens": 121176040.0, "step": 3191 }, { "epoch": 0.40605520926090827, "ewc_loss": 0.048583984375, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.1191165447235107, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8441011905670166, "num_tokens": 121214675.0, "step": 3192 }, { "epoch": 0.4061824195394988, "ewc_loss": 0.047119140625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0208091735839844, "learning_rate": 1e-05, "loss": 0.5336, "mean_token_accuracy": 0.8396855592727661, "num_tokens": 121252834.0, "step": 3193 }, { "epoch": 0.4063096298180893, "ewc_loss": 0.049072265625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.132781982421875, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8506578207015991, "num_tokens": 121292837.0, "step": 3194 }, { "epoch": 0.4064368400966798, "ewc_loss": 0.047119140625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9787368774414062, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8589920997619629, "num_tokens": 121332353.0, "step": 3195 }, { "epoch": 0.4065640503752703, "ewc_loss": 0.05029296875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.1754603385925293, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.862553596496582, "num_tokens": 121371095.0, "step": 3196 }, { "epoch": 0.40669126065386085, "ewc_loss": 0.04638671875, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.9322524070739746, "learning_rate": 1e-05, "loss": 0.5409, "mean_token_accuracy": 0.8371003866195679, "num_tokens": 121409653.0, "step": 3197 }, { "epoch": 0.4068184709324513, "ewc_loss": 0.0517578125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2523422241210938, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8660286664962769, "num_tokens": 121449309.0, "step": 3198 }, { "epoch": 0.40694568121104185, "ewc_loss": 0.046142578125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.851677656173706, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8609045743942261, "num_tokens": 121495004.0, "step": 3199 }, { "epoch": 0.4070728914896324, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3993587493896484, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8632094264030457, "num_tokens": 121537679.0, "step": 3200 }, { "epoch": 0.40720010176822286, "ewc_loss": 0.045654296875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 3.743171691894531e-05, "grad_norm": 2.7790753841400146, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8535054922103882, "num_tokens": 121579156.0, "step": 3201 }, { "epoch": 0.4073273120468134, "ewc_loss": 0.0595703125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.660259246826172, "learning_rate": 1e-05, "loss": 0.5192, "mean_token_accuracy": 0.8509775400161743, "num_tokens": 121611056.0, "step": 3202 }, { "epoch": 0.4074545223254039, "ewc_loss": 0.047607421875, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.790618658065796, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8647518157958984, "num_tokens": 121642821.0, "step": 3203 }, { "epoch": 0.4075817326039944, "ewc_loss": 0.0673828125, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.970323085784912, "learning_rate": 1e-05, "loss": 0.5165, "mean_token_accuracy": 0.8498329520225525, "num_tokens": 121674021.0, "step": 3204 }, { "epoch": 0.4077089428825849, "ewc_loss": 0.056884765625, "ewc_loss_diag": 8.344650268554688e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.0374677181243896, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8571765422821045, "num_tokens": 121711434.0, "step": 3205 }, { "epoch": 0.40783615316117544, "ewc_loss": 0.0703125, "ewc_loss_diag": 8.392333984375e-05, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.116396427154541, "learning_rate": 1e-05, "loss": 0.5497, "mean_token_accuracy": 0.8400816917419434, "num_tokens": 121755357.0, "step": 3206 }, { "epoch": 0.4079633634397659, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.4175350666046143, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8618569374084473, "num_tokens": 121791703.0, "step": 3207 }, { "epoch": 0.40809057371835644, "ewc_loss": 0.06689453125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.954345703125, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8672788143157959, "num_tokens": 121824940.0, "step": 3208 }, { "epoch": 0.40821778399694697, "ewc_loss": 0.06494140625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.53275465965271, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.857987642288208, "num_tokens": 121860434.0, "step": 3209 }, { "epoch": 0.40834499427553744, "ewc_loss": 0.06201171875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.7330212593078613, "learning_rate": 1e-05, "loss": 0.5331, "mean_token_accuracy": 0.8425695896148682, "num_tokens": 121897536.0, "step": 3210 }, { "epoch": 0.40847220455412797, "ewc_loss": 0.06103515625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4464545249938965, "learning_rate": 1e-05, "loss": 0.5528, "mean_token_accuracy": 0.8384731411933899, "num_tokens": 121936721.0, "step": 3211 }, { "epoch": 0.4085994148327185, "ewc_loss": 0.0576171875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.554643392562866, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8485226631164551, "num_tokens": 121977407.0, "step": 3212 }, { "epoch": 0.40872662511130897, "ewc_loss": 0.05615234375, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.316380739212036, "learning_rate": 1e-05, "loss": 0.551, "mean_token_accuracy": 0.8380252718925476, "num_tokens": 122015068.0, "step": 3213 }, { "epoch": 0.4088538353898995, "ewc_loss": 0.05517578125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4448556900024414, "learning_rate": 1e-05, "loss": 0.4996, "mean_token_accuracy": 0.8516404032707214, "num_tokens": 122058899.0, "step": 3214 }, { "epoch": 0.40898104566849003, "ewc_loss": 0.05224609375, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1868550777435303, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.858257532119751, "num_tokens": 122095407.0, "step": 3215 }, { "epoch": 0.4091082559470805, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.401512384414673, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8597617149353027, "num_tokens": 122134836.0, "step": 3216 }, { "epoch": 0.40923546622567103, "ewc_loss": 0.04931640625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0955913066864014, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8552204966545105, "num_tokens": 122165959.0, "step": 3217 }, { "epoch": 0.40936267650426156, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.421725273132324, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8492753505706787, "num_tokens": 122206223.0, "step": 3218 }, { "epoch": 0.40948988678285203, "ewc_loss": 0.04833984375, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.9884915351867676, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8617740273475647, "num_tokens": 122245416.0, "step": 3219 }, { "epoch": 0.40961709706144256, "ewc_loss": 0.05615234375, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.513046979904175, "learning_rate": 1e-05, "loss": 0.5321, "mean_token_accuracy": 0.8393610119819641, "num_tokens": 122285264.0, "step": 3220 }, { "epoch": 0.4097443073400331, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9350407123565674, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8515770435333252, "num_tokens": 122326956.0, "step": 3221 }, { "epoch": 0.40987151761862356, "ewc_loss": 0.06103515625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6759140491485596, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8617955446243286, "num_tokens": 122362409.0, "step": 3222 }, { "epoch": 0.4099987278972141, "ewc_loss": 0.05029296875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9986064434051514, "learning_rate": 1e-05, "loss": 0.5499, "mean_token_accuracy": 0.8320609331130981, "num_tokens": 122405011.0, "step": 3223 }, { "epoch": 0.4101259381758046, "ewc_loss": 0.06396484375, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7889204025268555, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8596717119216919, "num_tokens": 122445482.0, "step": 3224 }, { "epoch": 0.4102531484543951, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.158026695251465, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8596609830856323, "num_tokens": 122485024.0, "step": 3225 }, { "epoch": 0.4103803587329856, "ewc_loss": 0.06298828125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7514495849609375, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.856482207775116, "num_tokens": 122524766.0, "step": 3226 }, { "epoch": 0.41050756901157615, "ewc_loss": 0.056640625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3047072887420654, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8554712533950806, "num_tokens": 122566664.0, "step": 3227 }, { "epoch": 0.4106347792901666, "ewc_loss": 0.0576171875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5738377571105957, "learning_rate": 1e-05, "loss": 0.5926, "mean_token_accuracy": 0.8208407759666443, "num_tokens": 122606724.0, "step": 3228 }, { "epoch": 0.41076198956875715, "ewc_loss": 0.05712890625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.358534097671509, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8485857844352722, "num_tokens": 122641664.0, "step": 3229 }, { "epoch": 0.4108891998473477, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.330320358276367, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8724729418754578, "num_tokens": 122673370.0, "step": 3230 }, { "epoch": 0.4110164101259382, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.337296724319458, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.853225588798523, "num_tokens": 122711545.0, "step": 3231 }, { "epoch": 0.4111436204045287, "ewc_loss": 0.04931640625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.131805658340454, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.857395589351654, "num_tokens": 122743656.0, "step": 3232 }, { "epoch": 0.4112708306831192, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.3415372371673584, "learning_rate": 1e-05, "loss": 0.5056, "mean_token_accuracy": 0.8482999205589294, "num_tokens": 122778233.0, "step": 3233 }, { "epoch": 0.41139804096170973, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9810540676116943, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8635324835777283, "num_tokens": 122821063.0, "step": 3234 }, { "epoch": 0.4115252512403002, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.440017700195312e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.4032657146453857, "learning_rate": 1e-05, "loss": 0.512, "mean_token_accuracy": 0.850671648979187, "num_tokens": 122855871.0, "step": 3235 }, { "epoch": 0.41165246151889073, "ewc_loss": 0.047119140625, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.90973162651062, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8562161922454834, "num_tokens": 122893990.0, "step": 3236 }, { "epoch": 0.41177967179748126, "ewc_loss": 0.056396484375, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.505997657775879, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8578913807868958, "num_tokens": 122932017.0, "step": 3237 }, { "epoch": 0.41190688207607173, "ewc_loss": 0.047119140625, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.872999668121338, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8601663112640381, "num_tokens": 122975581.0, "step": 3238 }, { "epoch": 0.41203409235466226, "ewc_loss": 0.0615234375, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6859817504882812, "learning_rate": 1e-05, "loss": 0.5032, "mean_token_accuracy": 0.8498870134353638, "num_tokens": 123006051.0, "step": 3239 }, { "epoch": 0.4121613026332528, "ewc_loss": 0.05078125, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.947782516479492, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8507107496261597, "num_tokens": 123049756.0, "step": 3240 }, { "epoch": 0.41228851291184326, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7971887588500977, "learning_rate": 1e-05, "loss": 0.5272, "mean_token_accuracy": 0.8467389345169067, "num_tokens": 123091059.0, "step": 3241 }, { "epoch": 0.4124157231904338, "ewc_loss": 0.054931640625, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1251139640808105, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8616621494293213, "num_tokens": 123130369.0, "step": 3242 }, { "epoch": 0.4125429334690243, "ewc_loss": 0.06396484375, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7769572734832764, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8558737635612488, "num_tokens": 123173507.0, "step": 3243 }, { "epoch": 0.4126701437476148, "ewc_loss": 0.05810546875, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.267679214477539, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8640503883361816, "num_tokens": 123211428.0, "step": 3244 }, { "epoch": 0.4127973540262053, "ewc_loss": 0.060791015625, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.628256320953369, "learning_rate": 1e-05, "loss": 0.529, "mean_token_accuracy": 0.8486646413803101, "num_tokens": 123256115.0, "step": 3245 }, { "epoch": 0.41292456430479585, "ewc_loss": 0.057373046875, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2887818813323975, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.8559567332267761, "num_tokens": 123297319.0, "step": 3246 }, { "epoch": 0.4130517745833863, "ewc_loss": 0.056640625, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4648149013519287, "learning_rate": 1e-05, "loss": 0.5387, "mean_token_accuracy": 0.8394465446472168, "num_tokens": 123341241.0, "step": 3247 }, { "epoch": 0.41317898486197685, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2525787353515625, "learning_rate": 1e-05, "loss": 0.5211, "mean_token_accuracy": 0.8437209129333496, "num_tokens": 123381832.0, "step": 3248 }, { "epoch": 0.4133061951405674, "ewc_loss": 0.0537109375, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.4419305324554443, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8564025163650513, "num_tokens": 123418393.0, "step": 3249 }, { "epoch": 0.41343340541915785, "ewc_loss": 0.052978515625, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1836934089660645, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8645690679550171, "num_tokens": 123449916.0, "step": 3250 }, { "epoch": 0.4135606156977484, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2960305213928223, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8765215873718262, "num_tokens": 123484025.0, "step": 3251 }, { "epoch": 0.4136878259763389, "ewc_loss": 0.050048828125, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0991201400756836, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8590089678764343, "num_tokens": 123522345.0, "step": 3252 }, { "epoch": 0.4138150362549294, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.272387742996216, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8540551066398621, "num_tokens": 123562123.0, "step": 3253 }, { "epoch": 0.4139422465335199, "ewc_loss": 0.048583984375, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0442183017730713, "learning_rate": 1e-05, "loss": 0.5067, "mean_token_accuracy": 0.8478544354438782, "num_tokens": 123606014.0, "step": 3254 }, { "epoch": 0.41406945681211044, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2863316535949707, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8649399876594543, "num_tokens": 123643483.0, "step": 3255 }, { "epoch": 0.4141966670907009, "ewc_loss": 0.047607421875, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.974175453186035, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8578355312347412, "num_tokens": 123689303.0, "step": 3256 }, { "epoch": 0.41432387736929144, "ewc_loss": 0.053955078125, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3593266010284424, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8609236478805542, "num_tokens": 123722388.0, "step": 3257 }, { "epoch": 0.41445108764788197, "ewc_loss": 0.047607421875, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9308085441589355, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8547155857086182, "num_tokens": 123761761.0, "step": 3258 }, { "epoch": 0.41457829792647244, "ewc_loss": 0.055419921875, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.448619842529297, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8776264190673828, "num_tokens": 123799138.0, "step": 3259 }, { "epoch": 0.41470550820506297, "ewc_loss": 0.0478515625, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.8925201892852783, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8614854216575623, "num_tokens": 123840803.0, "step": 3260 }, { "epoch": 0.4148327184836535, "ewc_loss": 0.05908203125, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.598566770553589, "learning_rate": 1e-05, "loss": 0.5194, "mean_token_accuracy": 0.8516936898231506, "num_tokens": 123875705.0, "step": 3261 }, { "epoch": 0.41495992876224397, "ewc_loss": 0.049072265625, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.983311891555786, "learning_rate": 1e-05, "loss": 0.5306, "mean_token_accuracy": 0.8411475419998169, "num_tokens": 123915830.0, "step": 3262 }, { "epoch": 0.4150871390408345, "ewc_loss": 0.061279296875, "ewc_loss_diag": 8.487701416015625e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6522674560546875, "learning_rate": 1e-05, "loss": 0.5252, "mean_token_accuracy": 0.8436740636825562, "num_tokens": 123953855.0, "step": 3263 }, { "epoch": 0.415214349319425, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0764846801757812, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8553907871246338, "num_tokens": 123985697.0, "step": 3264 }, { "epoch": 0.4153415595980155, "ewc_loss": 0.06005859375, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6069834232330322, "learning_rate": 1e-05, "loss": 0.5033, "mean_token_accuracy": 0.8529952764511108, "num_tokens": 124021682.0, "step": 3265 }, { "epoch": 0.415468769876606, "ewc_loss": 0.05517578125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.166682481765747, "learning_rate": 1e-05, "loss": 0.5367, "mean_token_accuracy": 0.8413991928100586, "num_tokens": 124057472.0, "step": 3266 }, { "epoch": 0.41559598015519655, "ewc_loss": 0.056884765625, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4808130264282227, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8568843603134155, "num_tokens": 124100160.0, "step": 3267 }, { "epoch": 0.415723190433787, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.19650936126709, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8534533977508545, "num_tokens": 124138503.0, "step": 3268 }, { "epoch": 0.41585040071237755, "ewc_loss": 0.0537109375, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.326136350631714, "learning_rate": 1e-05, "loss": 0.5379, "mean_token_accuracy": 0.8441194891929626, "num_tokens": 124176082.0, "step": 3269 }, { "epoch": 0.4159776109909681, "ewc_loss": 0.053466796875, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2121450901031494, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8487724661827087, "num_tokens": 124213853.0, "step": 3270 }, { "epoch": 0.41610482126955856, "ewc_loss": 0.05078125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1543450355529785, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8719245195388794, "num_tokens": 124250562.0, "step": 3271 }, { "epoch": 0.4162320315481491, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2397758960723877, "learning_rate": 1e-05, "loss": 0.5364, "mean_token_accuracy": 0.8380831480026245, "num_tokens": 124282321.0, "step": 3272 }, { "epoch": 0.4163592418267396, "ewc_loss": 0.04833984375, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.033341884613037, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8592578172683716, "num_tokens": 124323172.0, "step": 3273 }, { "epoch": 0.4164864521053301, "ewc_loss": 0.052978515625, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.272550582885742, "learning_rate": 1e-05, "loss": 0.4981, "mean_token_accuracy": 0.8535394668579102, "num_tokens": 124361585.0, "step": 3274 }, { "epoch": 0.4166136623839206, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.949829578399658, "learning_rate": 1e-05, "loss": 0.5662, "mean_token_accuracy": 0.832820475101471, "num_tokens": 124398926.0, "step": 3275 }, { "epoch": 0.41674087266251114, "ewc_loss": 0.0537109375, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3327901363372803, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8690234422683716, "num_tokens": 124437914.0, "step": 3276 }, { "epoch": 0.4168680829411016, "ewc_loss": 0.046875, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.876701831817627, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8562608957290649, "num_tokens": 124478850.0, "step": 3277 }, { "epoch": 0.41699529321969214, "ewc_loss": 0.05517578125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4572784900665283, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.869555652141571, "num_tokens": 124519361.0, "step": 3278 }, { "epoch": 0.41712250349828267, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.8467886447906494, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8522834777832031, "num_tokens": 124555616.0, "step": 3279 }, { "epoch": 0.4172497137768732, "ewc_loss": 0.060791015625, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.642360210418701, "learning_rate": 1e-05, "loss": 0.5471, "mean_token_accuracy": 0.8384736776351929, "num_tokens": 124593045.0, "step": 3280 }, { "epoch": 0.41737692405546367, "ewc_loss": 0.04931640625, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.88944673538208, "learning_rate": 1e-05, "loss": 0.5511, "mean_token_accuracy": 0.8327171802520752, "num_tokens": 124631842.0, "step": 3281 }, { "epoch": 0.4175041343340542, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.535385131835938e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.794445276260376, "learning_rate": 1e-05, "loss": 0.5191, "mean_token_accuracy": 0.8505319356918335, "num_tokens": 124674466.0, "step": 3282 }, { "epoch": 0.41763134461264473, "ewc_loss": 0.055419921875, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.108022451400757, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8583074808120728, "num_tokens": 124707893.0, "step": 3283 }, { "epoch": 0.4177585548912352, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7737061977386475, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.852966845035553, "num_tokens": 124746230.0, "step": 3284 }, { "epoch": 0.41788576516982573, "ewc_loss": 0.05859375, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1772968769073486, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.853899359703064, "num_tokens": 124786988.0, "step": 3285 }, { "epoch": 0.41801297544841626, "ewc_loss": 0.0634765625, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6954333782196045, "learning_rate": 1e-05, "loss": 0.527, "mean_token_accuracy": 0.8468402624130249, "num_tokens": 124831061.0, "step": 3286 }, { "epoch": 0.41814018572700673, "ewc_loss": 0.058837890625, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.243095874786377, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8500572443008423, "num_tokens": 124870486.0, "step": 3287 }, { "epoch": 0.41826739600559726, "ewc_loss": 0.059814453125, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5577738285064697, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8673967123031616, "num_tokens": 124909675.0, "step": 3288 }, { "epoch": 0.4183946062841878, "ewc_loss": 0.056884765625, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2337615489959717, "learning_rate": 1e-05, "loss": 0.5316, "mean_token_accuracy": 0.8431726098060608, "num_tokens": 124953854.0, "step": 3289 }, { "epoch": 0.41852181656277826, "ewc_loss": 0.05615234375, "ewc_loss_diag": 8.58306884765625e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.423436403274536, "learning_rate": 1e-05, "loss": 0.5085, "mean_token_accuracy": 0.852412760257721, "num_tokens": 124991117.0, "step": 3290 }, { "epoch": 0.4186490268413688, "ewc_loss": 0.05517578125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1977334022521973, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8526731133460999, "num_tokens": 125027160.0, "step": 3291 }, { "epoch": 0.4187762371199593, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3203818798065186, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8606564402580261, "num_tokens": 125061324.0, "step": 3292 }, { "epoch": 0.4189034473985498, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1420297622680664, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.852519690990448, "num_tokens": 125099616.0, "step": 3293 }, { "epoch": 0.4190306576771403, "ewc_loss": 0.052978515625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.267066240310669, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.850302517414093, "num_tokens": 125132855.0, "step": 3294 }, { "epoch": 0.41915786795573085, "ewc_loss": 0.0498046875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0987019538879395, "learning_rate": 1e-05, "loss": 0.529, "mean_token_accuracy": 0.846721887588501, "num_tokens": 125168040.0, "step": 3295 }, { "epoch": 0.4192850782343213, "ewc_loss": 0.052490234375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.24017071723938, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8595339059829712, "num_tokens": 125198788.0, "step": 3296 }, { "epoch": 0.41941228851291185, "ewc_loss": 0.049072265625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0311737060546875, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8617543578147888, "num_tokens": 125238387.0, "step": 3297 }, { "epoch": 0.4195394987915024, "ewc_loss": 0.052490234375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.257059097290039, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8657364845275879, "num_tokens": 125273666.0, "step": 3298 }, { "epoch": 0.41966670907009285, "ewc_loss": 0.048095703125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9833786487579346, "learning_rate": 1e-05, "loss": 0.5166, "mean_token_accuracy": 0.8449547290802002, "num_tokens": 125315961.0, "step": 3299 }, { "epoch": 0.4197939193486834, "ewc_loss": 0.053955078125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.316441059112549, "learning_rate": 1e-05, "loss": 0.5274, "mean_token_accuracy": 0.8427313566207886, "num_tokens": 125358830.0, "step": 3300 }, { "epoch": 0.4199211296272739, "ewc_loss": 0.047607421875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9249727725982666, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8467094898223877, "num_tokens": 125397497.0, "step": 3301 }, { "epoch": 0.4200483399058644, "ewc_loss": 0.054443359375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.414116144180298, "learning_rate": 1e-05, "loss": 0.5527, "mean_token_accuracy": 0.8343640565872192, "num_tokens": 125437337.0, "step": 3302 }, { "epoch": 0.4201755501844549, "ewc_loss": 0.048095703125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.901916742324829, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8636506795883179, "num_tokens": 125470535.0, "step": 3303 }, { "epoch": 0.42030276046304543, "ewc_loss": 0.057861328125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.533477306365967, "learning_rate": 1e-05, "loss": 0.5375, "mean_token_accuracy": 0.8402475118637085, "num_tokens": 125510692.0, "step": 3304 }, { "epoch": 0.4204299707416359, "ewc_loss": 0.048583984375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.8885562419891357, "learning_rate": 1e-05, "loss": 0.5294, "mean_token_accuracy": 0.8389527797698975, "num_tokens": 125557328.0, "step": 3305 }, { "epoch": 0.42055718102022643, "ewc_loss": 0.06298828125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6900997161865234, "learning_rate": 1e-05, "loss": 0.5277, "mean_token_accuracy": 0.850375771522522, "num_tokens": 125594006.0, "step": 3306 }, { "epoch": 0.42068439129881696, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.9991037845611572, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.8495838642120361, "num_tokens": 125627692.0, "step": 3307 }, { "epoch": 0.42081160157740743, "ewc_loss": 0.06396484375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.744591474533081, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8567598462104797, "num_tokens": 125660218.0, "step": 3308 }, { "epoch": 0.42093881185599796, "ewc_loss": 0.05712890625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.141282796859741, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8639525175094604, "num_tokens": 125695793.0, "step": 3309 }, { "epoch": 0.4210660221345885, "ewc_loss": 0.0634765625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.66737961769104, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8567938804626465, "num_tokens": 125732438.0, "step": 3310 }, { "epoch": 0.42119323241317896, "ewc_loss": 0.057861328125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2767157554626465, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8614957332611084, "num_tokens": 125764919.0, "step": 3311 }, { "epoch": 0.4213204426917695, "ewc_loss": 0.05712890625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4842634201049805, "learning_rate": 1e-05, "loss": 0.5045, "mean_token_accuracy": 0.8517836332321167, "num_tokens": 125797172.0, "step": 3312 }, { "epoch": 0.42144765297036, "ewc_loss": 0.056640625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2517244815826416, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.857369065284729, "num_tokens": 125838144.0, "step": 3313 }, { "epoch": 0.4215748632489505, "ewc_loss": 0.053955078125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2933590412139893, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.865516722202301, "num_tokens": 125873897.0, "step": 3314 }, { "epoch": 0.421702073527541, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.251546859741211, "learning_rate": 1e-05, "loss": 0.5303, "mean_token_accuracy": 0.840326189994812, "num_tokens": 125909773.0, "step": 3315 }, { "epoch": 0.42182928380613155, "ewc_loss": 0.05126953125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1467504501342773, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8573075532913208, "num_tokens": 125947661.0, "step": 3316 }, { "epoch": 0.421956494084722, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2474234104156494, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8587065935134888, "num_tokens": 125979483.0, "step": 3317 }, { "epoch": 0.42208370436331255, "ewc_loss": 0.049072265625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0358781814575195, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8652150630950928, "num_tokens": 126016907.0, "step": 3318 }, { "epoch": 0.4222109146419031, "ewc_loss": 0.053466796875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.269871234893799, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8501538038253784, "num_tokens": 126059475.0, "step": 3319 }, { "epoch": 0.42233812492049355, "ewc_loss": 0.048095703125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9622104167938232, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8719667792320251, "num_tokens": 126097373.0, "step": 3320 }, { "epoch": 0.4224653351990841, "ewc_loss": 0.053955078125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3221919536590576, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8663352727890015, "num_tokens": 126138884.0, "step": 3321 }, { "epoch": 0.4225925454776746, "ewc_loss": 0.047607421875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.904886245727539, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8608980774879456, "num_tokens": 126178453.0, "step": 3322 }, { "epoch": 0.4227197557562651, "ewc_loss": 0.055419921875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4529480934143066, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8492248058319092, "num_tokens": 126208979.0, "step": 3323 }, { "epoch": 0.4228469660348556, "ewc_loss": 0.048095703125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.8517374992370605, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8588940501213074, "num_tokens": 126252195.0, "step": 3324 }, { "epoch": 0.42297417631344614, "ewc_loss": 0.059814453125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.6148085594177246, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8597409129142761, "num_tokens": 126287526.0, "step": 3325 }, { "epoch": 0.4231013865920366, "ewc_loss": 0.04931640625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.8752083778381348, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8659950494766235, "num_tokens": 126325868.0, "step": 3326 }, { "epoch": 0.42322859687062714, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.796609401702881, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.858260452747345, "num_tokens": 126364613.0, "step": 3327 }, { "epoch": 0.42335580714921767, "ewc_loss": 0.05517578125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.0246670246124268, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8671402335166931, "num_tokens": 126404373.0, "step": 3328 }, { "epoch": 0.42348301742780814, "ewc_loss": 0.06591796875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8782124519348145, "learning_rate": 1e-05, "loss": 0.5662, "mean_token_accuracy": 0.8347252011299133, "num_tokens": 126444557.0, "step": 3329 }, { "epoch": 0.42361022770639867, "ewc_loss": 0.06005859375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.251209020614624, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.8498209714889526, "num_tokens": 126479657.0, "step": 3330 }, { "epoch": 0.4237374379849892, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.757127523422241, "learning_rate": 1e-05, "loss": 0.5312, "mean_token_accuracy": 0.8470689058303833, "num_tokens": 126518004.0, "step": 3331 }, { "epoch": 0.4238646482635797, "ewc_loss": 0.060302734375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.321704864501953, "learning_rate": 1e-05, "loss": 0.5302, "mean_token_accuracy": 0.8480090498924255, "num_tokens": 126557854.0, "step": 3332 }, { "epoch": 0.4239918585421702, "ewc_loss": 0.059326171875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5834126472473145, "learning_rate": 1e-05, "loss": 0.5338, "mean_token_accuracy": 0.8509756326675415, "num_tokens": 126593657.0, "step": 3333 }, { "epoch": 0.4241190688207607, "ewc_loss": 0.0576171875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.291720390319824, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8779702186584473, "num_tokens": 126635913.0, "step": 3334 }, { "epoch": 0.42424627909935125, "ewc_loss": 0.055419921875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.434997081756592, "learning_rate": 1e-05, "loss": 0.5343, "mean_token_accuracy": 0.841492772102356, "num_tokens": 126668703.0, "step": 3335 }, { "epoch": 0.4243734893779417, "ewc_loss": 0.05517578125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2329041957855225, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8692057728767395, "num_tokens": 126707952.0, "step": 3336 }, { "epoch": 0.42450069965653225, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.31856632232666, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8542624711990356, "num_tokens": 126748756.0, "step": 3337 }, { "epoch": 0.4246279099351228, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1679763793945312, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8657728433609009, "num_tokens": 126784780.0, "step": 3338 }, { "epoch": 0.42475512021371326, "ewc_loss": 0.052490234375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.24960994720459, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8599279522895813, "num_tokens": 126827371.0, "step": 3339 }, { "epoch": 0.4248823304923038, "ewc_loss": 0.049560546875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.1219489574432373, "learning_rate": 1e-05, "loss": 0.5041, "mean_token_accuracy": 0.8497172594070435, "num_tokens": 126862357.0, "step": 3340 }, { "epoch": 0.4250095407708943, "ewc_loss": 0.05224609375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.20894455909729, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8755963444709778, "num_tokens": 126902433.0, "step": 3341 }, { "epoch": 0.4251367510494848, "ewc_loss": 0.048828125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0577189922332764, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8487000465393066, "num_tokens": 126946958.0, "step": 3342 }, { "epoch": 0.4252639613280753, "ewc_loss": 0.0517578125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.2302303314208984, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.865134596824646, "num_tokens": 126980267.0, "step": 3343 }, { "epoch": 0.42539117160666584, "ewc_loss": 0.047607421875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9895741939544678, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.87403404712677, "num_tokens": 127018215.0, "step": 3344 }, { "epoch": 0.4255183818852563, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2872183322906494, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8461155891418457, "num_tokens": 127055844.0, "step": 3345 }, { "epoch": 0.42564559216384684, "ewc_loss": 0.046875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.916041851043701, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8593851327896118, "num_tokens": 127099373.0, "step": 3346 }, { "epoch": 0.42577280244243737, "ewc_loss": 0.054443359375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3930017948150635, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8571339249610901, "num_tokens": 127143755.0, "step": 3347 }, { "epoch": 0.42590001272102784, "ewc_loss": 0.046875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.8630928993225098, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.864545464515686, "num_tokens": 127180543.0, "step": 3348 }, { "epoch": 0.42602722299961837, "ewc_loss": 0.0576171875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5595438480377197, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8500915169715881, "num_tokens": 127225203.0, "step": 3349 }, { "epoch": 0.4261544332782089, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.8770055770874023, "learning_rate": 1e-05, "loss": 0.5475, "mean_token_accuracy": 0.8363480567932129, "num_tokens": 127260949.0, "step": 3350 }, { "epoch": 0.4262816435567994, "ewc_loss": 0.0625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.733935832977295, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.871253252029419, "num_tokens": 127297487.0, "step": 3351 }, { "epoch": 0.4264088538353899, "ewc_loss": 0.052734375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.9706063270568848, "learning_rate": 1e-05, "loss": 0.5366, "mean_token_accuracy": 0.8402582406997681, "num_tokens": 127342612.0, "step": 3352 }, { "epoch": 0.42653606411398043, "ewc_loss": 0.0654296875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8300230503082275, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8647861480712891, "num_tokens": 127383915.0, "step": 3353 }, { "epoch": 0.4266632743925709, "ewc_loss": 0.0576171875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.175774574279785, "learning_rate": 1e-05, "loss": 0.5229, "mean_token_accuracy": 0.8488584160804749, "num_tokens": 127421564.0, "step": 3354 }, { "epoch": 0.42679048467116143, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.760227918624878, "learning_rate": 1e-05, "loss": 0.5515, "mean_token_accuracy": 0.8375934362411499, "num_tokens": 127463750.0, "step": 3355 }, { "epoch": 0.42691769494975196, "ewc_loss": 0.060302734375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3079521656036377, "learning_rate": 1e-05, "loss": 0.5278, "mean_token_accuracy": 0.8482779264450073, "num_tokens": 127502793.0, "step": 3356 }, { "epoch": 0.42704490522834243, "ewc_loss": 0.05908203125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5698626041412354, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8532047271728516, "num_tokens": 127542650.0, "step": 3357 }, { "epoch": 0.42717211550693296, "ewc_loss": 0.058349609375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.299984931945801, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8655637502670288, "num_tokens": 127584753.0, "step": 3358 }, { "epoch": 0.4272993257855235, "ewc_loss": 0.05517578125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3857858180999756, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8475573062896729, "num_tokens": 127624427.0, "step": 3359 }, { "epoch": 0.42742653606411396, "ewc_loss": 0.05517578125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.260082244873047, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8637549877166748, "num_tokens": 127665636.0, "step": 3360 }, { "epoch": 0.4275537463427045, "ewc_loss": 0.05224609375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2371575832366943, "learning_rate": 1e-05, "loss": 0.5082, "mean_token_accuracy": 0.8493456840515137, "num_tokens": 127706066.0, "step": 3361 }, { "epoch": 0.427680956621295, "ewc_loss": 0.05322265625, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2175683975219727, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8660311698913574, "num_tokens": 127741680.0, "step": 3362 }, { "epoch": 0.4278081668998855, "ewc_loss": 0.05078125, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.151089668273926, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8494385480880737, "num_tokens": 127781147.0, "step": 3363 }, { "epoch": 0.427935377178476, "ewc_loss": 0.05078125, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1891345977783203, "learning_rate": 1e-05, "loss": 0.5524, "mean_token_accuracy": 0.8349283933639526, "num_tokens": 127823787.0, "step": 3364 }, { "epoch": 0.42806258745706655, "ewc_loss": 0.048828125, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0988898277282715, "learning_rate": 1e-05, "loss": 0.5546, "mean_token_accuracy": 0.8339953422546387, "num_tokens": 127862239.0, "step": 3365 }, { "epoch": 0.428189797735657, "ewc_loss": 0.05029296875, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1601569652557373, "learning_rate": 1e-05, "loss": 0.5384, "mean_token_accuracy": 0.8428521752357483, "num_tokens": 127904376.0, "step": 3366 }, { "epoch": 0.42831700801424755, "ewc_loss": 0.048583984375, "ewc_loss_diag": 8.726119995117188e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0586462020874023, "learning_rate": 1e-05, "loss": 0.5222, "mean_token_accuracy": 0.8429387211799622, "num_tokens": 127944132.0, "step": 3367 }, { "epoch": 0.4284442182928381, "ewc_loss": 0.050048828125, "ewc_loss_diag": 8.726119995117188e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.1577653884887695, "learning_rate": 1e-05, "loss": 0.5163, "mean_token_accuracy": 0.8442294597625732, "num_tokens": 127983512.0, "step": 3368 }, { "epoch": 0.42857142857142855, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0246055126190186, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8626470565795898, "num_tokens": 128018864.0, "step": 3369 }, { "epoch": 0.4286986388500191, "ewc_loss": 0.05029296875, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.163039445877075, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8700780868530273, "num_tokens": 128056586.0, "step": 3370 }, { "epoch": 0.4288258491286096, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0036914348602295, "learning_rate": 1e-05, "loss": 0.4891, "mean_token_accuracy": 0.8507342338562012, "num_tokens": 128091131.0, "step": 3371 }, { "epoch": 0.4289530594072001, "ewc_loss": 0.05078125, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.20162034034729, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8482484817504883, "num_tokens": 128125455.0, "step": 3372 }, { "epoch": 0.4290802696857906, "ewc_loss": 0.047119140625, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.9463069438934326, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8506768941879272, "num_tokens": 128164357.0, "step": 3373 }, { "epoch": 0.42920747996438113, "ewc_loss": 0.052490234375, "ewc_loss_diag": 8.630752563476562e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2646727561950684, "learning_rate": 1e-05, "loss": 0.5072, "mean_token_accuracy": 0.8502302765846252, "num_tokens": 128203590.0, "step": 3374 }, { "epoch": 0.4293346902429716, "ewc_loss": 0.046875, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.8900578022003174, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8506022095680237, "num_tokens": 128238819.0, "step": 3375 }, { "epoch": 0.42946190052156213, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.389244794845581, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8711072206497192, "num_tokens": 128273879.0, "step": 3376 }, { "epoch": 0.42958911080015266, "ewc_loss": 0.04638671875, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.8115522861480713, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.856820821762085, "num_tokens": 128309688.0, "step": 3377 }, { "epoch": 0.42971632107874314, "ewc_loss": 0.05908203125, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.6060564517974854, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8559795022010803, "num_tokens": 128341515.0, "step": 3378 }, { "epoch": 0.42984353135733366, "ewc_loss": 0.04833984375, "ewc_loss_diag": 8.678436279296875e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.7869861125946045, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.860907256603241, "num_tokens": 128382676.0, "step": 3379 }, { "epoch": 0.4299707416359242, "ewc_loss": 0.06640625, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8886518478393555, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8723883032798767, "num_tokens": 128417512.0, "step": 3380 }, { "epoch": 0.4300979519145147, "ewc_loss": 0.054931640625, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 2.957062244415283, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8592613935470581, "num_tokens": 128457695.0, "step": 3381 }, { "epoch": 0.4302251621931052, "ewc_loss": 0.07080078125, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.114264965057373, "learning_rate": 1e-05, "loss": 0.5011, "mean_token_accuracy": 0.8536657691001892, "num_tokens": 128487459.0, "step": 3382 }, { "epoch": 0.4303523724716957, "ewc_loss": 0.06396484375, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3588671684265137, "learning_rate": 1e-05, "loss": 0.5606, "mean_token_accuracy": 0.8308160305023193, "num_tokens": 128526823.0, "step": 3383 }, { "epoch": 0.43047958275028625, "ewc_loss": 0.068359375, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9716968536376953, "learning_rate": 1e-05, "loss": 0.5033, "mean_token_accuracy": 0.851556658744812, "num_tokens": 128566841.0, "step": 3384 }, { "epoch": 0.4306067930288767, "ewc_loss": 0.0654296875, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5155625343322754, "learning_rate": 1e-05, "loss": 0.5246, "mean_token_accuracy": 0.844673216342926, "num_tokens": 128608147.0, "step": 3385 }, { "epoch": 0.43073400330746725, "ewc_loss": 0.0634765625, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7335362434387207, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.858844518661499, "num_tokens": 128648277.0, "step": 3386 }, { "epoch": 0.4308612135860578, "ewc_loss": 0.0625, "ewc_loss_diag": 8.7738037109375e-05, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.458040237426758, "learning_rate": 1e-05, "loss": 0.5556, "mean_token_accuracy": 0.8377494812011719, "num_tokens": 128682547.0, "step": 3387 }, { "epoch": 0.43098842386464825, "ewc_loss": 0.059326171875, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5587191581726074, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8584904670715332, "num_tokens": 128715103.0, "step": 3388 }, { "epoch": 0.4311156341432388, "ewc_loss": 0.057373046875, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3433148860931396, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8500627279281616, "num_tokens": 128748924.0, "step": 3389 }, { "epoch": 0.4312428444218293, "ewc_loss": 0.055908203125, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4189465045928955, "learning_rate": 1e-05, "loss": 0.5066, "mean_token_accuracy": 0.851464033126831, "num_tokens": 128786879.0, "step": 3390 }, { "epoch": 0.4313700547004198, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2297770977020264, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8512247800827026, "num_tokens": 128822121.0, "step": 3391 }, { "epoch": 0.4314972649790103, "ewc_loss": 0.053955078125, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3613381385803223, "learning_rate": 1e-05, "loss": 0.5525, "mean_token_accuracy": 0.8346641063690186, "num_tokens": 128858594.0, "step": 3392 }, { "epoch": 0.43162447525760084, "ewc_loss": 0.05029296875, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1192519664764404, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8535295724868774, "num_tokens": 128896147.0, "step": 3393 }, { "epoch": 0.4317516855361913, "ewc_loss": 0.054443359375, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3583240509033203, "learning_rate": 1e-05, "loss": 0.5247, "mean_token_accuracy": 0.8421846032142639, "num_tokens": 128930055.0, "step": 3394 }, { "epoch": 0.43187889581478184, "ewc_loss": 0.049072265625, "ewc_loss_diag": 8.821487426757812e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0233981609344482, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8520738482475281, "num_tokens": 128968577.0, "step": 3395 }, { "epoch": 0.43200610609337237, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.4084436893463135, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8487805128097534, "num_tokens": 129007074.0, "step": 3396 }, { "epoch": 0.43213331637196284, "ewc_loss": 0.0478515625, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.950099229812622, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.8455359935760498, "num_tokens": 129044722.0, "step": 3397 }, { "epoch": 0.43226052665055337, "ewc_loss": 0.0576171875, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.537086248397827, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8545850515365601, "num_tokens": 129081567.0, "step": 3398 }, { "epoch": 0.4323877369291439, "ewc_loss": 0.04833984375, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.918672561645508, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8501673340797424, "num_tokens": 129119354.0, "step": 3399 }, { "epoch": 0.43251494720773437, "ewc_loss": 0.0625, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6992478370666504, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8661139607429504, "num_tokens": 129157871.0, "step": 3400 }, { "epoch": 0.4326421574863249, "ewc_loss": 0.05224609375, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.98828387260437, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8638237714767456, "num_tokens": 129200164.0, "step": 3401 }, { "epoch": 0.4327693677649154, "ewc_loss": 0.0654296875, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.817603349685669, "learning_rate": 1e-05, "loss": 0.5142, "mean_token_accuracy": 0.850863516330719, "num_tokens": 129239382.0, "step": 3402 }, { "epoch": 0.4328965780435059, "ewc_loss": 0.056640625, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1920082569122314, "learning_rate": 1e-05, "loss": 0.5127, "mean_token_accuracy": 0.8478503823280334, "num_tokens": 129273201.0, "step": 3403 }, { "epoch": 0.4330237883220964, "ewc_loss": 0.064453125, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7438604831695557, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8633533716201782, "num_tokens": 129309636.0, "step": 3404 }, { "epoch": 0.43315099860068695, "ewc_loss": 0.05908203125, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3301503658294678, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.863782525062561, "num_tokens": 129345803.0, "step": 3405 }, { "epoch": 0.4332782088792774, "ewc_loss": 0.0595703125, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5373127460479736, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8600766062736511, "num_tokens": 129384955.0, "step": 3406 }, { "epoch": 0.43340541915786795, "ewc_loss": 0.0576171875, "ewc_loss_diag": 8.916854858398438e-05, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3293402194976807, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8711354732513428, "num_tokens": 129425129.0, "step": 3407 }, { "epoch": 0.4335326294364585, "ewc_loss": 0.05419921875, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3337762355804443, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8509732484817505, "num_tokens": 129465931.0, "step": 3408 }, { "epoch": 0.43365983971504896, "ewc_loss": 0.0556640625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3137288093566895, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.863516628742218, "num_tokens": 129497296.0, "step": 3409 }, { "epoch": 0.4337870499936395, "ewc_loss": 0.051513671875, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.14628005027771, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8588968515396118, "num_tokens": 129539037.0, "step": 3410 }, { "epoch": 0.43391426027223, "ewc_loss": 0.054443359375, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3000240325927734, "learning_rate": 1e-05, "loss": 0.5131, "mean_token_accuracy": 0.8491652607917786, "num_tokens": 129581036.0, "step": 3411 }, { "epoch": 0.4340414705508205, "ewc_loss": 0.049072265625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.041828155517578, "learning_rate": 1e-05, "loss": 0.5049, "mean_token_accuracy": 0.8487215042114258, "num_tokens": 129617515.0, "step": 3412 }, { "epoch": 0.434168680829411, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3170742988586426, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8742854595184326, "num_tokens": 129659348.0, "step": 3413 }, { "epoch": 0.43429589110800154, "ewc_loss": 0.0478515625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9771265983581543, "learning_rate": 1e-05, "loss": 0.5431, "mean_token_accuracy": 0.8347077369689941, "num_tokens": 129695000.0, "step": 3414 }, { "epoch": 0.434423101386592, "ewc_loss": 0.054931640625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.376765012741089, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8486981391906738, "num_tokens": 129734125.0, "step": 3415 }, { "epoch": 0.43455031166518254, "ewc_loss": 0.04736328125, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.8982222080230713, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8531261682510376, "num_tokens": 129774799.0, "step": 3416 }, { "epoch": 0.43467752194377307, "ewc_loss": 0.056884765625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4875223636627197, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8599890470504761, "num_tokens": 129814390.0, "step": 3417 }, { "epoch": 0.43480473222236354, "ewc_loss": 0.0478515625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.879680871963501, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8619865775108337, "num_tokens": 129848172.0, "step": 3418 }, { "epoch": 0.43493194250095407, "ewc_loss": 0.0615234375, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6576926708221436, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.869491457939148, "num_tokens": 129878728.0, "step": 3419 }, { "epoch": 0.4350591527795446, "ewc_loss": 0.051025390625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9251644611358643, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8525403738021851, "num_tokens": 129916843.0, "step": 3420 }, { "epoch": 0.4351863630581351, "ewc_loss": 0.06494140625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8005471229553223, "learning_rate": 1e-05, "loss": 0.5337, "mean_token_accuracy": 0.8430037498474121, "num_tokens": 129954614.0, "step": 3421 }, { "epoch": 0.4353135733367256, "ewc_loss": 0.056640625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.096855640411377, "learning_rate": 1e-05, "loss": 0.5646, "mean_token_accuracy": 0.8361007571220398, "num_tokens": 129993084.0, "step": 3422 }, { "epoch": 0.43544078361531613, "ewc_loss": 0.0654296875, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.799121379852295, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8617430925369263, "num_tokens": 130028313.0, "step": 3423 }, { "epoch": 0.4355679938939066, "ewc_loss": 0.05908203125, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2401554584503174, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.851688027381897, "num_tokens": 130066722.0, "step": 3424 }, { "epoch": 0.43569520417249713, "ewc_loss": 0.06298828125, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6796746253967285, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8508822917938232, "num_tokens": 130108628.0, "step": 3425 }, { "epoch": 0.43582241445108766, "ewc_loss": 0.058349609375, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2826178073883057, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8678799271583557, "num_tokens": 130145770.0, "step": 3426 }, { "epoch": 0.43594962472967813, "ewc_loss": 0.05908203125, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.507840871810913, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8541735410690308, "num_tokens": 130191539.0, "step": 3427 }, { "epoch": 0.43607683500826866, "ewc_loss": 0.056640625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2704951763153076, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8660291433334351, "num_tokens": 130229358.0, "step": 3428 }, { "epoch": 0.4362040452868592, "ewc_loss": 0.054931640625, "ewc_loss_diag": 9.012222290039062e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.349217414855957, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8563316464424133, "num_tokens": 130272173.0, "step": 3429 }, { "epoch": 0.4363312555654497, "ewc_loss": 0.0546875, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.222876787185669, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8527571558952332, "num_tokens": 130313135.0, "step": 3430 }, { "epoch": 0.4364584658440402, "ewc_loss": 0.052978515625, "ewc_loss_diag": 8.96453857421875e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2438278198242188, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8640696406364441, "num_tokens": 130350595.0, "step": 3431 }, { "epoch": 0.4365856761226307, "ewc_loss": 0.0517578125, "ewc_loss_diag": 9.059906005859375e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1695311069488525, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8619102239608765, "num_tokens": 130390194.0, "step": 3432 }, { "epoch": 0.43671288640122125, "ewc_loss": 0.05224609375, "ewc_loss_diag": 9.059906005859375e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1866209506988525, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8563765287399292, "num_tokens": 130424768.0, "step": 3433 }, { "epoch": 0.4368400966798117, "ewc_loss": 0.05029296875, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.1191489696502686, "learning_rate": 1e-05, "loss": 0.5367, "mean_token_accuracy": 0.8408451676368713, "num_tokens": 130469457.0, "step": 3434 }, { "epoch": 0.43696730695840225, "ewc_loss": 0.05126953125, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1571292877197266, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.869935154914856, "num_tokens": 130506250.0, "step": 3435 }, { "epoch": 0.4370945172369928, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.10001277923584, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8551660180091858, "num_tokens": 130538605.0, "step": 3436 }, { "epoch": 0.43722172751558325, "ewc_loss": 0.050048828125, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.1337881088256836, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.869861900806427, "num_tokens": 130577172.0, "step": 3437 }, { "epoch": 0.4373489377941738, "ewc_loss": 0.04833984375, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.076437473297119, "learning_rate": 1e-05, "loss": 0.4891, "mean_token_accuracy": 0.852787435054779, "num_tokens": 130613067.0, "step": 3438 }, { "epoch": 0.4374761480727643, "ewc_loss": 0.0498046875, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1368937492370605, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8722367286682129, "num_tokens": 130649176.0, "step": 3439 }, { "epoch": 0.4376033583513548, "ewc_loss": 0.04833984375, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.040153980255127, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8582065105438232, "num_tokens": 130685904.0, "step": 3440 }, { "epoch": 0.4377305686299453, "ewc_loss": 0.05029296875, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.1469810009002686, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.868570864200592, "num_tokens": 130727502.0, "step": 3441 }, { "epoch": 0.43785777890853583, "ewc_loss": 0.0478515625, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 3.0097708702087402, "learning_rate": 1e-05, "loss": 0.5039, "mean_token_accuracy": 0.8465316295623779, "num_tokens": 130765038.0, "step": 3442 }, { "epoch": 0.4379849891871263, "ewc_loss": 0.051025390625, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.174128532409668, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8616293668746948, "num_tokens": 130810349.0, "step": 3443 }, { "epoch": 0.43811219946571683, "ewc_loss": 0.0478515625, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9861648082733154, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.8474488258361816, "num_tokens": 130843383.0, "step": 3444 }, { "epoch": 0.43823940974430736, "ewc_loss": 0.052001953125, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.2309317588806152, "learning_rate": 1e-05, "loss": 0.5036, "mean_token_accuracy": 0.8507495522499084, "num_tokens": 130880399.0, "step": 3445 }, { "epoch": 0.43836662002289783, "ewc_loss": 0.047119140625, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.924870014190674, "learning_rate": 1e-05, "loss": 0.5402, "mean_token_accuracy": 0.8403007984161377, "num_tokens": 130915537.0, "step": 3446 }, { "epoch": 0.43849383030148836, "ewc_loss": 0.0537109375, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.3382625579833984, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8647297620773315, "num_tokens": 130948560.0, "step": 3447 }, { "epoch": 0.4386210405800789, "ewc_loss": 0.046875, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 3.790855407714844e-05, "grad_norm": 2.8452415466308594, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8644204139709473, "num_tokens": 130982405.0, "step": 3448 }, { "epoch": 0.43874825085866936, "ewc_loss": 0.056640625, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.5001468658447266, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8609088063240051, "num_tokens": 131022673.0, "step": 3449 }, { "epoch": 0.4388754611372599, "ewc_loss": 0.04736328125, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.7956507205963135, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8526703119277954, "num_tokens": 131054647.0, "step": 3450 }, { "epoch": 0.4390026714158504, "ewc_loss": 0.064453125, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.762653350830078, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8612124919891357, "num_tokens": 131091616.0, "step": 3451 }, { "epoch": 0.4391298816944409, "ewc_loss": 0.05126953125, "ewc_loss_diag": 9.107589721679688e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.8579251766204834, "learning_rate": 1e-05, "loss": 0.5618, "mean_token_accuracy": 0.8297497034072876, "num_tokens": 131129663.0, "step": 3452 }, { "epoch": 0.4392570919730314, "ewc_loss": 0.0703125, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.022037506103516, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8596647381782532, "num_tokens": 131170858.0, "step": 3453 }, { "epoch": 0.43938430225162195, "ewc_loss": 0.06103515625, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.1580917835235596, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8539183139801025, "num_tokens": 131210892.0, "step": 3454 }, { "epoch": 0.4395115125302124, "ewc_loss": 0.0703125, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.080862998962402, "learning_rate": 1e-05, "loss": 0.5156, "mean_token_accuracy": 0.8535324335098267, "num_tokens": 131246616.0, "step": 3455 }, { "epoch": 0.43963872280880295, "ewc_loss": 0.06640625, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.4679203033447266, "learning_rate": 1e-05, "loss": 0.4891, "mean_token_accuracy": 0.8581130504608154, "num_tokens": 131286515.0, "step": 3456 }, { "epoch": 0.4397659330873935, "ewc_loss": 0.06689453125, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.878267288208008, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.854404091835022, "num_tokens": 131321715.0, "step": 3457 }, { "epoch": 0.43989314336598395, "ewc_loss": 0.06494140625, "ewc_loss_diag": 9.1552734375e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.507995128631592, "learning_rate": 1e-05, "loss": 0.5419, "mean_token_accuracy": 0.8465137481689453, "num_tokens": 131359235.0, "step": 3458 }, { "epoch": 0.4400203536445745, "ewc_loss": 0.06201171875, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.671177864074707, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8747230768203735, "num_tokens": 131390551.0, "step": 3459 }, { "epoch": 0.440147563923165, "ewc_loss": 0.060302734375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4035356044769287, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8583009243011475, "num_tokens": 131424107.0, "step": 3460 }, { "epoch": 0.4402747742017555, "ewc_loss": 0.057861328125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.517568349838257, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.862876832485199, "num_tokens": 131459797.0, "step": 3461 }, { "epoch": 0.440401984480346, "ewc_loss": 0.056396484375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2824156284332275, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8540114760398865, "num_tokens": 131496877.0, "step": 3462 }, { "epoch": 0.44052919475893654, "ewc_loss": 0.05517578125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.426332712173462, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8644246459007263, "num_tokens": 131528958.0, "step": 3463 }, { "epoch": 0.440656405037527, "ewc_loss": 0.052978515625, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1710970401763916, "learning_rate": 1e-05, "loss": 0.5471, "mean_token_accuracy": 0.8367503881454468, "num_tokens": 131572157.0, "step": 3464 }, { "epoch": 0.44078361531611754, "ewc_loss": 0.05517578125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.4016501903533936, "learning_rate": 1e-05, "loss": 0.5056, "mean_token_accuracy": 0.8493788242340088, "num_tokens": 131613855.0, "step": 3465 }, { "epoch": 0.44091082559470807, "ewc_loss": 0.050537109375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0587174892425537, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.854769229888916, "num_tokens": 131654056.0, "step": 3466 }, { "epoch": 0.44103803587329854, "ewc_loss": 0.055419921875, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4460465908050537, "learning_rate": 1e-05, "loss": 0.5488, "mean_token_accuracy": 0.8341095447540283, "num_tokens": 131694865.0, "step": 3467 }, { "epoch": 0.44116524615188907, "ewc_loss": 0.04931640625, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9861581325531006, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8546825051307678, "num_tokens": 131732883.0, "step": 3468 }, { "epoch": 0.4412924564304796, "ewc_loss": 0.058349609375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5572493076324463, "learning_rate": 1e-05, "loss": 0.5148, "mean_token_accuracy": 0.8489850759506226, "num_tokens": 131770384.0, "step": 3469 }, { "epoch": 0.44141966670907007, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9753761291503906, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8515757918357849, "num_tokens": 131801611.0, "step": 3470 }, { "epoch": 0.4415468769876606, "ewc_loss": 0.062255859375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.685758352279663, "learning_rate": 1e-05, "loss": 0.6006, "mean_token_accuracy": 0.8313149809837341, "num_tokens": 131842522.0, "step": 3471 }, { "epoch": 0.4416740872662511, "ewc_loss": 0.052001953125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0260276794433594, "learning_rate": 1e-05, "loss": 0.5524, "mean_token_accuracy": 0.8409191966056824, "num_tokens": 131880041.0, "step": 3472 }, { "epoch": 0.4418012975448416, "ewc_loss": 0.064453125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7760045528411865, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8584314584732056, "num_tokens": 131917682.0, "step": 3473 }, { "epoch": 0.4419285078234321, "ewc_loss": 0.055908203125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1680212020874023, "learning_rate": 1e-05, "loss": 0.4969, "mean_token_accuracy": 0.8533037900924683, "num_tokens": 131958149.0, "step": 3474 }, { "epoch": 0.44205571810202265, "ewc_loss": 0.064453125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.727560043334961, "learning_rate": 1e-05, "loss": 0.5334, "mean_token_accuracy": 0.8422589302062988, "num_tokens": 131996648.0, "step": 3475 }, { "epoch": 0.4421829283806131, "ewc_loss": 0.058349609375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2837579250335693, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8570536375045776, "num_tokens": 132036524.0, "step": 3476 }, { "epoch": 0.44231013865920366, "ewc_loss": 0.05859375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.554405450820923, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8572978377342224, "num_tokens": 132077474.0, "step": 3477 }, { "epoch": 0.4424373489377942, "ewc_loss": 0.057373046875, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.317413091659546, "learning_rate": 1e-05, "loss": 0.5301, "mean_token_accuracy": 0.8395226001739502, "num_tokens": 132117153.0, "step": 3478 }, { "epoch": 0.44256455921638466, "ewc_loss": 0.0546875, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3416666984558105, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8686838746070862, "num_tokens": 132156832.0, "step": 3479 }, { "epoch": 0.4426917694949752, "ewc_loss": 0.055908203125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3070220947265625, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8564066886901855, "num_tokens": 132189245.0, "step": 3480 }, { "epoch": 0.4428189797735657, "ewc_loss": 0.052001953125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.163489818572998, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8550111055374146, "num_tokens": 132230572.0, "step": 3481 }, { "epoch": 0.44294619005215624, "ewc_loss": 0.054443359375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.298555850982666, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8545463681221008, "num_tokens": 132271850.0, "step": 3482 }, { "epoch": 0.4430734003307467, "ewc_loss": 0.049072265625, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0363216400146484, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8586432933807373, "num_tokens": 132312461.0, "step": 3483 }, { "epoch": 0.44320061060933724, "ewc_loss": 0.053955078125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.3330318927764893, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8616273403167725, "num_tokens": 132350628.0, "step": 3484 }, { "epoch": 0.44332782088792777, "ewc_loss": 0.0478515625, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.9468142986297607, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8531883955001831, "num_tokens": 132394040.0, "step": 3485 }, { "epoch": 0.44345503116651824, "ewc_loss": 0.05517578125, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.4159719944000244, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8551669120788574, "num_tokens": 132435727.0, "step": 3486 }, { "epoch": 0.44358224144510877, "ewc_loss": 0.047607421875, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.884840250015259, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8520529866218567, "num_tokens": 132475769.0, "step": 3487 }, { "epoch": 0.4437094517236993, "ewc_loss": 0.058349609375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5582778453826904, "learning_rate": 1e-05, "loss": 0.5011, "mean_token_accuracy": 0.8524266481399536, "num_tokens": 132520944.0, "step": 3488 }, { "epoch": 0.4438366620022898, "ewc_loss": 0.04833984375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.8803882598876953, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8526536822319031, "num_tokens": 132558138.0, "step": 3489 }, { "epoch": 0.4439638722808803, "ewc_loss": 0.06396484375, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7507059574127197, "learning_rate": 1e-05, "loss": 0.5694, "mean_token_accuracy": 0.8304365873336792, "num_tokens": 132594937.0, "step": 3490 }, { "epoch": 0.44409108255947083, "ewc_loss": 0.053466796875, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.996257781982422, "learning_rate": 1e-05, "loss": 0.5457, "mean_token_accuracy": 0.8369897603988647, "num_tokens": 132629168.0, "step": 3491 }, { "epoch": 0.4442182928380613, "ewc_loss": 0.06591796875, "ewc_loss_diag": 9.202957153320312e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.85387921333313, "learning_rate": 1e-05, "loss": 0.5918, "mean_token_accuracy": 0.8271454572677612, "num_tokens": 132673239.0, "step": 3492 }, { "epoch": 0.44434550311665183, "ewc_loss": 0.058349609375, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.1736788749694824, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8616899847984314, "num_tokens": 132707429.0, "step": 3493 }, { "epoch": 0.44447271339524236, "ewc_loss": 0.0654296875, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.803222894668579, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8565303683280945, "num_tokens": 132745592.0, "step": 3494 }, { "epoch": 0.44459992367383283, "ewc_loss": 0.060302734375, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.309206962585449, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8636490106582642, "num_tokens": 132778260.0, "step": 3495 }, { "epoch": 0.44472713395242336, "ewc_loss": 0.062255859375, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6314210891723633, "learning_rate": 1e-05, "loss": 0.3977, "mean_token_accuracy": 0.8835473656654358, "num_tokens": 132813156.0, "step": 3496 }, { "epoch": 0.4448543442310139, "ewc_loss": 0.058837890625, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3366661071777344, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8530921936035156, "num_tokens": 132847654.0, "step": 3497 }, { "epoch": 0.44498155450960436, "ewc_loss": 0.056640625, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4432380199432373, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8619298934936523, "num_tokens": 132889258.0, "step": 3498 }, { "epoch": 0.4451087647881949, "ewc_loss": 0.056884765625, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3040904998779297, "learning_rate": 1e-05, "loss": 0.5574, "mean_token_accuracy": 0.8338796496391296, "num_tokens": 132919962.0, "step": 3499 }, { "epoch": 0.4452359750667854, "ewc_loss": 0.05419921875, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.292450189590454, "learning_rate": 1e-05, "loss": 0.523, "mean_token_accuracy": 0.8481153249740601, "num_tokens": 132957333.0, "step": 3500 }, { "epoch": 0.4453631853453759, "ewc_loss": 0.052978515625, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2410976886749268, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8544546365737915, "num_tokens": 132990645.0, "step": 3501 }, { "epoch": 0.4454903956239664, "ewc_loss": 0.0517578125, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.188054323196411, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8656409978866577, "num_tokens": 133026234.0, "step": 3502 }, { "epoch": 0.44561760590255695, "ewc_loss": 0.052734375, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.194403648376465, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8574065566062927, "num_tokens": 133062833.0, "step": 3503 }, { "epoch": 0.4457448161811474, "ewc_loss": 0.051025390625, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1366043090820312, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.868412971496582, "num_tokens": 133097868.0, "step": 3504 }, { "epoch": 0.44587202645973795, "ewc_loss": 0.051025390625, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1449127197265625, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8620272278785706, "num_tokens": 133142062.0, "step": 3505 }, { "epoch": 0.4459992367383285, "ewc_loss": 0.050048828125, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1106417179107666, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8509119749069214, "num_tokens": 133181415.0, "step": 3506 }, { "epoch": 0.44612644701691895, "ewc_loss": 0.0498046875, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.1221559047698975, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8571203947067261, "num_tokens": 133217849.0, "step": 3507 }, { "epoch": 0.4462536572955095, "ewc_loss": 0.0498046875, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0962066650390625, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.8539353609085083, "num_tokens": 133257639.0, "step": 3508 }, { "epoch": 0.4463808675741, "ewc_loss": 0.0498046875, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.1020686626434326, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8502030372619629, "num_tokens": 133299692.0, "step": 3509 }, { "epoch": 0.4465080778526905, "ewc_loss": 0.04931640625, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.092355489730835, "learning_rate": 1e-05, "loss": 0.5349, "mean_token_accuracy": 0.8394138813018799, "num_tokens": 133339315.0, "step": 3510 }, { "epoch": 0.446635288131281, "ewc_loss": 0.04931640625, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.093364715576172, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8598606586456299, "num_tokens": 133378477.0, "step": 3511 }, { "epoch": 0.44676249840987153, "ewc_loss": 0.049072265625, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0937745571136475, "learning_rate": 1e-05, "loss": 0.5644, "mean_token_accuracy": 0.8310026526451111, "num_tokens": 133416041.0, "step": 3512 }, { "epoch": 0.446889708688462, "ewc_loss": 0.049072265625, "ewc_loss_diag": 9.250640869140625e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.078173875808716, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.859126091003418, "num_tokens": 133458201.0, "step": 3513 }, { "epoch": 0.44701691896705253, "ewc_loss": 0.049560546875, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0941596031188965, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.8575249910354614, "num_tokens": 133495645.0, "step": 3514 }, { "epoch": 0.44714412924564306, "ewc_loss": 0.048583984375, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0562000274658203, "learning_rate": 1e-05, "loss": 0.5073, "mean_token_accuracy": 0.8497415781021118, "num_tokens": 133535385.0, "step": 3515 }, { "epoch": 0.44727133952423354, "ewc_loss": 0.050048828125, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.101834774017334, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8602859973907471, "num_tokens": 133575093.0, "step": 3516 }, { "epoch": 0.44739854980282406, "ewc_loss": 0.048583984375, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0451807975769043, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8465721607208252, "num_tokens": 133610447.0, "step": 3517 }, { "epoch": 0.4475257600814146, "ewc_loss": 0.050048828125, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.1207964420318604, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.863568902015686, "num_tokens": 133643506.0, "step": 3518 }, { "epoch": 0.44765297036000506, "ewc_loss": 0.048583984375, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9997522830963135, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8767780661582947, "num_tokens": 133683081.0, "step": 3519 }, { "epoch": 0.4477801806385956, "ewc_loss": 0.05078125, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.162931203842163, "learning_rate": 1e-05, "loss": 0.5168, "mean_token_accuracy": 0.8462448716163635, "num_tokens": 133721836.0, "step": 3520 }, { "epoch": 0.4479073909171861, "ewc_loss": 0.048095703125, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9669716358184814, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8606826066970825, "num_tokens": 133756243.0, "step": 3521 }, { "epoch": 0.4480346011957766, "ewc_loss": 0.052001953125, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.2168824672698975, "learning_rate": 1e-05, "loss": 0.5195, "mean_token_accuracy": 0.8433557152748108, "num_tokens": 133807126.0, "step": 3522 }, { "epoch": 0.4481618114743671, "ewc_loss": 0.047607421875, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9210731983184814, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.852103054523468, "num_tokens": 133842036.0, "step": 3523 }, { "epoch": 0.44828902175295765, "ewc_loss": 0.05419921875, "ewc_loss_diag": 9.298324584960938e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.333444833755493, "learning_rate": 1e-05, "loss": 0.5375, "mean_token_accuracy": 0.8395883440971375, "num_tokens": 133881948.0, "step": 3524 }, { "epoch": 0.4484162320315481, "ewc_loss": 0.04736328125, "ewc_loss_diag": 9.34600830078125e-05, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.8317697048187256, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8624180555343628, "num_tokens": 133917765.0, "step": 3525 }, { "epoch": 0.44854344231013865, "ewc_loss": 0.0576171875, "ewc_loss_diag": 9.34600830078125e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5186562538146973, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8523058891296387, "num_tokens": 133959199.0, "step": 3526 }, { "epoch": 0.4486706525887292, "ewc_loss": 0.047607421875, "ewc_loss_diag": 9.34600830078125e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.787788152694702, "learning_rate": 1e-05, "loss": 0.5191, "mean_token_accuracy": 0.8438202142715454, "num_tokens": 133991484.0, "step": 3527 }, { "epoch": 0.44879786286731965, "ewc_loss": 0.06494140625, "ewc_loss_diag": 9.34600830078125e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.805832624435425, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.86509770154953, "num_tokens": 134025817.0, "step": 3528 }, { "epoch": 0.4489250731459102, "ewc_loss": 0.0517578125, "ewc_loss_diag": 9.393692016601562e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.8244574069976807, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8550910949707031, "num_tokens": 134063956.0, "step": 3529 }, { "epoch": 0.4490522834245007, "ewc_loss": 0.072265625, "ewc_loss_diag": 9.393692016601562e-05, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.150183200836182, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8538632392883301, "num_tokens": 134101959.0, "step": 3530 }, { "epoch": 0.44917949370309124, "ewc_loss": 0.0634765625, "ewc_loss_diag": 9.393692016601562e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.2254865169525146, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.8493419289588928, "num_tokens": 134139961.0, "step": 3531 }, { "epoch": 0.4493067039816817, "ewc_loss": 0.07275390625, "ewc_loss_diag": 9.393692016601562e-05, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.189967155456543, "learning_rate": 1e-05, "loss": 0.5286, "mean_token_accuracy": 0.8486438393592834, "num_tokens": 134180368.0, "step": 3532 }, { "epoch": 0.44943391426027224, "ewc_loss": 0.07177734375, "ewc_loss_diag": 9.393692016601562e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6010804176330566, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8657045364379883, "num_tokens": 134216842.0, "step": 3533 }, { "epoch": 0.44956112453886277, "ewc_loss": 0.06689453125, "ewc_loss_diag": 9.393692016601562e-05, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.917983293533325, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.8595764636993408, "num_tokens": 134254089.0, "step": 3534 }, { "epoch": 0.44968833481745324, "ewc_loss": 0.06689453125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.5947964191436768, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8623627424240112, "num_tokens": 134289559.0, "step": 3535 }, { "epoch": 0.44981554509604377, "ewc_loss": 0.06298828125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.7052035331726074, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.8561625480651855, "num_tokens": 134323029.0, "step": 3536 }, { "epoch": 0.4499427553746343, "ewc_loss": 0.061767578125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4531216621398926, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8529295325279236, "num_tokens": 134362657.0, "step": 3537 }, { "epoch": 0.45006996565322477, "ewc_loss": 0.059814453125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5453169345855713, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8589984178543091, "num_tokens": 134403555.0, "step": 3538 }, { "epoch": 0.4501971759318153, "ewc_loss": 0.056640625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.309190273284912, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8586927652359009, "num_tokens": 134438780.0, "step": 3539 }, { "epoch": 0.4503243862104058, "ewc_loss": 0.05712890625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.470261335372925, "learning_rate": 1e-05, "loss": 0.509, "mean_token_accuracy": 0.8505117893218994, "num_tokens": 134477331.0, "step": 3540 }, { "epoch": 0.4504515964889963, "ewc_loss": 0.053466796875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.182816982269287, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8682776689529419, "num_tokens": 134513898.0, "step": 3541 }, { "epoch": 0.4505788067675868, "ewc_loss": 0.05615234375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.458544969558716, "learning_rate": 1e-05, "loss": 0.5537, "mean_token_accuracy": 0.8401250243186951, "num_tokens": 134548507.0, "step": 3542 }, { "epoch": 0.45070601704617735, "ewc_loss": 0.051025390625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.068765878677368, "learning_rate": 1e-05, "loss": 0.5083, "mean_token_accuracy": 0.8497726321220398, "num_tokens": 134590243.0, "step": 3543 }, { "epoch": 0.4508332273247678, "ewc_loss": 0.057861328125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5248119831085205, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8575579524040222, "num_tokens": 134624693.0, "step": 3544 }, { "epoch": 0.45096043760335836, "ewc_loss": 0.049560546875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.985163927078247, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8763209581375122, "num_tokens": 134661365.0, "step": 3545 }, { "epoch": 0.4510876478819489, "ewc_loss": 0.061279296875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.670595407485962, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8596210479736328, "num_tokens": 134698218.0, "step": 3546 }, { "epoch": 0.45121485816053936, "ewc_loss": 0.052001953125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0146567821502686, "learning_rate": 1e-05, "loss": 0.5068, "mean_token_accuracy": 0.8493856191635132, "num_tokens": 134736011.0, "step": 3547 }, { "epoch": 0.4513420684391299, "ewc_loss": 0.06494140625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.796884059906006, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.859198808670044, "num_tokens": 134774457.0, "step": 3548 }, { "epoch": 0.4514692787177204, "ewc_loss": 0.055419921875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.155972719192505, "learning_rate": 1e-05, "loss": 0.5511, "mean_token_accuracy": 0.833572506904602, "num_tokens": 134813788.0, "step": 3549 }, { "epoch": 0.4515964889963109, "ewc_loss": 0.06494140625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.8013672828674316, "learning_rate": 1e-05, "loss": 0.5296, "mean_token_accuracy": 0.8436117172241211, "num_tokens": 134851916.0, "step": 3550 }, { "epoch": 0.4517236992749014, "ewc_loss": 0.05859375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3069469928741455, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8618545532226562, "num_tokens": 134892655.0, "step": 3551 }, { "epoch": 0.45185090955349194, "ewc_loss": 0.061279296875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6316120624542236, "learning_rate": 1e-05, "loss": 0.4975, "mean_token_accuracy": 0.8533693552017212, "num_tokens": 134931391.0, "step": 3552 }, { "epoch": 0.4519781198320824, "ewc_loss": 0.05908203125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3663833141326904, "learning_rate": 1e-05, "loss": 0.5084, "mean_token_accuracy": 0.8515921831130981, "num_tokens": 134966285.0, "step": 3553 }, { "epoch": 0.45210533011067294, "ewc_loss": 0.055419921875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.4158453941345215, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.856887936592102, "num_tokens": 135003893.0, "step": 3554 }, { "epoch": 0.45223254038926347, "ewc_loss": 0.056640625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3310067653656006, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.856493353843689, "num_tokens": 135043245.0, "step": 3555 }, { "epoch": 0.45235975066785394, "ewc_loss": 0.052978515625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2306952476501465, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8530793786048889, "num_tokens": 135082867.0, "step": 3556 }, { "epoch": 0.45248696094644447, "ewc_loss": 0.054443359375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2930309772491455, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8769767880439758, "num_tokens": 135121153.0, "step": 3557 }, { "epoch": 0.452614171225035, "ewc_loss": 0.0498046875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.1147947311401367, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8521636128425598, "num_tokens": 135154221.0, "step": 3558 }, { "epoch": 0.4527413815036255, "ewc_loss": 0.053955078125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.285522222518921, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8577697277069092, "num_tokens": 135196065.0, "step": 3559 }, { "epoch": 0.452868591782216, "ewc_loss": 0.048583984375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.025108575820923, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8650944232940674, "num_tokens": 135233692.0, "step": 3560 }, { "epoch": 0.45299580206080653, "ewc_loss": 0.05419921875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.314899444580078, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8717886209487915, "num_tokens": 135267538.0, "step": 3561 }, { "epoch": 0.453123012339397, "ewc_loss": 0.0478515625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.9453163146972656, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8640127182006836, "num_tokens": 135303946.0, "step": 3562 }, { "epoch": 0.45325022261798753, "ewc_loss": 0.055419921875, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.4027867317199707, "learning_rate": 1e-05, "loss": 0.5142, "mean_token_accuracy": 0.8461612462997437, "num_tokens": 135340940.0, "step": 3563 }, { "epoch": 0.45337743289657806, "ewc_loss": 0.0478515625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.8916873931884766, "learning_rate": 1e-05, "loss": 0.5146, "mean_token_accuracy": 0.8460670113563538, "num_tokens": 135377472.0, "step": 3564 }, { "epoch": 0.45350464317516853, "ewc_loss": 0.057861328125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5444586277008057, "learning_rate": 1e-05, "loss": 0.5437, "mean_token_accuracy": 0.8364969491958618, "num_tokens": 135411396.0, "step": 3565 }, { "epoch": 0.45363185345375906, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.873608112335205, "learning_rate": 1e-05, "loss": 0.5085, "mean_token_accuracy": 0.8460990786552429, "num_tokens": 135445323.0, "step": 3566 }, { "epoch": 0.4537590637323496, "ewc_loss": 0.06396484375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7117059230804443, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8600010871887207, "num_tokens": 135487364.0, "step": 3567 }, { "epoch": 0.45388627401094006, "ewc_loss": 0.05224609375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.9382693767547607, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8566757440567017, "num_tokens": 135529879.0, "step": 3568 }, { "epoch": 0.4540134842895306, "ewc_loss": 0.0673828125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.866492986679077, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.861016571521759, "num_tokens": 135568812.0, "step": 3569 }, { "epoch": 0.4541406945681211, "ewc_loss": 0.058837890625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.148726224899292, "learning_rate": 1e-05, "loss": 0.5052, "mean_token_accuracy": 0.8514124155044556, "num_tokens": 135608096.0, "step": 3570 }, { "epoch": 0.4542679048467116, "ewc_loss": 0.0673828125, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8631951808929443, "learning_rate": 1e-05, "loss": 0.5865, "mean_token_accuracy": 0.8298287391662598, "num_tokens": 135645796.0, "step": 3571 }, { "epoch": 0.4543951151253021, "ewc_loss": 0.062255859375, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3329100608825684, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8533191084861755, "num_tokens": 135686535.0, "step": 3572 }, { "epoch": 0.45452232540389265, "ewc_loss": 0.0634765625, "ewc_loss_diag": 9.489059448242188e-05, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.688764810562134, "learning_rate": 1e-05, "loss": 0.4966, "mean_token_accuracy": 0.8560770153999329, "num_tokens": 135719555.0, "step": 3573 }, { "epoch": 0.4546495356824831, "ewc_loss": 0.060791015625, "ewc_loss_diag": 9.5367431640625e-05, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3609931468963623, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8671931028366089, "num_tokens": 135756498.0, "step": 3574 }, { "epoch": 0.45477674596107365, "ewc_loss": 0.058837890625, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4824509620666504, "learning_rate": 1e-05, "loss": 0.5545, "mean_token_accuracy": 0.8366800546646118, "num_tokens": 135797438.0, "step": 3575 }, { "epoch": 0.4549039562396642, "ewc_loss": 0.05810546875, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3408076763153076, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8514341115951538, "num_tokens": 135836085.0, "step": 3576 }, { "epoch": 0.45503116651825465, "ewc_loss": 0.054931640625, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3030166625976562, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8609238266944885, "num_tokens": 135872272.0, "step": 3577 }, { "epoch": 0.4551583767968452, "ewc_loss": 0.055419921875, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.282898187637329, "learning_rate": 1e-05, "loss": 0.5525, "mean_token_accuracy": 0.837396502494812, "num_tokens": 135908667.0, "step": 3578 }, { "epoch": 0.4552855870754357, "ewc_loss": 0.052490234375, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1995434761047363, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8659205436706543, "num_tokens": 135939205.0, "step": 3579 }, { "epoch": 0.4554127973540262, "ewc_loss": 0.053466796875, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.227039337158203, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8591628074645996, "num_tokens": 135977084.0, "step": 3580 }, { "epoch": 0.4555400076326167, "ewc_loss": 0.051025390625, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.11771559715271, "learning_rate": 1e-05, "loss": 0.5033, "mean_token_accuracy": 0.8480165004730225, "num_tokens": 136012712.0, "step": 3581 }, { "epoch": 0.45566721791120723, "ewc_loss": 0.052490234375, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1958580017089844, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8611615896224976, "num_tokens": 136050253.0, "step": 3582 }, { "epoch": 0.45579442818979776, "ewc_loss": 0.0498046875, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0674214363098145, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8554421663284302, "num_tokens": 136088072.0, "step": 3583 }, { "epoch": 0.45592163846838824, "ewc_loss": 0.052001953125, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.203599691390991, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8506602048873901, "num_tokens": 136122434.0, "step": 3584 }, { "epoch": 0.45604884874697876, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0112552642822266, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8634635806083679, "num_tokens": 136164489.0, "step": 3585 }, { "epoch": 0.4561760590255693, "ewc_loss": 0.053466796875, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2481625080108643, "learning_rate": 1e-05, "loss": 0.4945, "mean_token_accuracy": 0.8521949052810669, "num_tokens": 136198534.0, "step": 3586 }, { "epoch": 0.45630326930415976, "ewc_loss": 0.04833984375, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.938890218734741, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8569918870925903, "num_tokens": 136242261.0, "step": 3587 }, { "epoch": 0.4564304795827503, "ewc_loss": 0.054931640625, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3466451168060303, "learning_rate": 1e-05, "loss": 0.5234, "mean_token_accuracy": 0.8467904329299927, "num_tokens": 136277106.0, "step": 3588 }, { "epoch": 0.4565576898613408, "ewc_loss": 0.0478515625, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.874372720718384, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8583313226699829, "num_tokens": 136315872.0, "step": 3589 }, { "epoch": 0.4566849001399313, "ewc_loss": 0.05712890625, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.6126646995544434, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8560743927955627, "num_tokens": 136359513.0, "step": 3590 }, { "epoch": 0.4568121104185218, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.820971727371216, "learning_rate": 1e-05, "loss": 0.5092, "mean_token_accuracy": 0.8498004674911499, "num_tokens": 136398575.0, "step": 3591 }, { "epoch": 0.45693932069711235, "ewc_loss": 0.0654296875, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.812778949737549, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8663833737373352, "num_tokens": 136439062.0, "step": 3592 }, { "epoch": 0.4570665309757028, "ewc_loss": 0.053955078125, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 2.939955472946167, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8548024296760559, "num_tokens": 136475044.0, "step": 3593 }, { "epoch": 0.45719374125429335, "ewc_loss": 0.0712890625, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.052437782287598, "learning_rate": 1e-05, "loss": 0.5443, "mean_token_accuracy": 0.8448713421821594, "num_tokens": 136513674.0, "step": 3594 }, { "epoch": 0.4573209515328839, "ewc_loss": 0.06298828125, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.269963502883911, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8600250482559204, "num_tokens": 136551218.0, "step": 3595 }, { "epoch": 0.45744816181147435, "ewc_loss": 0.07080078125, "ewc_loss_diag": 9.584426879882812e-05, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9753007888793945, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8694989085197449, "num_tokens": 136589088.0, "step": 3596 }, { "epoch": 0.4575753720900649, "ewc_loss": 0.06591796875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.474971294403076, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.857062816619873, "num_tokens": 136625269.0, "step": 3597 }, { "epoch": 0.4577025823686554, "ewc_loss": 0.0654296875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.764495611190796, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8602661490440369, "num_tokens": 136660212.0, "step": 3598 }, { "epoch": 0.4578297926472459, "ewc_loss": 0.06396484375, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.44478702545166, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.84897381067276, "num_tokens": 136703372.0, "step": 3599 }, { "epoch": 0.4579570029258364, "ewc_loss": 0.061279296875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.572807788848877, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8503443002700806, "num_tokens": 136741241.0, "step": 3600 }, { "epoch": 0.45808421320442694, "ewc_loss": 0.0595703125, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3457581996917725, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.855584442615509, "num_tokens": 136780329.0, "step": 3601 }, { "epoch": 0.4582114234830174, "ewc_loss": 0.05712890625, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4532930850982666, "learning_rate": 1e-05, "loss": 0.5385, "mean_token_accuracy": 0.8406386375427246, "num_tokens": 136812567.0, "step": 3602 }, { "epoch": 0.45833863376160794, "ewc_loss": 0.0556640625, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2436892986297607, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8478190898895264, "num_tokens": 136854370.0, "step": 3603 }, { "epoch": 0.45846584404019847, "ewc_loss": 0.0556640625, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.37138295173645, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.8536677360534668, "num_tokens": 136890536.0, "step": 3604 }, { "epoch": 0.45859305431878894, "ewc_loss": 0.05322265625, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.144832134246826, "learning_rate": 1e-05, "loss": 0.575, "mean_token_accuracy": 0.8294895887374878, "num_tokens": 136929376.0, "step": 3605 }, { "epoch": 0.45872026459737947, "ewc_loss": 0.055419921875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.342351198196411, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8614609241485596, "num_tokens": 136967857.0, "step": 3606 }, { "epoch": 0.45884747487597, "ewc_loss": 0.05029296875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0570688247680664, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8549826145172119, "num_tokens": 137007853.0, "step": 3607 }, { "epoch": 0.45897468515456047, "ewc_loss": 0.055419921875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3749053478240967, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8578304052352905, "num_tokens": 137048172.0, "step": 3608 }, { "epoch": 0.459101895433151, "ewc_loss": 0.0498046875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9852850437164307, "learning_rate": 1e-05, "loss": 0.5362, "mean_token_accuracy": 0.8384990692138672, "num_tokens": 137083598.0, "step": 3609 }, { "epoch": 0.4592291057117415, "ewc_loss": 0.057373046875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4874825477600098, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8446629047393799, "num_tokens": 137117626.0, "step": 3610 }, { "epoch": 0.459356315990332, "ewc_loss": 0.04931640625, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9286692142486572, "learning_rate": 1e-05, "loss": 0.5157, "mean_token_accuracy": 0.8444580435752869, "num_tokens": 137161536.0, "step": 3611 }, { "epoch": 0.4594835262689225, "ewc_loss": 0.0615234375, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6325137615203857, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8637244701385498, "num_tokens": 137195932.0, "step": 3612 }, { "epoch": 0.45961073654751305, "ewc_loss": 0.050537109375, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.965207099914551, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8645531535148621, "num_tokens": 137229263.0, "step": 3613 }, { "epoch": 0.4597379468261035, "ewc_loss": 0.0654296875, "ewc_loss_diag": 9.679794311523438e-05, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7684121131896973, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8558273315429688, "num_tokens": 137265250.0, "step": 3614 }, { "epoch": 0.45986515710469406, "ewc_loss": 0.055908203125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.089911937713623, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8674015998840332, "num_tokens": 137307612.0, "step": 3615 }, { "epoch": 0.4599923673832846, "ewc_loss": 0.0654296875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.803311824798584, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8688355684280396, "num_tokens": 137336075.0, "step": 3616 }, { "epoch": 0.46011957766187506, "ewc_loss": 0.0595703125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.274698495864868, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8547604084014893, "num_tokens": 137374571.0, "step": 3617 }, { "epoch": 0.4602467879404656, "ewc_loss": 0.06298828125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6487655639648438, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8482716083526611, "num_tokens": 137409275.0, "step": 3618 }, { "epoch": 0.4603739982190561, "ewc_loss": 0.06005859375, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3305282592773438, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8575918674468994, "num_tokens": 137447665.0, "step": 3619 }, { "epoch": 0.4605012084976466, "ewc_loss": 0.056396484375, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4314301013946533, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8644959330558777, "num_tokens": 137484256.0, "step": 3620 }, { "epoch": 0.4606284187762371, "ewc_loss": 0.057373046875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.313387870788574, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.858488917350769, "num_tokens": 137523868.0, "step": 3621 }, { "epoch": 0.46075562905482764, "ewc_loss": 0.0537109375, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2431440353393555, "learning_rate": 1e-05, "loss": 0.5565, "mean_token_accuracy": 0.8313143253326416, "num_tokens": 137564707.0, "step": 3622 }, { "epoch": 0.4608828393334181, "ewc_loss": 0.0556640625, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.302321672439575, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.862994909286499, "num_tokens": 137600347.0, "step": 3623 }, { "epoch": 0.46101004961200864, "ewc_loss": 0.05126953125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0996901988983154, "learning_rate": 1e-05, "loss": 0.5062, "mean_token_accuracy": 0.850228488445282, "num_tokens": 137642874.0, "step": 3624 }, { "epoch": 0.46113725989059917, "ewc_loss": 0.0546875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.309685707092285, "learning_rate": 1e-05, "loss": 0.5348, "mean_token_accuracy": 0.8483972549438477, "num_tokens": 137685253.0, "step": 3625 }, { "epoch": 0.46126447016918964, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 3.0050132274627686, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8731704354286194, "num_tokens": 137722180.0, "step": 3626 }, { "epoch": 0.4613916804477802, "ewc_loss": 0.055419921875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3721365928649902, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8621970415115356, "num_tokens": 137756665.0, "step": 3627 }, { "epoch": 0.4615188907263707, "ewc_loss": 0.048583984375, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.931185483932495, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8626879453659058, "num_tokens": 137795928.0, "step": 3628 }, { "epoch": 0.4616461010049612, "ewc_loss": 0.056884765625, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.476853609085083, "learning_rate": 1e-05, "loss": 0.5225, "mean_token_accuracy": 0.8418629169464111, "num_tokens": 137833695.0, "step": 3629 }, { "epoch": 0.4617733112835517, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.8838067054748535, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8595297932624817, "num_tokens": 137871894.0, "step": 3630 }, { "epoch": 0.46190052156214223, "ewc_loss": 0.062255859375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6365950107574463, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8498002290725708, "num_tokens": 137907818.0, "step": 3631 }, { "epoch": 0.46202773184073276, "ewc_loss": 0.050537109375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.919147491455078, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8691968321800232, "num_tokens": 137943506.0, "step": 3632 }, { "epoch": 0.46215494211932323, "ewc_loss": 0.0654296875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8094797134399414, "learning_rate": 1e-05, "loss": 0.542, "mean_token_accuracy": 0.8486857414245605, "num_tokens": 137975113.0, "step": 3633 }, { "epoch": 0.46228215239791376, "ewc_loss": 0.056396484375, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.0901849269866943, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8617078065872192, "num_tokens": 138006503.0, "step": 3634 }, { "epoch": 0.4624093626765043, "ewc_loss": 0.06591796875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8204410076141357, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8652885556221008, "num_tokens": 138046667.0, "step": 3635 }, { "epoch": 0.46253657295509476, "ewc_loss": 0.059814453125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.228703260421753, "learning_rate": 1e-05, "loss": 0.5349, "mean_token_accuracy": 0.8438511490821838, "num_tokens": 138084821.0, "step": 3636 }, { "epoch": 0.4626637832336853, "ewc_loss": 0.064453125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.716383934020996, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8546000719070435, "num_tokens": 138123994.0, "step": 3637 }, { "epoch": 0.4627909935122758, "ewc_loss": 0.059814453125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2989325523376465, "learning_rate": 1e-05, "loss": 0.5083, "mean_token_accuracy": 0.8522237539291382, "num_tokens": 138161225.0, "step": 3638 }, { "epoch": 0.4629182037908663, "ewc_loss": 0.060791015625, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5646541118621826, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8638259172439575, "num_tokens": 138197441.0, "step": 3639 }, { "epoch": 0.4630454140694568, "ewc_loss": 0.05810546875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.290743112564087, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8595575094223022, "num_tokens": 138229663.0, "step": 3640 }, { "epoch": 0.46317262434804735, "ewc_loss": 0.05615234375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4038519859313965, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.8533492088317871, "num_tokens": 138267955.0, "step": 3641 }, { "epoch": 0.4632998346266378, "ewc_loss": 0.0556640625, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2354161739349365, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8672572374343872, "num_tokens": 138303968.0, "step": 3642 }, { "epoch": 0.46342704490522835, "ewc_loss": 0.053955078125, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.288792610168457, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8575162887573242, "num_tokens": 138344667.0, "step": 3643 }, { "epoch": 0.4635542551838189, "ewc_loss": 0.053466796875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.16937518119812, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.867713451385498, "num_tokens": 138387513.0, "step": 3644 }, { "epoch": 0.46368146546240935, "ewc_loss": 0.0537109375, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.225877046585083, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8744195699691772, "num_tokens": 138428070.0, "step": 3645 }, { "epoch": 0.4638086757409999, "ewc_loss": 0.051513671875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1296939849853516, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8594762682914734, "num_tokens": 138462303.0, "step": 3646 }, { "epoch": 0.4639358860195904, "ewc_loss": 0.05322265625, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.20953369140625, "learning_rate": 1e-05, "loss": 0.5553, "mean_token_accuracy": 0.8371426463127136, "num_tokens": 138502099.0, "step": 3647 }, { "epoch": 0.4640630962981809, "ewc_loss": 0.05029296875, "ewc_loss_diag": 9.72747802734375e-05, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0903422832489014, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8491916656494141, "num_tokens": 138539400.0, "step": 3648 }, { "epoch": 0.4641903065767714, "ewc_loss": 0.05322265625, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.213124990463257, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8564997315406799, "num_tokens": 138573082.0, "step": 3649 }, { "epoch": 0.46431751685536193, "ewc_loss": 0.049072265625, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.0234851837158203, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8621451258659363, "num_tokens": 138613693.0, "step": 3650 }, { "epoch": 0.4644447271339524, "ewc_loss": 0.0537109375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2361035346984863, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8708126544952393, "num_tokens": 138655138.0, "step": 3651 }, { "epoch": 0.46457193741254293, "ewc_loss": 0.048583984375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.98952317237854, "learning_rate": 1e-05, "loss": 0.5767, "mean_token_accuracy": 0.8263580203056335, "num_tokens": 138697117.0, "step": 3652 }, { "epoch": 0.46469914769113346, "ewc_loss": 0.054931640625, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3027074337005615, "learning_rate": 1e-05, "loss": 0.4997, "mean_token_accuracy": 0.8537274599075317, "num_tokens": 138732723.0, "step": 3653 }, { "epoch": 0.46482635796972394, "ewc_loss": 0.04833984375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.923243522644043, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.858723521232605, "num_tokens": 138767680.0, "step": 3654 }, { "epoch": 0.46495356824831446, "ewc_loss": 0.0556640625, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.4139482975006104, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8639936447143555, "num_tokens": 138803119.0, "step": 3655 }, { "epoch": 0.465080778526905, "ewc_loss": 0.04833984375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.8746252059936523, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.863324761390686, "num_tokens": 138845032.0, "step": 3656 }, { "epoch": 0.46520798880549546, "ewc_loss": 0.0595703125, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5728185176849365, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.8623762130737305, "num_tokens": 138886521.0, "step": 3657 }, { "epoch": 0.465335199084086, "ewc_loss": 0.049560546875, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.857450246810913, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8569232225418091, "num_tokens": 138927586.0, "step": 3658 }, { "epoch": 0.4654624093626765, "ewc_loss": 0.0654296875, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.780186891555786, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8516935110092163, "num_tokens": 138967382.0, "step": 3659 }, { "epoch": 0.465589619641267, "ewc_loss": 0.054443359375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.9914941787719727, "learning_rate": 1e-05, "loss": 0.5275, "mean_token_accuracy": 0.8410726189613342, "num_tokens": 139003498.0, "step": 3660 }, { "epoch": 0.4657168299198575, "ewc_loss": 0.068359375, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8906750679016113, "learning_rate": 1e-05, "loss": 0.5266, "mean_token_accuracy": 0.8459200859069824, "num_tokens": 139046724.0, "step": 3661 }, { "epoch": 0.46584404019844805, "ewc_loss": 0.060546875, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.1935970783233643, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8542686700820923, "num_tokens": 139091513.0, "step": 3662 }, { "epoch": 0.4659712504770385, "ewc_loss": 0.06640625, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.839813470840454, "learning_rate": 1e-05, "loss": 0.57, "mean_token_accuracy": 0.8332771062850952, "num_tokens": 139130082.0, "step": 3663 }, { "epoch": 0.46609846075562905, "ewc_loss": 0.0625, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.342303991317749, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8585518598556519, "num_tokens": 139172974.0, "step": 3664 }, { "epoch": 0.4662256710342196, "ewc_loss": 0.06298828125, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6522700786590576, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.864983856678009, "num_tokens": 139212164.0, "step": 3665 }, { "epoch": 0.46635288131281005, "ewc_loss": 0.060546875, "ewc_loss_diag": 9.775161743164062e-05, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3543858528137207, "learning_rate": 1e-05, "loss": 0.578, "mean_token_accuracy": 0.827834963798523, "num_tokens": 139252960.0, "step": 3666 }, { "epoch": 0.4664800915914006, "ewc_loss": 0.05810546875, "ewc_loss_diag": 9.822845458984375e-05, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.478187322616577, "learning_rate": 1e-05, "loss": 0.5283, "mean_token_accuracy": 0.8482770919799805, "num_tokens": 139287078.0, "step": 3667 }, { "epoch": 0.4666073018699911, "ewc_loss": 0.057373046875, "ewc_loss_diag": 9.822845458984375e-05, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3308165073394775, "learning_rate": 1e-05, "loss": 0.5539, "mean_token_accuracy": 0.845413863658905, "num_tokens": 139321545.0, "step": 3668 }, { "epoch": 0.4667345121485816, "ewc_loss": 0.0556640625, "ewc_loss_diag": 9.822845458984375e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3092260360717773, "learning_rate": 1e-05, "loss": 0.5393, "mean_token_accuracy": 0.8400692343711853, "num_tokens": 139360829.0, "step": 3669 }, { "epoch": 0.4668617224271721, "ewc_loss": 0.0556640625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2417407035827637, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8580749034881592, "num_tokens": 139396347.0, "step": 3670 }, { "epoch": 0.46698893270576264, "ewc_loss": 0.05322265625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2157328128814697, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8594516515731812, "num_tokens": 139430367.0, "step": 3671 }, { "epoch": 0.4671161429843531, "ewc_loss": 0.05322265625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1838366985321045, "learning_rate": 1e-05, "loss": 0.4962, "mean_token_accuracy": 0.847946047782898, "num_tokens": 139468233.0, "step": 3672 }, { "epoch": 0.46724335326294364, "ewc_loss": 0.052734375, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1749610900878906, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8522064685821533, "num_tokens": 139502377.0, "step": 3673 }, { "epoch": 0.46737056354153417, "ewc_loss": 0.0517578125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.126518726348877, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.846495509147644, "num_tokens": 139540503.0, "step": 3674 }, { "epoch": 0.46749777382012464, "ewc_loss": 0.05224609375, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1503913402557373, "learning_rate": 1e-05, "loss": 0.5167, "mean_token_accuracy": 0.8435502052307129, "num_tokens": 139578063.0, "step": 3675 }, { "epoch": 0.46762498409871517, "ewc_loss": 0.05078125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0963141918182373, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.871094286441803, "num_tokens": 139615451.0, "step": 3676 }, { "epoch": 0.4677521943773057, "ewc_loss": 0.05224609375, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.14931058883667, "learning_rate": 1e-05, "loss": 0.529, "mean_token_accuracy": 0.845639705657959, "num_tokens": 139657779.0, "step": 3677 }, { "epoch": 0.46787940465589617, "ewc_loss": 0.04931640625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.057351589202881, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8579875230789185, "num_tokens": 139698166.0, "step": 3678 }, { "epoch": 0.4680066149344867, "ewc_loss": 0.05224609375, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1814699172973633, "learning_rate": 1e-05, "loss": 0.5368, "mean_token_accuracy": 0.8404002785682678, "num_tokens": 139736970.0, "step": 3679 }, { "epoch": 0.4681338252130772, "ewc_loss": 0.04931640625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.014331579208374, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8555088043212891, "num_tokens": 139771570.0, "step": 3680 }, { "epoch": 0.46826103549166775, "ewc_loss": 0.05322265625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2220101356506348, "learning_rate": 1e-05, "loss": 0.5607, "mean_token_accuracy": 0.833246111869812, "num_tokens": 139811927.0, "step": 3681 }, { "epoch": 0.4683882457702582, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.953986167907715, "learning_rate": 1e-05, "loss": 0.5125, "mean_token_accuracy": 0.8437929749488831, "num_tokens": 139852903.0, "step": 3682 }, { "epoch": 0.46851545604884876, "ewc_loss": 0.05517578125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3017048835754395, "learning_rate": 1e-05, "loss": 0.5168, "mean_token_accuracy": 0.8450610041618347, "num_tokens": 139895375.0, "step": 3683 }, { "epoch": 0.4686426663274393, "ewc_loss": 0.048828125, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.917592763900757, "learning_rate": 1e-05, "loss": 0.5799, "mean_token_accuracy": 0.8325878381729126, "num_tokens": 139927878.0, "step": 3684 }, { "epoch": 0.46876987660602976, "ewc_loss": 0.056640625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4312326908111572, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8534228205680847, "num_tokens": 139959864.0, "step": 3685 }, { "epoch": 0.4688970868846203, "ewc_loss": 0.04833984375, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.8168296813964844, "learning_rate": 1e-05, "loss": 0.5239, "mean_token_accuracy": 0.8429080247879028, "num_tokens": 140003815.0, "step": 3686 }, { "epoch": 0.4690242971632108, "ewc_loss": 0.06201171875, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6523377895355225, "learning_rate": 1e-05, "loss": 0.5436, "mean_token_accuracy": 0.8393267393112183, "num_tokens": 140043126.0, "step": 3687 }, { "epoch": 0.4691515074418013, "ewc_loss": 0.0498046875, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.8310904502868652, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8514819741249084, "num_tokens": 140078466.0, "step": 3688 }, { "epoch": 0.4692787177203918, "ewc_loss": 0.0693359375, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9248390197753906, "learning_rate": 1e-05, "loss": 0.5286, "mean_token_accuracy": 0.8445821404457092, "num_tokens": 140114733.0, "step": 3689 }, { "epoch": 0.46940592799898234, "ewc_loss": 0.05810546875, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.0102977752685547, "learning_rate": 1e-05, "loss": 0.5092, "mean_token_accuracy": 0.8515737652778625, "num_tokens": 140157017.0, "step": 3690 }, { "epoch": 0.4695331382775728, "ewc_loss": 0.072265625, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.09584903717041, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8500254154205322, "num_tokens": 140197408.0, "step": 3691 }, { "epoch": 0.46966034855616334, "ewc_loss": 0.06591796875, "ewc_loss_diag": 9.918212890625e-05, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.385401725769043, "learning_rate": 1e-05, "loss": 0.5374, "mean_token_accuracy": 0.8472679853439331, "num_tokens": 140232458.0, "step": 3692 }, { "epoch": 0.46978755883475387, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9540963172912598, "learning_rate": 1e-05, "loss": 0.5207, "mean_token_accuracy": 0.8526703715324402, "num_tokens": 140273790.0, "step": 3693 }, { "epoch": 0.46991476911334434, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.496419906616211, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8709889650344849, "num_tokens": 140313339.0, "step": 3694 }, { "epoch": 0.47004197939193487, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7466094493865967, "learning_rate": 1e-05, "loss": 0.5088, "mean_token_accuracy": 0.8516727685928345, "num_tokens": 140352572.0, "step": 3695 }, { "epoch": 0.4701691896705254, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4588193893432617, "learning_rate": 1e-05, "loss": 0.5306, "mean_token_accuracy": 0.8460056781768799, "num_tokens": 140385221.0, "step": 3696 }, { "epoch": 0.4702963999491159, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5543501377105713, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8701146841049194, "num_tokens": 140427041.0, "step": 3697 }, { "epoch": 0.4704236102277064, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3330137729644775, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8704648613929749, "num_tokens": 140460956.0, "step": 3698 }, { "epoch": 0.47055082050629693, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4429385662078857, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8699125051498413, "num_tokens": 140497644.0, "step": 3699 }, { "epoch": 0.4706780307848874, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2295329570770264, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8564143180847168, "num_tokens": 140532764.0, "step": 3700 }, { "epoch": 0.47080524106347793, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.399017095565796, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8538615107536316, "num_tokens": 140564705.0, "step": 3701 }, { "epoch": 0.47093245134206846, "ewc_loss": 0.052490234375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.121278762817383, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8519671559333801, "num_tokens": 140606959.0, "step": 3702 }, { "epoch": 0.47105966162065893, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3890702724456787, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8619689345359802, "num_tokens": 140646034.0, "step": 3703 }, { "epoch": 0.47118687189924946, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.041571855545044, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8543508648872375, "num_tokens": 140679948.0, "step": 3704 }, { "epoch": 0.47131408217784, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.466002941131592, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.8457221984863281, "num_tokens": 140713071.0, "step": 3705 }, { "epoch": 0.47144129245643046, "ewc_loss": 0.050048828125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.978607416152954, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8628530502319336, "num_tokens": 140744243.0, "step": 3706 }, { "epoch": 0.471568502735021, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.616525888442993, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8738706707954407, "num_tokens": 140776716.0, "step": 3707 }, { "epoch": 0.4716957130136115, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.9674243927001953, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8750247955322266, "num_tokens": 140808874.0, "step": 3708 }, { "epoch": 0.471822923292202, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7244491577148438, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8661739230155945, "num_tokens": 140846198.0, "step": 3709 }, { "epoch": 0.4719501335707925, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0566272735595703, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8584384918212891, "num_tokens": 140889228.0, "step": 3710 }, { "epoch": 0.47207734384938305, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7909648418426514, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8586792945861816, "num_tokens": 140924549.0, "step": 3711 }, { "epoch": 0.4722045541279735, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.24444580078125, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8681890368461609, "num_tokens": 140960995.0, "step": 3712 }, { "epoch": 0.47233176440656405, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6624960899353027, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8572540283203125, "num_tokens": 141000899.0, "step": 3713 }, { "epoch": 0.4724589746851546, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3175740242004395, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8636680841445923, "num_tokens": 141037736.0, "step": 3714 }, { "epoch": 0.47258618496374505, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4828011989593506, "learning_rate": 1e-05, "loss": 0.5413, "mean_token_accuracy": 0.8387197256088257, "num_tokens": 141068916.0, "step": 3715 }, { "epoch": 0.4727133952423356, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00010013580322265625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3284189701080322, "learning_rate": 1e-05, "loss": 0.5407, "mean_token_accuracy": 0.8389278054237366, "num_tokens": 141102353.0, "step": 3716 }, { "epoch": 0.4728406055209261, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.25742244720459, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.865371823310852, "num_tokens": 141142653.0, "step": 3717 }, { "epoch": 0.4729678157995166, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3040988445281982, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8593555688858032, "num_tokens": 141180091.0, "step": 3718 }, { "epoch": 0.4730950260781071, "ewc_loss": 0.051513671875, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1047933101654053, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8733052611351013, "num_tokens": 141218674.0, "step": 3719 }, { "epoch": 0.47322223635669763, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.328110933303833, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8667899370193481, "num_tokens": 141258680.0, "step": 3720 }, { "epoch": 0.4733494466352881, "ewc_loss": 0.04931640625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.99285626411438, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8505909442901611, "num_tokens": 141297883.0, "step": 3721 }, { "epoch": 0.47347665691387864, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.4077091217041016, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8574352264404297, "num_tokens": 141336549.0, "step": 3722 }, { "epoch": 0.47360386719246916, "ewc_loss": 0.049072265625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9350199699401855, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8574130535125732, "num_tokens": 141368697.0, "step": 3723 }, { "epoch": 0.47373107747105964, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.5098536014556885, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8609943985939026, "num_tokens": 141399349.0, "step": 3724 }, { "epoch": 0.47385828774965016, "ewc_loss": 0.049560546875, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.894972324371338, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8709613680839539, "num_tokens": 141435198.0, "step": 3725 }, { "epoch": 0.4739854980282407, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.668367385864258, "learning_rate": 1e-05, "loss": 0.5598, "mean_token_accuracy": 0.8414598703384399, "num_tokens": 141470659.0, "step": 3726 }, { "epoch": 0.47411270830683117, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.9365358352661133, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8541628122329712, "num_tokens": 141514774.0, "step": 3727 }, { "epoch": 0.4742399185854217, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8237600326538086, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8606778383255005, "num_tokens": 141551071.0, "step": 3728 }, { "epoch": 0.4743671288640122, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1039931774139404, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8571070432662964, "num_tokens": 141590201.0, "step": 3729 }, { "epoch": 0.4744943391426027, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.826549768447876, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8639515042304993, "num_tokens": 141637580.0, "step": 3730 }, { "epoch": 0.4746215494211932, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.27864933013916, "learning_rate": 1e-05, "loss": 0.5506, "mean_token_accuracy": 0.8446335196495056, "num_tokens": 141674092.0, "step": 3731 }, { "epoch": 0.47474875969978375, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00010061264038085938, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.717664957046509, "learning_rate": 1e-05, "loss": 0.5537, "mean_token_accuracy": 0.837394118309021, "num_tokens": 141709416.0, "step": 3732 }, { "epoch": 0.4748759699783743, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.329117774963379, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8558053970336914, "num_tokens": 141742445.0, "step": 3733 }, { "epoch": 0.47500318025696475, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.530984878540039, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8661607503890991, "num_tokens": 141778224.0, "step": 3734 }, { "epoch": 0.4751303905355553, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.291982889175415, "learning_rate": 1e-05, "loss": 0.5244, "mean_token_accuracy": 0.8467402458190918, "num_tokens": 141818239.0, "step": 3735 }, { "epoch": 0.4752576008141458, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3954615592956543, "learning_rate": 1e-05, "loss": 0.5386, "mean_token_accuracy": 0.8417993783950806, "num_tokens": 141853624.0, "step": 3736 }, { "epoch": 0.4753848110927363, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.229215383529663, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8560964465141296, "num_tokens": 141893620.0, "step": 3737 }, { "epoch": 0.4755120213713268, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.3067855834960938, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8640173673629761, "num_tokens": 141930504.0, "step": 3738 }, { "epoch": 0.47563923164991734, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.166090726852417, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8484164476394653, "num_tokens": 141968195.0, "step": 3739 }, { "epoch": 0.4757664419285078, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2567150592803955, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8590916395187378, "num_tokens": 142010143.0, "step": 3740 }, { "epoch": 0.47589365220709834, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.1210622787475586, "learning_rate": 1e-05, "loss": 0.521, "mean_token_accuracy": 0.8452261686325073, "num_tokens": 142051170.0, "step": 3741 }, { "epoch": 0.47602086248568887, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2527682781219482, "learning_rate": 1e-05, "loss": 0.538, "mean_token_accuracy": 0.8371875286102295, "num_tokens": 142084327.0, "step": 3742 }, { "epoch": 0.47614807276427934, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0563786029815674, "learning_rate": 1e-05, "loss": 0.5162, "mean_token_accuracy": 0.8459775447845459, "num_tokens": 142125512.0, "step": 3743 }, { "epoch": 0.47627528304286987, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2743091583251953, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8577466607093811, "num_tokens": 142159117.0, "step": 3744 }, { "epoch": 0.4764024933214604, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9970827102661133, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8567382097244263, "num_tokens": 142194580.0, "step": 3745 }, { "epoch": 0.47652970360005087, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3316519260406494, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8713833689689636, "num_tokens": 142230945.0, "step": 3746 }, { "epoch": 0.4766569138786414, "ewc_loss": 0.04931640625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.932354688644409, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8653618097305298, "num_tokens": 142264792.0, "step": 3747 }, { "epoch": 0.4767841241572319, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4427289962768555, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8670716285705566, "num_tokens": 142302210.0, "step": 3748 }, { "epoch": 0.4769113344358224, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.888411283493042, "learning_rate": 1e-05, "loss": 0.5093, "mean_token_accuracy": 0.8484926223754883, "num_tokens": 142345646.0, "step": 3749 }, { "epoch": 0.4770385447144129, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6047780513763428, "learning_rate": 1e-05, "loss": 0.5694, "mean_token_accuracy": 0.8320866823196411, "num_tokens": 142387097.0, "step": 3750 }, { "epoch": 0.47716575499300345, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00010156631469726562, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.9113030433654785, "learning_rate": 1e-05, "loss": 0.5465, "mean_token_accuracy": 0.8348076343536377, "num_tokens": 142424087.0, "step": 3751 }, { "epoch": 0.4772929652715939, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.764343738555908, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8655146360397339, "num_tokens": 142458837.0, "step": 3752 }, { "epoch": 0.47742017555018446, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.0281026363372803, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8648921251296997, "num_tokens": 142495661.0, "step": 3753 }, { "epoch": 0.477547385828775, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8378748893737793, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8456301689147949, "num_tokens": 142532439.0, "step": 3754 }, { "epoch": 0.47767459610736546, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.216726541519165, "learning_rate": 1e-05, "loss": 0.5255, "mean_token_accuracy": 0.8442411422729492, "num_tokens": 142574032.0, "step": 3755 }, { "epoch": 0.477801806385956, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7409989833831787, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.860936164855957, "num_tokens": 142607438.0, "step": 3756 }, { "epoch": 0.4779290166645465, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3273379802703857, "learning_rate": 1e-05, "loss": 0.5227, "mean_token_accuracy": 0.8457645773887634, "num_tokens": 142642527.0, "step": 3757 }, { "epoch": 0.478056226943137, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5499558448791504, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8561865091323853, "num_tokens": 142677447.0, "step": 3758 }, { "epoch": 0.4781834372217275, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3248603343963623, "learning_rate": 1e-05, "loss": 0.5166, "mean_token_accuracy": 0.8474858403205872, "num_tokens": 142714429.0, "step": 3759 }, { "epoch": 0.47831064750031804, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3588831424713135, "learning_rate": 1e-05, "loss": 0.4982, "mean_token_accuracy": 0.8524192571640015, "num_tokens": 142752742.0, "step": 3760 }, { "epoch": 0.4784378577789085, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.286052942276001, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8568567037582397, "num_tokens": 142788925.0, "step": 3761 }, { "epoch": 0.47856506805749904, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.211681365966797, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8657565712928772, "num_tokens": 142823215.0, "step": 3762 }, { "epoch": 0.47869227833608957, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.26070499420166, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8453018069267273, "num_tokens": 142860346.0, "step": 3763 }, { "epoch": 0.47881948861468004, "ewc_loss": 0.052490234375, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1244802474975586, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8542167544364929, "num_tokens": 142895814.0, "step": 3764 }, { "epoch": 0.4789466988932706, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.253587245941162, "learning_rate": 1e-05, "loss": 0.5562, "mean_token_accuracy": 0.832857608795166, "num_tokens": 142934379.0, "step": 3765 }, { "epoch": 0.4790739091718611, "ewc_loss": 0.051025390625, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0741114616394043, "learning_rate": 1e-05, "loss": 0.5403, "mean_token_accuracy": 0.8358901143074036, "num_tokens": 142970512.0, "step": 3766 }, { "epoch": 0.4792011194504516, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2487969398498535, "learning_rate": 1e-05, "loss": 0.529, "mean_token_accuracy": 0.8412089347839355, "num_tokens": 143008635.0, "step": 3767 }, { "epoch": 0.4793283297290421, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0080606937408447, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8617837429046631, "num_tokens": 143047638.0, "step": 3768 }, { "epoch": 0.47945554000763263, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.284276008605957, "learning_rate": 1e-05, "loss": 0.514, "mean_token_accuracy": 0.8476417064666748, "num_tokens": 143082470.0, "step": 3769 }, { "epoch": 0.4795827502862231, "ewc_loss": 0.049560546875, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9412662982940674, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8619004487991333, "num_tokens": 143126869.0, "step": 3770 }, { "epoch": 0.47970996056481363, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3782646656036377, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.874969482421875, "num_tokens": 143161371.0, "step": 3771 }, { "epoch": 0.47983717084340416, "ewc_loss": 0.04931640625, "ewc_loss_diag": 0.00010251998901367188, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.875882387161255, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8705769181251526, "num_tokens": 143199896.0, "step": 3772 }, { "epoch": 0.47996438112199463, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5529065132141113, "learning_rate": 1e-05, "loss": 0.5128, "mean_token_accuracy": 0.8509262204170227, "num_tokens": 143236577.0, "step": 3773 }, { "epoch": 0.48009159140058516, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.8583176136016846, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.8555172085762024, "num_tokens": 143279780.0, "step": 3774 }, { "epoch": 0.4802188016791757, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.752251148223877, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8607802391052246, "num_tokens": 143321946.0, "step": 3775 }, { "epoch": 0.48034601195776616, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.9310734272003174, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8662236332893372, "num_tokens": 143359278.0, "step": 3776 }, { "epoch": 0.4804732222363567, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.936230421066284, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8504766821861267, "num_tokens": 143398968.0, "step": 3777 }, { "epoch": 0.4806004325149472, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.206981658935547, "learning_rate": 1e-05, "loss": 0.5109, "mean_token_accuracy": 0.849342942237854, "num_tokens": 143438020.0, "step": 3778 }, { "epoch": 0.4807276427935377, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8947031497955322, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8610490560531616, "num_tokens": 143478813.0, "step": 3779 }, { "epoch": 0.4808548530721282, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.338789701461792, "learning_rate": 1e-05, "loss": 0.5284, "mean_token_accuracy": 0.8481601476669312, "num_tokens": 143519766.0, "step": 3780 }, { "epoch": 0.48098206335071875, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.742677688598633, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8621960878372192, "num_tokens": 143562570.0, "step": 3781 }, { "epoch": 0.4811092736293093, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3864939212799072, "learning_rate": 1e-05, "loss": 0.5531, "mean_token_accuracy": 0.8340651988983154, "num_tokens": 143597203.0, "step": 3782 }, { "epoch": 0.48123648390789975, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5677084922790527, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.8539954423904419, "num_tokens": 143635979.0, "step": 3783 }, { "epoch": 0.4813636941864903, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.308015823364258, "learning_rate": 1e-05, "loss": 0.5295, "mean_token_accuracy": 0.8452637791633606, "num_tokens": 143679915.0, "step": 3784 }, { "epoch": 0.4814909044650808, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4272940158843994, "learning_rate": 1e-05, "loss": 0.5375, "mean_token_accuracy": 0.8398962616920471, "num_tokens": 143716275.0, "step": 3785 }, { "epoch": 0.4816181147436713, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2462382316589355, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8470648527145386, "num_tokens": 143752029.0, "step": 3786 }, { "epoch": 0.4817453250222618, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.327439546585083, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8620062470436096, "num_tokens": 143793139.0, "step": 3787 }, { "epoch": 0.48187253530085233, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.162259817123413, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8689152598381042, "num_tokens": 143832790.0, "step": 3788 }, { "epoch": 0.4819997455794428, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.288794755935669, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.851822018623352, "num_tokens": 143876046.0, "step": 3789 }, { "epoch": 0.48212695585803333, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.098144054412842, "learning_rate": 1e-05, "loss": 0.5011, "mean_token_accuracy": 0.8487370014190674, "num_tokens": 143917019.0, "step": 3790 }, { "epoch": 0.48225416613662386, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.288949489593506, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8619841933250427, "num_tokens": 143957789.0, "step": 3791 }, { "epoch": 0.48238137641521434, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0360312461853027, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8603202104568481, "num_tokens": 143993797.0, "step": 3792 }, { "epoch": 0.48250858669380486, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.340909004211426, "learning_rate": 1e-05, "loss": 0.5536, "mean_token_accuracy": 0.838502049446106, "num_tokens": 144033583.0, "step": 3793 }, { "epoch": 0.4826357969723954, "ewc_loss": 0.049560546875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.964341640472412, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8571109771728516, "num_tokens": 144074358.0, "step": 3794 }, { "epoch": 0.48276300725098586, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4509201049804688, "learning_rate": 1e-05, "loss": 0.5264, "mean_token_accuracy": 0.8480992913246155, "num_tokens": 144114480.0, "step": 3795 }, { "epoch": 0.4828902175295764, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.908769130706787, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8621648550033569, "num_tokens": 144152306.0, "step": 3796 }, { "epoch": 0.4830174278081669, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.6061644554138184, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8574838042259216, "num_tokens": 144191134.0, "step": 3797 }, { "epoch": 0.4831446380867574, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.917379856109619, "learning_rate": 1e-05, "loss": 0.4997, "mean_token_accuracy": 0.849213719367981, "num_tokens": 144230165.0, "step": 3798 }, { "epoch": 0.4832718483653479, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7748451232910156, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8594087362289429, "num_tokens": 144265035.0, "step": 3799 }, { "epoch": 0.48339905864393845, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.0348782539367676, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8738710880279541, "num_tokens": 144304692.0, "step": 3800 }, { "epoch": 0.4835262689225289, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.849186897277832, "learning_rate": 1e-05, "loss": 0.5083, "mean_token_accuracy": 0.8500762581825256, "num_tokens": 144345288.0, "step": 3801 }, { "epoch": 0.48365347920111945, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2591145038604736, "learning_rate": 1e-05, "loss": 0.5189, "mean_token_accuracy": 0.8473196625709534, "num_tokens": 144377427.0, "step": 3802 }, { "epoch": 0.48378068947971, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.718238592147827, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.864967942237854, "num_tokens": 144412405.0, "step": 3803 }, { "epoch": 0.48390789975830045, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3362507820129395, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8673961162567139, "num_tokens": 144448093.0, "step": 3804 }, { "epoch": 0.484035110036891, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5228819847106934, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8529860973358154, "num_tokens": 144483280.0, "step": 3805 }, { "epoch": 0.4841623203154815, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3186697959899902, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8756216764450073, "num_tokens": 144520827.0, "step": 3806 }, { "epoch": 0.484289530594072, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3225009441375732, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8637996315956116, "num_tokens": 144562314.0, "step": 3807 }, { "epoch": 0.4844167408726625, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3066246509552, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8614121079444885, "num_tokens": 144599101.0, "step": 3808 }, { "epoch": 0.48454395115125304, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1671783924102783, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8599409461021423, "num_tokens": 144641308.0, "step": 3809 }, { "epoch": 0.4846711614298435, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.298663854598999, "learning_rate": 1e-05, "loss": 0.5336, "mean_token_accuracy": 0.8449285626411438, "num_tokens": 144677267.0, "step": 3810 }, { "epoch": 0.48479837170843404, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0776565074920654, "learning_rate": 1e-05, "loss": 0.5209, "mean_token_accuracy": 0.8445689082145691, "num_tokens": 144713904.0, "step": 3811 }, { "epoch": 0.48492558198702457, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.30692982673645, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8545244336128235, "num_tokens": 144750523.0, "step": 3812 }, { "epoch": 0.48505279226561504, "ewc_loss": 0.050048828125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.9905805587768555, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8554548025131226, "num_tokens": 144791569.0, "step": 3813 }, { "epoch": 0.48518000254420557, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.363340139389038, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8639065027236938, "num_tokens": 144827919.0, "step": 3814 }, { "epoch": 0.4853072128227961, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9288246631622314, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8728952407836914, "num_tokens": 144866859.0, "step": 3815 }, { "epoch": 0.48543442310138657, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.472968578338623, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8706285953521729, "num_tokens": 144903088.0, "step": 3816 }, { "epoch": 0.4855616333799771, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.89221453666687, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.869284987449646, "num_tokens": 144938485.0, "step": 3817 }, { "epoch": 0.4856888436585676, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6449923515319824, "learning_rate": 1e-05, "loss": 0.576, "mean_token_accuracy": 0.839518666267395, "num_tokens": 144974296.0, "step": 3818 }, { "epoch": 0.4858160539371581, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.9214251041412354, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8583483099937439, "num_tokens": 145012391.0, "step": 3819 }, { "epoch": 0.4859432642157486, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.8054120540618896, "learning_rate": 1e-05, "loss": 0.5423, "mean_token_accuracy": 0.8408688306808472, "num_tokens": 145045409.0, "step": 3820 }, { "epoch": 0.48607047449433916, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.0563430786132812, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8634660840034485, "num_tokens": 145081162.0, "step": 3821 }, { "epoch": 0.48619768477292963, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.858376979827881, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8622980117797852, "num_tokens": 145117045.0, "step": 3822 }, { "epoch": 0.48632489505152016, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2383556365966797, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8638145327568054, "num_tokens": 145153931.0, "step": 3823 }, { "epoch": 0.4864521053301107, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7611114978790283, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8623594641685486, "num_tokens": 145193812.0, "step": 3824 }, { "epoch": 0.48657931560870116, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.328864336013794, "learning_rate": 1e-05, "loss": 0.5346, "mean_token_accuracy": 0.8488019704818726, "num_tokens": 145231814.0, "step": 3825 }, { "epoch": 0.4867065258872917, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.584218740463257, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.8515883684158325, "num_tokens": 145279821.0, "step": 3826 }, { "epoch": 0.4868337361658822, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3212008476257324, "learning_rate": 1e-05, "loss": 0.5132, "mean_token_accuracy": 0.8502663969993591, "num_tokens": 145318972.0, "step": 3827 }, { "epoch": 0.4869609464444727, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4628069400787354, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8586252927780151, "num_tokens": 145359479.0, "step": 3828 }, { "epoch": 0.4870881567230632, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.000102996826171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2569491863250732, "learning_rate": 1e-05, "loss": 0.5184, "mean_token_accuracy": 0.8473330736160278, "num_tokens": 145403119.0, "step": 3829 }, { "epoch": 0.48721536700165374, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.323375701904297, "learning_rate": 1e-05, "loss": 0.5028, "mean_token_accuracy": 0.8473268747329712, "num_tokens": 145439715.0, "step": 3830 }, { "epoch": 0.48734257728024427, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2175285816192627, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8652024269104004, "num_tokens": 145480715.0, "step": 3831 }, { "epoch": 0.48746978755883474, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.235288381576538, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.8546913862228394, "num_tokens": 145523501.0, "step": 3832 }, { "epoch": 0.48759699783742527, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.165126085281372, "learning_rate": 1e-05, "loss": 0.5513, "mean_token_accuracy": 0.8327937126159668, "num_tokens": 145563387.0, "step": 3833 }, { "epoch": 0.4877242081160158, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1885273456573486, "learning_rate": 1e-05, "loss": 0.5311, "mean_token_accuracy": 0.841771125793457, "num_tokens": 145607818.0, "step": 3834 }, { "epoch": 0.4878514183946063, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1241729259490967, "learning_rate": 1e-05, "loss": 0.5188, "mean_token_accuracy": 0.8441929817199707, "num_tokens": 145646115.0, "step": 3835 }, { "epoch": 0.4879786286731968, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.17612361907959, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8618916869163513, "num_tokens": 145681555.0, "step": 3836 }, { "epoch": 0.48810583895178733, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0815505981445312, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8654279112815857, "num_tokens": 145717459.0, "step": 3837 }, { "epoch": 0.4882330492303778, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.187433958053589, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8483803868293762, "num_tokens": 145753143.0, "step": 3838 }, { "epoch": 0.48836025950896833, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.036757707595825, "learning_rate": 1e-05, "loss": 0.5085, "mean_token_accuracy": 0.847395658493042, "num_tokens": 145790345.0, "step": 3839 }, { "epoch": 0.48848746978755886, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2177882194519043, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.8497363328933716, "num_tokens": 145832900.0, "step": 3840 }, { "epoch": 0.48861468006614933, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9850151538848877, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8605642318725586, "num_tokens": 145870309.0, "step": 3841 }, { "epoch": 0.48874189034473986, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2869749069213867, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8631808757781982, "num_tokens": 145910276.0, "step": 3842 }, { "epoch": 0.4888691006233304, "ewc_loss": 0.04931640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9065659046173096, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8661829233169556, "num_tokens": 145952945.0, "step": 3843 }, { "epoch": 0.48899631090192086, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4169254302978516, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8525751829147339, "num_tokens": 145987339.0, "step": 3844 }, { "epoch": 0.4891235211805114, "ewc_loss": 0.04931640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.8519797325134277, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8573086261749268, "num_tokens": 146027447.0, "step": 3845 }, { "epoch": 0.4892507314591019, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6078975200653076, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8602347373962402, "num_tokens": 146064192.0, "step": 3846 }, { "epoch": 0.4893779417376924, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.831988573074341, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8577462434768677, "num_tokens": 146103468.0, "step": 3847 }, { "epoch": 0.4895051520162829, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.866734266281128, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8646729588508606, "num_tokens": 146142499.0, "step": 3848 }, { "epoch": 0.48963236229487345, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 2.995589256286621, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8519065380096436, "num_tokens": 146177538.0, "step": 3849 }, { "epoch": 0.4897595725734639, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.030429840087891, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8653591871261597, "num_tokens": 146218259.0, "step": 3850 }, { "epoch": 0.48988678285205445, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.299794912338257, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8573082089424133, "num_tokens": 146256630.0, "step": 3851 }, { "epoch": 0.490013993130645, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9588351249694824, "learning_rate": 1e-05, "loss": 0.5647, "mean_token_accuracy": 0.8385419845581055, "num_tokens": 146290891.0, "step": 3852 }, { "epoch": 0.49014120340923545, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.471937656402588, "learning_rate": 1e-05, "loss": 0.5079, "mean_token_accuracy": 0.8512915372848511, "num_tokens": 146327148.0, "step": 3853 }, { "epoch": 0.490268413687826, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.729398012161255, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.858703076839447, "num_tokens": 146367932.0, "step": 3854 }, { "epoch": 0.4903956239664165, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.422189950942993, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.862248957157135, "num_tokens": 146407639.0, "step": 3855 }, { "epoch": 0.490522834245007, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5561399459838867, "learning_rate": 1e-05, "loss": 0.5118, "mean_token_accuracy": 0.848206639289856, "num_tokens": 146442974.0, "step": 3856 }, { "epoch": 0.4906500445235975, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.329481840133667, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8632739782333374, "num_tokens": 146482510.0, "step": 3857 }, { "epoch": 0.49077725480218803, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4275779724121094, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8700072765350342, "num_tokens": 146522616.0, "step": 3858 }, { "epoch": 0.4909044650807785, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.234020233154297, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8640238046646118, "num_tokens": 146558226.0, "step": 3859 }, { "epoch": 0.49103167535936904, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.356597423553467, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8519459962844849, "num_tokens": 146601873.0, "step": 3860 }, { "epoch": 0.49115888563795956, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.139094591140747, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8585512638092041, "num_tokens": 146645284.0, "step": 3861 }, { "epoch": 0.49128609591655004, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.34600567817688, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8596744537353516, "num_tokens": 146680374.0, "step": 3862 }, { "epoch": 0.49141330619514056, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.049821376800537, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8508310317993164, "num_tokens": 146721915.0, "step": 3863 }, { "epoch": 0.4915405164737311, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3910293579101562, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8671249747276306, "num_tokens": 146758981.0, "step": 3864 }, { "epoch": 0.49166772675232157, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9749622344970703, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8509998321533203, "num_tokens": 146795606.0, "step": 3865 }, { "epoch": 0.4917949370309121, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5140762329101562, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8670032024383545, "num_tokens": 146829681.0, "step": 3866 }, { "epoch": 0.4919221473095026, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9101171493530273, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8722559809684753, "num_tokens": 146874142.0, "step": 3867 }, { "epoch": 0.4920493575880931, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.706212282180786, "learning_rate": 1e-05, "loss": 0.6013, "mean_token_accuracy": 0.8256144523620605, "num_tokens": 146911414.0, "step": 3868 }, { "epoch": 0.4921765678666836, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0331709384918213, "learning_rate": 1e-05, "loss": 0.5366, "mean_token_accuracy": 0.8390182852745056, "num_tokens": 146956376.0, "step": 3869 }, { "epoch": 0.49230377814527415, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7818126678466797, "learning_rate": 1e-05, "loss": 0.5337, "mean_token_accuracy": 0.8457878232002258, "num_tokens": 147001539.0, "step": 3870 }, { "epoch": 0.4924309884238646, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1299028396606445, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8447547554969788, "num_tokens": 147043782.0, "step": 3871 }, { "epoch": 0.49255819870245515, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.781172752380371, "learning_rate": 1e-05, "loss": 0.5488, "mean_token_accuracy": 0.8369308114051819, "num_tokens": 147077431.0, "step": 3872 }, { "epoch": 0.4926854089810457, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2843518257141113, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8661518096923828, "num_tokens": 147113325.0, "step": 3873 }, { "epoch": 0.49281261925963615, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6123647689819336, "learning_rate": 1e-05, "loss": 0.5507, "mean_token_accuracy": 0.8376636505126953, "num_tokens": 147158363.0, "step": 3874 }, { "epoch": 0.4929398295382267, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.335324764251709, "learning_rate": 1e-05, "loss": 0.5372, "mean_token_accuracy": 0.8466812372207642, "num_tokens": 147195868.0, "step": 3875 }, { "epoch": 0.4930670398168172, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.410680055618286, "learning_rate": 1e-05, "loss": 0.5595, "mean_token_accuracy": 0.8340761661529541, "num_tokens": 147237782.0, "step": 3876 }, { "epoch": 0.4931942500954077, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3536458015441895, "learning_rate": 1e-05, "loss": 0.5102, "mean_token_accuracy": 0.8476280570030212, "num_tokens": 147274709.0, "step": 3877 }, { "epoch": 0.4933214603739982, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2098758220672607, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8620432019233704, "num_tokens": 147312261.0, "step": 3878 }, { "epoch": 0.49344867065258874, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.319631814956665, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.857185959815979, "num_tokens": 147352290.0, "step": 3879 }, { "epoch": 0.4935758809311792, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0811145305633545, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8585878610610962, "num_tokens": 147390558.0, "step": 3880 }, { "epoch": 0.49370309120976974, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3434808254241943, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8506877422332764, "num_tokens": 147429597.0, "step": 3881 }, { "epoch": 0.49383030148836027, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.00010347366333007812, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.9863998889923096, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8493967056274414, "num_tokens": 147472302.0, "step": 3882 }, { "epoch": 0.4939575117669508, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.420814037322998, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8696529269218445, "num_tokens": 147502846.0, "step": 3883 }, { "epoch": 0.49408472204554127, "ewc_loss": 0.050048828125, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.918429136276245, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8639366030693054, "num_tokens": 147542351.0, "step": 3884 }, { "epoch": 0.4942119323241318, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5502383708953857, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8616921901702881, "num_tokens": 147576912.0, "step": 3885 }, { "epoch": 0.4943391426027223, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.907378673553467, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8525353074073792, "num_tokens": 147614687.0, "step": 3886 }, { "epoch": 0.4944663528813128, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6993863582611084, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8645334243774414, "num_tokens": 147649048.0, "step": 3887 }, { "epoch": 0.4945935631599033, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.9723196029663086, "learning_rate": 1e-05, "loss": 0.5161, "mean_token_accuracy": 0.8470489382743835, "num_tokens": 147689337.0, "step": 3888 }, { "epoch": 0.49472077343849385, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8217906951904297, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8571189045906067, "num_tokens": 147725040.0, "step": 3889 }, { "epoch": 0.4948479837170843, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.13407564163208, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8555201292037964, "num_tokens": 147761442.0, "step": 3890 }, { "epoch": 0.49497519399567486, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8192343711853027, "learning_rate": 1e-05, "loss": 0.5049, "mean_token_accuracy": 0.8512905240058899, "num_tokens": 147795421.0, "step": 3891 }, { "epoch": 0.4951024042742654, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2674922943115234, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8557422161102295, "num_tokens": 147833591.0, "step": 3892 }, { "epoch": 0.49522961455285586, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.695530414581299, "learning_rate": 1e-05, "loss": 0.5218, "mean_token_accuracy": 0.8450067043304443, "num_tokens": 147870707.0, "step": 3893 }, { "epoch": 0.4953568248314464, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3160367012023926, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8553048968315125, "num_tokens": 147905231.0, "step": 3894 }, { "epoch": 0.4954840351100369, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5322113037109375, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8587639927864075, "num_tokens": 147945890.0, "step": 3895 }, { "epoch": 0.4956112453886274, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.29111909866333, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8582122325897217, "num_tokens": 147984572.0, "step": 3896 }, { "epoch": 0.4957384556672179, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.388653039932251, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8655407428741455, "num_tokens": 148020059.0, "step": 3897 }, { "epoch": 0.49586566594580844, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2350220680236816, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8701273202896118, "num_tokens": 148058442.0, "step": 3898 }, { "epoch": 0.4959928762243989, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2969553470611572, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8646869659423828, "num_tokens": 148093710.0, "step": 3899 }, { "epoch": 0.49612008650298944, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.163255214691162, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8780616521835327, "num_tokens": 148135763.0, "step": 3900 }, { "epoch": 0.49624729678157997, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2550082206726074, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8650023937225342, "num_tokens": 148169065.0, "step": 3901 }, { "epoch": 0.49637450706017044, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00010442733764648438, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.119096517562866, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8449491262435913, "num_tokens": 148205779.0, "step": 3902 }, { "epoch": 0.496501717338761, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2654807567596436, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8414546847343445, "num_tokens": 148241893.0, "step": 3903 }, { "epoch": 0.4966289276173515, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0639781951904297, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.86025071144104, "num_tokens": 148275384.0, "step": 3904 }, { "epoch": 0.496756137895942, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.268829584121704, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8575206995010376, "num_tokens": 148311248.0, "step": 3905 }, { "epoch": 0.4968833481745325, "ewc_loss": 0.050048828125, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.0131287574768066, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8569757342338562, "num_tokens": 148347717.0, "step": 3906 }, { "epoch": 0.49701055845312303, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3279924392700195, "learning_rate": 1e-05, "loss": 0.5466, "mean_token_accuracy": 0.8390100598335266, "num_tokens": 148393395.0, "step": 3907 }, { "epoch": 0.4971377687317135, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.0001049041748046875, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9498307704925537, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8660643696784973, "num_tokens": 148427843.0, "step": 3908 }, { "epoch": 0.49726497901030403, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00010538101196289062, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.43537974357605, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8449499011039734, "num_tokens": 148470813.0, "step": 3909 }, { "epoch": 0.49739218928889456, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00010538101196289062, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9061901569366455, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8553322553634644, "num_tokens": 148504491.0, "step": 3910 }, { "epoch": 0.49751939956748503, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00010538101196289062, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.580125331878662, "learning_rate": 1e-05, "loss": 0.5052, "mean_token_accuracy": 0.8512389659881592, "num_tokens": 148545934.0, "step": 3911 }, { "epoch": 0.49764660984607556, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.9011943340301514, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8629892468452454, "num_tokens": 148583521.0, "step": 3912 }, { "epoch": 0.4977738201246661, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7490487098693848, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8524885773658752, "num_tokens": 148621484.0, "step": 3913 }, { "epoch": 0.49790103040325656, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.007352828979492, "learning_rate": 1e-05, "loss": 0.5232, "mean_token_accuracy": 0.8456172943115234, "num_tokens": 148661198.0, "step": 3914 }, { "epoch": 0.4980282406818471, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8553853034973145, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8556487560272217, "num_tokens": 148702620.0, "step": 3915 }, { "epoch": 0.4981554509604376, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.1916913986206055, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.852454423904419, "num_tokens": 148739991.0, "step": 3916 }, { "epoch": 0.4982826612390281, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7795188426971436, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8733147978782654, "num_tokens": 148780520.0, "step": 3917 }, { "epoch": 0.4984098715176186, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3218600749969482, "learning_rate": 1e-05, "loss": 0.5119, "mean_token_accuracy": 0.8485472202301025, "num_tokens": 148817604.0, "step": 3918 }, { "epoch": 0.49853708179620915, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6050915718078613, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.853268027305603, "num_tokens": 148858028.0, "step": 3919 }, { "epoch": 0.4986642920747996, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.33245849609375, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8598735928535461, "num_tokens": 148895458.0, "step": 3920 }, { "epoch": 0.49879150235339015, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4143292903900146, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.8506640791893005, "num_tokens": 148933305.0, "step": 3921 }, { "epoch": 0.4989187126319807, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.288801431655884, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8687556982040405, "num_tokens": 148970587.0, "step": 3922 }, { "epoch": 0.49904592291057115, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2739293575286865, "learning_rate": 1e-05, "loss": 0.5036, "mean_token_accuracy": 0.8507741093635559, "num_tokens": 149004874.0, "step": 3923 }, { "epoch": 0.4991731331891617, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2619941234588623, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8613529801368713, "num_tokens": 149045169.0, "step": 3924 }, { "epoch": 0.4993003434677522, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1963133811950684, "learning_rate": 1e-05, "loss": 0.5777, "mean_token_accuracy": 0.8256376385688782, "num_tokens": 149081554.0, "step": 3925 }, { "epoch": 0.4994275537463427, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.217250108718872, "learning_rate": 1e-05, "loss": 0.5146, "mean_token_accuracy": 0.8470422029495239, "num_tokens": 149124500.0, "step": 3926 }, { "epoch": 0.4995547640249332, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1210756301879883, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8502423763275146, "num_tokens": 149162522.0, "step": 3927 }, { "epoch": 0.49968197430352373, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.193990707397461, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.858784556388855, "num_tokens": 149198083.0, "step": 3928 }, { "epoch": 0.4998091845821142, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0777745246887207, "learning_rate": 1e-05, "loss": 0.5128, "mean_token_accuracy": 0.842779278755188, "num_tokens": 149234786.0, "step": 3929 }, { "epoch": 0.49993639486070474, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00010585784912109375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.195056676864624, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8732259273529053, "num_tokens": 149269859.0, "step": 3930 }, { "epoch": 0.5000636051392953, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0418548583984375, "learning_rate": 1e-05, "loss": 0.5452, "mean_token_accuracy": 0.8348828554153442, "num_tokens": 149311209.0, "step": 3931 }, { "epoch": 0.5001908154178858, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2187232971191406, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.850978672504425, "num_tokens": 149351283.0, "step": 3932 }, { "epoch": 0.5003180256964763, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9853310585021973, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8564271330833435, "num_tokens": 149391068.0, "step": 3933 }, { "epoch": 0.5004452359750667, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2838199138641357, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.8476389646530151, "num_tokens": 149435078.0, "step": 3934 }, { "epoch": 0.5005724462536573, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9265336990356445, "learning_rate": 1e-05, "loss": 0.528, "mean_token_accuracy": 0.8445099592208862, "num_tokens": 149481336.0, "step": 3935 }, { "epoch": 0.5006996565322478, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4027352333068848, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8568645119667053, "num_tokens": 149523278.0, "step": 3936 }, { "epoch": 0.5008268668108383, "ewc_loss": 0.04931640625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.8489792346954346, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8744118213653564, "num_tokens": 149563011.0, "step": 3937 }, { "epoch": 0.5009540770894289, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.600438117980957, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.854118824005127, "num_tokens": 149606609.0, "step": 3938 }, { "epoch": 0.5010812873680194, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.82607364654541, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8684264421463013, "num_tokens": 149647012.0, "step": 3939 }, { "epoch": 0.5012084976466098, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.869211435317993, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8597213625907898, "num_tokens": 149680153.0, "step": 3940 }, { "epoch": 0.5013357079252003, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 2.9652044773101807, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8627965450286865, "num_tokens": 149717080.0, "step": 3941 }, { "epoch": 0.5014629182037909, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.074735164642334, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8601726293563843, "num_tokens": 149755934.0, "step": 3942 }, { "epoch": 0.5015901284823814, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3011038303375244, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8571215867996216, "num_tokens": 149792714.0, "step": 3943 }, { "epoch": 0.5017173387609719, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.013270378112793, "learning_rate": 1e-05, "loss": 0.5253, "mean_token_accuracy": 0.847449541091919, "num_tokens": 149828647.0, "step": 3944 }, { "epoch": 0.5018445490395624, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.498394250869751, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8565539121627808, "num_tokens": 149864303.0, "step": 3945 }, { "epoch": 0.5019717593181529, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7961113452911377, "learning_rate": 1e-05, "loss": 0.5398, "mean_token_accuracy": 0.8417524099349976, "num_tokens": 149902154.0, "step": 3946 }, { "epoch": 0.5020989695967434, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.4801321029663086, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8628968596458435, "num_tokens": 149930902.0, "step": 3947 }, { "epoch": 0.5022261798753339, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.594938278198242, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8544946908950806, "num_tokens": 149972033.0, "step": 3948 }, { "epoch": 0.5023533901539244, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3640360832214355, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8665902018547058, "num_tokens": 150007278.0, "step": 3949 }, { "epoch": 0.502480600432515, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4581329822540283, "learning_rate": 1e-05, "loss": 0.4981, "mean_token_accuracy": 0.8535037040710449, "num_tokens": 150054512.0, "step": 3950 }, { "epoch": 0.5026078107111055, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.257204055786133, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.8544241189956665, "num_tokens": 150094614.0, "step": 3951 }, { "epoch": 0.5027350209896959, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.38857102394104, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8566074371337891, "num_tokens": 150133011.0, "step": 3952 }, { "epoch": 0.5028622312682864, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1671671867370605, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8523752093315125, "num_tokens": 150170041.0, "step": 3953 }, { "epoch": 0.502989441546877, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.377338171005249, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8530535101890564, "num_tokens": 150206034.0, "step": 3954 }, { "epoch": 0.5031166518254675, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00010633468627929688, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.075683355331421, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8555046319961548, "num_tokens": 150241601.0, "step": 3955 }, { "epoch": 0.503243862104058, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4109368324279785, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.867459774017334, "num_tokens": 150279219.0, "step": 3956 }, { "epoch": 0.5033710723826486, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.982224464416504, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.852781355381012, "num_tokens": 150323160.0, "step": 3957 }, { "epoch": 0.5034982826612391, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5117223262786865, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.859329342842102, "num_tokens": 150363788.0, "step": 3958 }, { "epoch": 0.5036254929398295, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9581081867218018, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8508918881416321, "num_tokens": 150395481.0, "step": 3959 }, { "epoch": 0.50375270321842, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6612446308135986, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8608753681182861, "num_tokens": 150425726.0, "step": 3960 }, { "epoch": 0.5038799134970106, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0706682205200195, "learning_rate": 1e-05, "loss": 0.5127, "mean_token_accuracy": 0.8514536619186401, "num_tokens": 150463073.0, "step": 3961 }, { "epoch": 0.5040071237756011, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.693448305130005, "learning_rate": 1e-05, "loss": 0.5286, "mean_token_accuracy": 0.8422788977622986, "num_tokens": 150502938.0, "step": 3962 }, { "epoch": 0.5041343340541916, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0853495597839355, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8541936874389648, "num_tokens": 150538601.0, "step": 3963 }, { "epoch": 0.5042615443327821, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7093536853790283, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8520317077636719, "num_tokens": 150573745.0, "step": 3964 }, { "epoch": 0.5043887546113726, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1883256435394287, "learning_rate": 1e-05, "loss": 0.5239, "mean_token_accuracy": 0.8429772853851318, "num_tokens": 150613029.0, "step": 3965 }, { "epoch": 0.5045159648899631, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6215198040008545, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8623870015144348, "num_tokens": 150649773.0, "step": 3966 }, { "epoch": 0.5046431751685536, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2419469356536865, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.858333170413971, "num_tokens": 150694764.0, "step": 3967 }, { "epoch": 0.5047703854471441, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4797885417938232, "learning_rate": 1e-05, "loss": 0.5097, "mean_token_accuracy": 0.8491000533103943, "num_tokens": 150735233.0, "step": 3968 }, { "epoch": 0.5048975957257347, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.260209083557129, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8769237995147705, "num_tokens": 150769596.0, "step": 3969 }, { "epoch": 0.5050248060043252, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.329993963241577, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8500033617019653, "num_tokens": 150806096.0, "step": 3970 }, { "epoch": 0.5051520162829156, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2522315979003906, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.854802131652832, "num_tokens": 150844721.0, "step": 3971 }, { "epoch": 0.5052792265615061, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1847219467163086, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8576554656028748, "num_tokens": 150885711.0, "step": 3972 }, { "epoch": 0.5054064368400967, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2570579051971436, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8527557849884033, "num_tokens": 150924993.0, "step": 3973 }, { "epoch": 0.5055336471186872, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0964393615722656, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8675580620765686, "num_tokens": 150956139.0, "step": 3974 }, { "epoch": 0.5056608573972777, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010776519775390625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2725963592529297, "learning_rate": 1e-05, "loss": 0.5544, "mean_token_accuracy": 0.8374572992324829, "num_tokens": 150995816.0, "step": 3975 }, { "epoch": 0.5057880676758683, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0193886756896973, "learning_rate": 1e-05, "loss": 0.5243, "mean_token_accuracy": 0.8397804498672485, "num_tokens": 151035191.0, "step": 3976 }, { "epoch": 0.5059152779544587, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3121535778045654, "learning_rate": 1e-05, "loss": 0.5358, "mean_token_accuracy": 0.8363519906997681, "num_tokens": 151069625.0, "step": 3977 }, { "epoch": 0.5060424882330492, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.96541166305542, "learning_rate": 1e-05, "loss": 0.5187, "mean_token_accuracy": 0.8463982343673706, "num_tokens": 151105999.0, "step": 3978 }, { "epoch": 0.5061696985116397, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.382154941558838, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8559041023254395, "num_tokens": 151148712.0, "step": 3979 }, { "epoch": 0.5062969087902303, "ewc_loss": 0.049560546875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9020729064941406, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8704285025596619, "num_tokens": 151182842.0, "step": 3980 }, { "epoch": 0.5064241190688208, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5179977416992188, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8541927933692932, "num_tokens": 151223731.0, "step": 3981 }, { "epoch": 0.5065513293474113, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.875495195388794, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.852522611618042, "num_tokens": 151259711.0, "step": 3982 }, { "epoch": 0.5066785396260017, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6964001655578613, "learning_rate": 1e-05, "loss": 0.5209, "mean_token_accuracy": 0.8487824201583862, "num_tokens": 151301486.0, "step": 3983 }, { "epoch": 0.5068057499045923, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.9204046726226807, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8544143438339233, "num_tokens": 151339068.0, "step": 3984 }, { "epoch": 0.5069329601831828, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8840885162353516, "learning_rate": 1e-05, "loss": 0.5315, "mean_token_accuracy": 0.8457559943199158, "num_tokens": 151380694.0, "step": 3985 }, { "epoch": 0.5070601704617733, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001087188720703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.106323003768921, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.8587615489959717, "num_tokens": 151416522.0, "step": 3986 }, { "epoch": 0.5071873807403638, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001087188720703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9317517280578613, "learning_rate": 1e-05, "loss": 0.5534, "mean_token_accuracy": 0.8393929600715637, "num_tokens": 151453786.0, "step": 3987 }, { "epoch": 0.5073145910189544, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3207921981811523, "learning_rate": 1e-05, "loss": 0.5058, "mean_token_accuracy": 0.855288028717041, "num_tokens": 151492155.0, "step": 3988 }, { "epoch": 0.5074418012975448, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7808616161346436, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8624603748321533, "num_tokens": 151534979.0, "step": 3989 }, { "epoch": 0.5075690115761353, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3865463733673096, "learning_rate": 1e-05, "loss": 0.524, "mean_token_accuracy": 0.8504815697669983, "num_tokens": 151573670.0, "step": 3990 }, { "epoch": 0.5076962218547258, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.594203233718872, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8811222910881042, "num_tokens": 151605232.0, "step": 3991 }, { "epoch": 0.5078234321333164, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3589670658111572, "learning_rate": 1e-05, "loss": 0.5462, "mean_token_accuracy": 0.838085949420929, "num_tokens": 151648696.0, "step": 3992 }, { "epoch": 0.5079506424119069, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4207329750061035, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8650942444801331, "num_tokens": 151684203.0, "step": 3993 }, { "epoch": 0.5080778526904974, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3033628463745117, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8482666015625, "num_tokens": 151716749.0, "step": 3994 }, { "epoch": 0.5082050629690879, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.293762445449829, "learning_rate": 1e-05, "loss": 0.5231, "mean_token_accuracy": 0.8439121246337891, "num_tokens": 151752567.0, "step": 3995 }, { "epoch": 0.5083322732476784, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2231340408325195, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8768645524978638, "num_tokens": 151794201.0, "step": 3996 }, { "epoch": 0.5084594835262689, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.216959238052368, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8602117300033569, "num_tokens": 151833731.0, "step": 3997 }, { "epoch": 0.5085866938048594, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1848771572113037, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8578514456748962, "num_tokens": 151872738.0, "step": 3998 }, { "epoch": 0.50871390408345, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1644949913024902, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8641453385353088, "num_tokens": 151916801.0, "step": 3999 }, { "epoch": 0.5088411143620405, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1591508388519287, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8522442579269409, "num_tokens": 151949661.0, "step": 4000 }, { "epoch": 0.5089683246406309, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.155519485473633, "learning_rate": 1e-05, "loss": 0.5006, "mean_token_accuracy": 0.8478166460990906, "num_tokens": 151986719.0, "step": 4001 }, { "epoch": 0.5090955349192214, "ewc_loss": 0.052001953125, "ewc_loss_diag": 0.00010919570922851562, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.128072738647461, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.8511221408843994, "num_tokens": 152023921.0, "step": 4002 }, { "epoch": 0.509222745197812, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00010824203491210938, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.132183074951172, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.864742636680603, "num_tokens": 152061246.0, "step": 4003 }, { "epoch": 0.5093499554764025, "ewc_loss": 0.051513671875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.100508689880371, "learning_rate": 1e-05, "loss": 0.5026, "mean_token_accuracy": 0.8458930253982544, "num_tokens": 152103299.0, "step": 4004 }, { "epoch": 0.509477165754993, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.133131265640259, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.8458573222160339, "num_tokens": 152148576.0, "step": 4005 }, { "epoch": 0.5096043760335836, "ewc_loss": 0.051025390625, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0868141651153564, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.854589581489563, "num_tokens": 152189221.0, "step": 4006 }, { "epoch": 0.5097315863121741, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1384778022766113, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.8506305813789368, "num_tokens": 152232732.0, "step": 4007 }, { "epoch": 0.5098587965907645, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.0665271282196045, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8611472845077515, "num_tokens": 152268715.0, "step": 4008 }, { "epoch": 0.509986006869355, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.162325382232666, "learning_rate": 1e-05, "loss": 0.503, "mean_token_accuracy": 0.8496749997138977, "num_tokens": 152311141.0, "step": 4009 }, { "epoch": 0.5101132171479456, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.037929058074951, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.870181679725647, "num_tokens": 152348892.0, "step": 4010 }, { "epoch": 0.5102404274265361, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.2032711505889893, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8595607280731201, "num_tokens": 152383300.0, "step": 4011 }, { "epoch": 0.5103676377051266, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9862163066864014, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8519272804260254, "num_tokens": 152424404.0, "step": 4012 }, { "epoch": 0.5104948479837171, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.263868570327759, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8725252151489258, "num_tokens": 152465226.0, "step": 4013 }, { "epoch": 0.5106220582623076, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.937171697616577, "learning_rate": 1e-05, "loss": 0.5108, "mean_token_accuracy": 0.8488454818725586, "num_tokens": 152507819.0, "step": 4014 }, { "epoch": 0.5107492685408981, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3709495067596436, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8777665495872498, "num_tokens": 152548194.0, "step": 4015 }, { "epoch": 0.5108764788194886, "ewc_loss": 0.049072265625, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 3.814697265625e-05, "grad_norm": 2.86087703704834, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8573683500289917, "num_tokens": 152580255.0, "step": 4016 }, { "epoch": 0.5110036890980791, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.559048891067505, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8549551963806152, "num_tokens": 152619813.0, "step": 4017 }, { "epoch": 0.5111308993766697, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.799112319946289, "learning_rate": 1e-05, "loss": 0.5128, "mean_token_accuracy": 0.8489426970481873, "num_tokens": 152655552.0, "step": 4018 }, { "epoch": 0.5112581096552602, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8532254695892334, "learning_rate": 1e-05, "loss": 0.5421, "mean_token_accuracy": 0.8434465527534485, "num_tokens": 152693695.0, "step": 4019 }, { "epoch": 0.5113853199338506, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.882096767425537, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8537038564682007, "num_tokens": 152732917.0, "step": 4020 }, { "epoch": 0.5115125302124411, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.169466018676758, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8627904653549194, "num_tokens": 152768052.0, "step": 4021 }, { "epoch": 0.5116397404910317, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.289306163787842, "learning_rate": 1e-05, "loss": 0.5426, "mean_token_accuracy": 0.8418550491333008, "num_tokens": 152808106.0, "step": 4022 }, { "epoch": 0.5117669507696222, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.150540351867676, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8788516521453857, "num_tokens": 152849989.0, "step": 4023 }, { "epoch": 0.5118941610482127, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.592547655105591, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8611540198326111, "num_tokens": 152886599.0, "step": 4024 }, { "epoch": 0.5120213713268033, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.9135630130767822, "learning_rate": 1e-05, "loss": 0.53, "mean_token_accuracy": 0.8496657013893127, "num_tokens": 152926383.0, "step": 4025 }, { "epoch": 0.5121485816053937, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.570742607116699, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8508078455924988, "num_tokens": 152966141.0, "step": 4026 }, { "epoch": 0.5122757918839842, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.71995210647583, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8609417080879211, "num_tokens": 152998217.0, "step": 4027 }, { "epoch": 0.5124030021625747, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.462648391723633, "learning_rate": 1e-05, "loss": 0.558, "mean_token_accuracy": 0.8415369987487793, "num_tokens": 153032258.0, "step": 4028 }, { "epoch": 0.5125302124411653, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5572400093078613, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8600678443908691, "num_tokens": 153075762.0, "step": 4029 }, { "epoch": 0.5126574227197558, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3068158626556396, "learning_rate": 1e-05, "loss": 0.5109, "mean_token_accuracy": 0.8518281579017639, "num_tokens": 153112571.0, "step": 4030 }, { "epoch": 0.5127846329983463, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4771082401275635, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8653315305709839, "num_tokens": 153153348.0, "step": 4031 }, { "epoch": 0.5129118432769367, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1889050006866455, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8660612106323242, "num_tokens": 153189616.0, "step": 4032 }, { "epoch": 0.5130390535555273, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.482877254486084, "learning_rate": 1e-05, "loss": 0.5464, "mean_token_accuracy": 0.8417096138000488, "num_tokens": 153230879.0, "step": 4033 }, { "epoch": 0.5131662638341178, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0733730792999268, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8511753678321838, "num_tokens": 153275234.0, "step": 4034 }, { "epoch": 0.5132934741127083, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00011014938354492188, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.556379795074463, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8668808341026306, "num_tokens": 153312606.0, "step": 4035 }, { "epoch": 0.5134206843912988, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0103654861450195, "learning_rate": 1e-05, "loss": 0.5114, "mean_token_accuracy": 0.8455016016960144, "num_tokens": 153352370.0, "step": 4036 }, { "epoch": 0.5135478946698894, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.7001006603240967, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8580117225646973, "num_tokens": 153393030.0, "step": 4037 }, { "epoch": 0.5136751049484798, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0535857677459717, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8538885116577148, "num_tokens": 153423831.0, "step": 4038 }, { "epoch": 0.5138023152270703, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.814220666885376, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8673062324523926, "num_tokens": 153458388.0, "step": 4039 }, { "epoch": 0.5139295255056608, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1736724376678467, "learning_rate": 1e-05, "loss": 0.5145, "mean_token_accuracy": 0.848982572555542, "num_tokens": 153503486.0, "step": 4040 }, { "epoch": 0.5140567357842514, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.812622308731079, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.860497236251831, "num_tokens": 153541474.0, "step": 4041 }, { "epoch": 0.5141839460628419, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.32582688331604, "learning_rate": 1e-05, "loss": 0.5295, "mean_token_accuracy": 0.8463512659072876, "num_tokens": 153583518.0, "step": 4042 }, { "epoch": 0.5143111563414324, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6536505222320557, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8547247648239136, "num_tokens": 153617552.0, "step": 4043 }, { "epoch": 0.5144383666200228, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.387833595275879, "learning_rate": 1e-05, "loss": 0.5401, "mean_token_accuracy": 0.8440688848495483, "num_tokens": 153655616.0, "step": 4044 }, { "epoch": 0.5145655768986134, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.426236152648926, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8631159663200378, "num_tokens": 153692835.0, "step": 4045 }, { "epoch": 0.5146927871772039, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3837337493896484, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.851158618927002, "num_tokens": 153720867.0, "step": 4046 }, { "epoch": 0.5148199974557944, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2239553928375244, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8613278865814209, "num_tokens": 153758798.0, "step": 4047 }, { "epoch": 0.514947207734385, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3591887950897217, "learning_rate": 1e-05, "loss": 0.5471, "mean_token_accuracy": 0.8410799503326416, "num_tokens": 153798257.0, "step": 4048 }, { "epoch": 0.5150744180129755, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0749828815460205, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8679826259613037, "num_tokens": 153839241.0, "step": 4049 }, { "epoch": 0.5152016282915659, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3725881576538086, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8554301261901855, "num_tokens": 153882522.0, "step": 4050 }, { "epoch": 0.5153288385701564, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 3.0010857582092285, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8688937425613403, "num_tokens": 153912173.0, "step": 4051 }, { "epoch": 0.515456048848747, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4453394412994385, "learning_rate": 1e-05, "loss": 0.5972, "mean_token_accuracy": 0.8211691379547119, "num_tokens": 153955070.0, "step": 4052 }, { "epoch": 0.5155832591273375, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9192357063293457, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8615375757217407, "num_tokens": 153995526.0, "step": 4053 }, { "epoch": 0.515710469405928, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5854640007019043, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8536490797996521, "num_tokens": 154035476.0, "step": 4054 }, { "epoch": 0.5158376796845185, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.927953004837036, "learning_rate": 1e-05, "loss": 0.5064, "mean_token_accuracy": 0.8485856056213379, "num_tokens": 154079108.0, "step": 4055 }, { "epoch": 0.5159648899631091, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.749551773071289, "learning_rate": 1e-05, "loss": 0.5873, "mean_token_accuracy": 0.829217255115509, "num_tokens": 154118484.0, "step": 4056 }, { "epoch": 0.5160921002416995, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.000110626220703125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 2.9924726486206055, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8560733795166016, "num_tokens": 154163981.0, "step": 4057 }, { "epoch": 0.51621931052029, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8724091053009033, "learning_rate": 1e-05, "loss": 0.5463, "mean_token_accuracy": 0.8437851667404175, "num_tokens": 154205035.0, "step": 4058 }, { "epoch": 0.5163465207988805, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.191532611846924, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8658573627471924, "num_tokens": 154241243.0, "step": 4059 }, { "epoch": 0.5164737310774711, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8303074836730957, "learning_rate": 1e-05, "loss": 0.5452, "mean_token_accuracy": 0.8416054844856262, "num_tokens": 154289182.0, "step": 4060 }, { "epoch": 0.5166009413560616, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3350634574890137, "learning_rate": 1e-05, "loss": 0.5231, "mean_token_accuracy": 0.8462708592414856, "num_tokens": 154326735.0, "step": 4061 }, { "epoch": 0.5167281516346521, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.666170835494995, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8566330671310425, "num_tokens": 154367795.0, "step": 4062 }, { "epoch": 0.5168553619132426, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.358442544937134, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8711361289024353, "num_tokens": 154401949.0, "step": 4063 }, { "epoch": 0.5169825721918331, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.482485294342041, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8669501543045044, "num_tokens": 154442693.0, "step": 4064 }, { "epoch": 0.5171097824704236, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2998859882354736, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8571312427520752, "num_tokens": 154489219.0, "step": 4065 }, { "epoch": 0.5172369927490141, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3451311588287354, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8710898160934448, "num_tokens": 154526639.0, "step": 4066 }, { "epoch": 0.5173642030276047, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.276108503341675, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8689482808113098, "num_tokens": 154557339.0, "step": 4067 }, { "epoch": 0.5174914133061952, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2454938888549805, "learning_rate": 1e-05, "loss": 0.5006, "mean_token_accuracy": 0.8512322902679443, "num_tokens": 154592738.0, "step": 4068 }, { "epoch": 0.5176186235847856, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2114250659942627, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8549529314041138, "num_tokens": 154631633.0, "step": 4069 }, { "epoch": 0.5177458338633761, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1784801483154297, "learning_rate": 1e-05, "loss": 0.5634, "mean_token_accuracy": 0.8327047228813171, "num_tokens": 154676274.0, "step": 4070 }, { "epoch": 0.5178730441419667, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.169393539428711, "learning_rate": 1e-05, "loss": 0.5224, "mean_token_accuracy": 0.8412870168685913, "num_tokens": 154716379.0, "step": 4071 }, { "epoch": 0.5180002544205572, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.15199875831604, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8509014844894409, "num_tokens": 154757038.0, "step": 4072 }, { "epoch": 0.5181274646991477, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.127415418624878, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8805118799209595, "num_tokens": 154795684.0, "step": 4073 }, { "epoch": 0.5182546749777382, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.15891695022583, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8594783544540405, "num_tokens": 154834858.0, "step": 4074 }, { "epoch": 0.5183818852563287, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.101470470428467, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8599856495857239, "num_tokens": 154868919.0, "step": 4075 }, { "epoch": 0.5185090955349192, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.165330171585083, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.852895200252533, "num_tokens": 154910511.0, "step": 4076 }, { "epoch": 0.5186363058135097, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0637035369873047, "learning_rate": 1e-05, "loss": 0.529, "mean_token_accuracy": 0.8422108888626099, "num_tokens": 154955345.0, "step": 4077 }, { "epoch": 0.5187635160921003, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1914563179016113, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8614695072174072, "num_tokens": 154992541.0, "step": 4078 }, { "epoch": 0.5188907263706908, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0183022022247314, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8556517958641052, "num_tokens": 155033592.0, "step": 4079 }, { "epoch": 0.5190179366492813, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2342305183410645, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8580389022827148, "num_tokens": 155080034.0, "step": 4080 }, { "epoch": 0.5191451469278717, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.9719717502593994, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8442064523696899, "num_tokens": 155116207.0, "step": 4081 }, { "epoch": 0.5192723572064623, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.319138288497925, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8528046607971191, "num_tokens": 155155787.0, "step": 4082 }, { "epoch": 0.5193995674850528, "ewc_loss": 0.05029296875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.8999624252319336, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8629636764526367, "num_tokens": 155194021.0, "step": 4083 }, { "epoch": 0.5195267777636433, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.478496789932251, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8612856268882751, "num_tokens": 155228685.0, "step": 4084 }, { "epoch": 0.5196539880422338, "ewc_loss": 0.0498046875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 3.8623809814453125e-05, "grad_norm": 2.8251922130584717, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8543589115142822, "num_tokens": 155269146.0, "step": 4085 }, { "epoch": 0.5197811983208244, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7216219902038574, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8566810488700867, "num_tokens": 155309080.0, "step": 4086 }, { "epoch": 0.5199084085994148, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.830247640609741, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8571377992630005, "num_tokens": 155348661.0, "step": 4087 }, { "epoch": 0.5200356188780053, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.036323547363281, "learning_rate": 1e-05, "loss": 0.5204, "mean_token_accuracy": 0.8495774269104004, "num_tokens": 155384984.0, "step": 4088 }, { "epoch": 0.5201628291565958, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00011110305786132812, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.1224167346954346, "learning_rate": 1e-05, "loss": 0.5125, "mean_token_accuracy": 0.8485307097434998, "num_tokens": 155421914.0, "step": 4089 }, { "epoch": 0.5202900394351864, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.141181468963623, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8618849515914917, "num_tokens": 155466460.0, "step": 4090 }, { "epoch": 0.5204172497137769, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.47969388961792, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8602511286735535, "num_tokens": 155504034.0, "step": 4091 }, { "epoch": 0.5205444599923674, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.965902090072632, "learning_rate": 1e-05, "loss": 0.5463, "mean_token_accuracy": 0.8447365760803223, "num_tokens": 155542521.0, "step": 4092 }, { "epoch": 0.5206716702709578, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5721988677978516, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.861587643623352, "num_tokens": 155574170.0, "step": 4093 }, { "epoch": 0.5207988805495484, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7467079162597656, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8590927124023438, "num_tokens": 155604314.0, "step": 4094 }, { "epoch": 0.5209260908281389, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.459768056869507, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8594121932983398, "num_tokens": 155643259.0, "step": 4095 }, { "epoch": 0.5210533011067294, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5794293880462646, "learning_rate": 1e-05, "loss": 0.4966, "mean_token_accuracy": 0.8535050749778748, "num_tokens": 155682522.0, "step": 4096 }, { "epoch": 0.52118051138532, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3456265926361084, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8648676872253418, "num_tokens": 155714021.0, "step": 4097 }, { "epoch": 0.5213077216639105, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.483415365219116, "learning_rate": 1e-05, "loss": 0.5484, "mean_token_accuracy": 0.837410569190979, "num_tokens": 155754064.0, "step": 4098 }, { "epoch": 0.5214349319425009, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.226806879043579, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8584572076797485, "num_tokens": 155788428.0, "step": 4099 }, { "epoch": 0.5215621422210914, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4450247287750244, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8517519235610962, "num_tokens": 155826608.0, "step": 4100 }, { "epoch": 0.521689352499682, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1069345474243164, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.8555070161819458, "num_tokens": 155870314.0, "step": 4101 }, { "epoch": 0.5218165627782725, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.482189416885376, "learning_rate": 1e-05, "loss": 0.5224, "mean_token_accuracy": 0.8440980911254883, "num_tokens": 155907742.0, "step": 4102 }, { "epoch": 0.521943773056863, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00011205673217773438, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.025726556777954, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8712098598480225, "num_tokens": 155949837.0, "step": 4103 }, { "epoch": 0.5220709833354535, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5867881774902344, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8539100289344788, "num_tokens": 155990407.0, "step": 4104 }, { "epoch": 0.522198193614044, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0016448497772217, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.852836012840271, "num_tokens": 156027157.0, "step": 4105 }, { "epoch": 0.5223254038926345, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7384018898010254, "learning_rate": 1e-05, "loss": 0.5044, "mean_token_accuracy": 0.8540005683898926, "num_tokens": 156065074.0, "step": 4106 }, { "epoch": 0.522452614171225, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0589981079101562, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8652344942092896, "num_tokens": 156103399.0, "step": 4107 }, { "epoch": 0.5225798244498155, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8218753337860107, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8605329990386963, "num_tokens": 156143825.0, "step": 4108 }, { "epoch": 0.5227070347284061, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2272074222564697, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8601744174957275, "num_tokens": 156183841.0, "step": 4109 }, { "epoch": 0.5228342450069966, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.752622127532959, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8619116544723511, "num_tokens": 156225288.0, "step": 4110 }, { "epoch": 0.5229614552855871, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.35036301612854, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8716641664505005, "num_tokens": 156257463.0, "step": 4111 }, { "epoch": 0.5230886655641775, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5641629695892334, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8559338450431824, "num_tokens": 156293059.0, "step": 4112 }, { "epoch": 0.5232158758427681, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.361575126647949, "learning_rate": 1e-05, "loss": 0.5642, "mean_token_accuracy": 0.8340457677841187, "num_tokens": 156338605.0, "step": 4113 }, { "epoch": 0.5233430861213586, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.355999231338501, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8526612520217896, "num_tokens": 156375472.0, "step": 4114 }, { "epoch": 0.5234702963999491, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.336880683898926, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8581912517547607, "num_tokens": 156412290.0, "step": 4115 }, { "epoch": 0.5235975066785397, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1879217624664307, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8544667959213257, "num_tokens": 156452268.0, "step": 4116 }, { "epoch": 0.5237247169571302, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3142647743225098, "learning_rate": 1e-05, "loss": 0.5113, "mean_token_accuracy": 0.8494941592216492, "num_tokens": 156500659.0, "step": 4117 }, { "epoch": 0.5238519272357206, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.087919235229492, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8610502481460571, "num_tokens": 156534529.0, "step": 4118 }, { "epoch": 0.5239791375143111, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3235721588134766, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8703141212463379, "num_tokens": 156574459.0, "step": 4119 }, { "epoch": 0.5241063477929017, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 3.009761095046997, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8576412796974182, "num_tokens": 156611626.0, "step": 4120 }, { "epoch": 0.5242335580714922, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.382253885269165, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8608298897743225, "num_tokens": 156649454.0, "step": 4121 }, { "epoch": 0.5243607683500827, "ewc_loss": 0.050048828125, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.941107749938965, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.854480504989624, "num_tokens": 156687263.0, "step": 4122 }, { "epoch": 0.5244879786286732, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.5052871704101562, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.8542591333389282, "num_tokens": 156720384.0, "step": 4123 }, { "epoch": 0.5246151889072637, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.8980624675750732, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8605270981788635, "num_tokens": 156758365.0, "step": 4124 }, { "epoch": 0.5247423991858542, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.674211025238037, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8556900024414062, "num_tokens": 156790411.0, "step": 4125 }, { "epoch": 0.5248696094644447, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9312708377838135, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8574732542037964, "num_tokens": 156829176.0, "step": 4126 }, { "epoch": 0.5249968197430352, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001125335693359375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8485207557678223, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8646091818809509, "num_tokens": 156871932.0, "step": 4127 }, { "epoch": 0.5251240300216258, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.0969150066375732, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8690502643585205, "num_tokens": 156905810.0, "step": 4128 }, { "epoch": 0.5252512403002163, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.906780242919922, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8649340867996216, "num_tokens": 156947218.0, "step": 4129 }, { "epoch": 0.5253784505788067, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.309941053390503, "learning_rate": 1e-05, "loss": 0.5195, "mean_token_accuracy": 0.8462932109832764, "num_tokens": 156978943.0, "step": 4130 }, { "epoch": 0.5255056608573972, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7658965587615967, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8669114708900452, "num_tokens": 157015015.0, "step": 4131 }, { "epoch": 0.5256328711359878, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.380300283432007, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8551145792007446, "num_tokens": 157051789.0, "step": 4132 }, { "epoch": 0.5257600814145783, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5717742443084717, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8664014339447021, "num_tokens": 157090568.0, "step": 4133 }, { "epoch": 0.5258872916931688, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3426997661590576, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.862349271774292, "num_tokens": 157125255.0, "step": 4134 }, { "epoch": 0.5260145019717594, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4033396244049072, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8550445437431335, "num_tokens": 157164917.0, "step": 4135 }, { "epoch": 0.5261417122503498, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.289677858352661, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.8508926630020142, "num_tokens": 157205186.0, "step": 4136 }, { "epoch": 0.5262689225289403, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.289541721343994, "learning_rate": 1e-05, "loss": 0.5214, "mean_token_accuracy": 0.8463589549064636, "num_tokens": 157241511.0, "step": 4137 }, { "epoch": 0.5263961328075308, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.216745376586914, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8630449771881104, "num_tokens": 157283381.0, "step": 4138 }, { "epoch": 0.5265233430861214, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.227856397628784, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8553306460380554, "num_tokens": 157321315.0, "step": 4139 }, { "epoch": 0.5266505533647119, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1705663204193115, "learning_rate": 1e-05, "loss": 0.5199, "mean_token_accuracy": 0.8447832465171814, "num_tokens": 157359205.0, "step": 4140 }, { "epoch": 0.5267777636433024, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1924755573272705, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8648276925086975, "num_tokens": 157398889.0, "step": 4141 }, { "epoch": 0.5269049739218928, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1229236125946045, "learning_rate": 1e-05, "loss": 0.5254, "mean_token_accuracy": 0.8462634086608887, "num_tokens": 157440831.0, "step": 4142 }, { "epoch": 0.5270321842004834, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1975150108337402, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8582165241241455, "num_tokens": 157476275.0, "step": 4143 }, { "epoch": 0.5271593944790739, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.081973075866699, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8715242147445679, "num_tokens": 157508830.0, "step": 4144 }, { "epoch": 0.5272866047576644, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2244300842285156, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8627576231956482, "num_tokens": 157539077.0, "step": 4145 }, { "epoch": 0.527413815036255, "ewc_loss": 0.051025390625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0165417194366455, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8682091236114502, "num_tokens": 157583501.0, "step": 4146 }, { "epoch": 0.5275410253148455, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.273625612258911, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8563206195831299, "num_tokens": 157629778.0, "step": 4147 }, { "epoch": 0.5276682355934359, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.960679292678833, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8610280156135559, "num_tokens": 157669375.0, "step": 4148 }, { "epoch": 0.5277954458720264, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3815765380859375, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8678880929946899, "num_tokens": 157707400.0, "step": 4149 }, { "epoch": 0.527922656150617, "ewc_loss": 0.049560546875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 3.838539123535156e-05, "grad_norm": 2.8988304138183594, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8557402491569519, "num_tokens": 157747039.0, "step": 4150 }, { "epoch": 0.5280498664292075, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.552095413208008, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8578034043312073, "num_tokens": 157782790.0, "step": 4151 }, { "epoch": 0.528177076707798, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.831686019897461, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.861228883266449, "num_tokens": 157832002.0, "step": 4152 }, { "epoch": 0.5283042869863885, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.814523458480835, "learning_rate": 1e-05, "loss": 0.53, "mean_token_accuracy": 0.8469215631484985, "num_tokens": 157869875.0, "step": 4153 }, { "epoch": 0.528431497264979, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.9348156452178955, "learning_rate": 1e-05, "loss": 0.4962, "mean_token_accuracy": 0.8504103422164917, "num_tokens": 157908910.0, "step": 4154 }, { "epoch": 0.5285587075435695, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.050212383270264, "learning_rate": 1e-05, "loss": 0.522, "mean_token_accuracy": 0.8423194885253906, "num_tokens": 157944366.0, "step": 4155 }, { "epoch": 0.52868591782216, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.245570659637451, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8585702776908875, "num_tokens": 157982053.0, "step": 4156 }, { "epoch": 0.5288131281007505, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.018905162811279, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8564803600311279, "num_tokens": 158022654.0, "step": 4157 }, { "epoch": 0.5289403383793411, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.486433267593384, "learning_rate": 1e-05, "loss": 0.5698, "mean_token_accuracy": 0.8339780569076538, "num_tokens": 158060508.0, "step": 4158 }, { "epoch": 0.5290675486579316, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.804762363433838, "learning_rate": 1e-05, "loss": 0.5325, "mean_token_accuracy": 0.8450653553009033, "num_tokens": 158098956.0, "step": 4159 }, { "epoch": 0.5291947589365221, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.479314088821411, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8528611063957214, "num_tokens": 158134781.0, "step": 4160 }, { "epoch": 0.5293219692151125, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00011301040649414062, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.595869541168213, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8650995492935181, "num_tokens": 158177495.0, "step": 4161 }, { "epoch": 0.5294491794937031, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3826100826263428, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8542865514755249, "num_tokens": 158216672.0, "step": 4162 }, { "epoch": 0.5295763897722936, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4452061653137207, "learning_rate": 1e-05, "loss": 0.5068, "mean_token_accuracy": 0.848446249961853, "num_tokens": 158255397.0, "step": 4163 }, { "epoch": 0.5297036000508841, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.286703109741211, "learning_rate": 1e-05, "loss": 0.4951, "mean_token_accuracy": 0.849672794342041, "num_tokens": 158292230.0, "step": 4164 }, { "epoch": 0.5298308103294747, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.358964204788208, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8586666584014893, "num_tokens": 158329884.0, "step": 4165 }, { "epoch": 0.5299580206080652, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.21178936958313, "learning_rate": 1e-05, "loss": 0.5815, "mean_token_accuracy": 0.8310187458992004, "num_tokens": 158367290.0, "step": 4166 }, { "epoch": 0.5300852308866556, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.3057503700256348, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8784130215644836, "num_tokens": 158407999.0, "step": 4167 }, { "epoch": 0.5302124411652461, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.112894296646118, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8706904649734497, "num_tokens": 158444283.0, "step": 4168 }, { "epoch": 0.5303396514438367, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.313624143600464, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.854398250579834, "num_tokens": 158484807.0, "step": 4169 }, { "epoch": 0.5304668617224272, "ewc_loss": 0.051513671875, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.043210983276367, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8621397018432617, "num_tokens": 158518863.0, "step": 4170 }, { "epoch": 0.5305940720010177, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3681697845458984, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8643936514854431, "num_tokens": 158555057.0, "step": 4171 }, { "epoch": 0.5307212822796082, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.969973564147949, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8604256510734558, "num_tokens": 158596032.0, "step": 4172 }, { "epoch": 0.5308484925581987, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.498185157775879, "learning_rate": 1e-05, "loss": 0.5205, "mean_token_accuracy": 0.8474246263504028, "num_tokens": 158631098.0, "step": 4173 }, { "epoch": 0.5309757028367892, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.927607774734497, "learning_rate": 1e-05, "loss": 0.505, "mean_token_accuracy": 0.8483563661575317, "num_tokens": 158664579.0, "step": 4174 }, { "epoch": 0.5311029131153797, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 4.069387435913086, "learning_rate": 1e-05, "loss": 0.5223, "mean_token_accuracy": 0.8478896617889404, "num_tokens": 158701220.0, "step": 4175 }, { "epoch": 0.5312301233939702, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.0730478763580322, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8638578653335571, "num_tokens": 158744235.0, "step": 4176 }, { "epoch": 0.5313573336725608, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9545788764953613, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8612302541732788, "num_tokens": 158782602.0, "step": 4177 }, { "epoch": 0.5314845439511513, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.2848784923553467, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8616466522216797, "num_tokens": 158813345.0, "step": 4178 }, { "epoch": 0.5316117542297417, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9324111938476562, "learning_rate": 1e-05, "loss": 0.5334, "mean_token_accuracy": 0.8431183099746704, "num_tokens": 158847204.0, "step": 4179 }, { "epoch": 0.5317389645083322, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.435962200164795, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8533371686935425, "num_tokens": 158889481.0, "step": 4180 }, { "epoch": 0.5318661747869228, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.727139472961426, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8659550547599792, "num_tokens": 158929267.0, "step": 4181 }, { "epoch": 0.5319933850655133, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.722099781036377, "learning_rate": 1e-05, "loss": 0.5256, "mean_token_accuracy": 0.8454530239105225, "num_tokens": 158974526.0, "step": 4182 }, { "epoch": 0.5321205953441038, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3958120346069336, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8616446256637573, "num_tokens": 159015428.0, "step": 4183 }, { "epoch": 0.5322478056226944, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4182722568511963, "learning_rate": 1e-05, "loss": 0.5297, "mean_token_accuracy": 0.8454029560089111, "num_tokens": 159050492.0, "step": 4184 }, { "epoch": 0.5323750159012848, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2306172847747803, "learning_rate": 1e-05, "loss": 0.5173, "mean_token_accuracy": 0.8470115065574646, "num_tokens": 159085644.0, "step": 4185 }, { "epoch": 0.5325022261798753, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.387809991836548, "learning_rate": 1e-05, "loss": 0.5453, "mean_token_accuracy": 0.8446444272994995, "num_tokens": 159123093.0, "step": 4186 }, { "epoch": 0.5326294364584658, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.132793664932251, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8555796146392822, "num_tokens": 159157688.0, "step": 4187 }, { "epoch": 0.5327566467370564, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.37953519821167, "learning_rate": 1e-05, "loss": 0.5043, "mean_token_accuracy": 0.8498942852020264, "num_tokens": 159194473.0, "step": 4188 }, { "epoch": 0.5328838570156469, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00011396408081054688, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.061732530593872, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.8505010604858398, "num_tokens": 159230314.0, "step": 4189 }, { "epoch": 0.5330110672942374, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.406688690185547, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8643746972084045, "num_tokens": 159266579.0, "step": 4190 }, { "epoch": 0.5331382775728278, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.973179340362549, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8664268255233765, "num_tokens": 159306502.0, "step": 4191 }, { "epoch": 0.5332654878514184, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5100507736206055, "learning_rate": 1e-05, "loss": 0.5233, "mean_token_accuracy": 0.840592086315155, "num_tokens": 159342512.0, "step": 4192 }, { "epoch": 0.5333926981300089, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.930596351623535, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.862433671951294, "num_tokens": 159378408.0, "step": 4193 }, { "epoch": 0.5335199084085994, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.669975519180298, "learning_rate": 1e-05, "loss": 0.563, "mean_token_accuracy": 0.8361116647720337, "num_tokens": 159415843.0, "step": 4194 }, { "epoch": 0.53364711868719, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.96636700630188, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8571139574050903, "num_tokens": 159446880.0, "step": 4195 }, { "epoch": 0.5337743289657805, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8027591705322266, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8593801259994507, "num_tokens": 159491960.0, "step": 4196 }, { "epoch": 0.5339015392443709, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00011444091796875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.0684235095977783, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8603423237800598, "num_tokens": 159529157.0, "step": 4197 }, { "epoch": 0.5340287495229614, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00011491775512695312, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.881734609603882, "learning_rate": 1e-05, "loss": 0.5142, "mean_token_accuracy": 0.8518413305282593, "num_tokens": 159563589.0, "step": 4198 }, { "epoch": 0.534155959801552, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00011491775512695312, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.237964153289795, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8754537105560303, "num_tokens": 159601789.0, "step": 4199 }, { "epoch": 0.5342831700801425, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00011491775512695312, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8312811851501465, "learning_rate": 1e-05, "loss": 0.5585, "mean_token_accuracy": 0.8342970609664917, "num_tokens": 159637357.0, "step": 4200 }, { "epoch": 0.534410380358733, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3494701385498047, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8702572584152222, "num_tokens": 159678919.0, "step": 4201 }, { "epoch": 0.5345375906373235, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6666481494903564, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.8517901301383972, "num_tokens": 159714109.0, "step": 4202 }, { "epoch": 0.534664800915914, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3496785163879395, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8647158145904541, "num_tokens": 159752171.0, "step": 4203 }, { "epoch": 0.5347920111945045, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4965665340423584, "learning_rate": 1e-05, "loss": 0.5212, "mean_token_accuracy": 0.8476574420928955, "num_tokens": 159791251.0, "step": 4204 }, { "epoch": 0.534919221473095, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.324274778366089, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8662548065185547, "num_tokens": 159819925.0, "step": 4205 }, { "epoch": 0.5350464317516855, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.352694511413574, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.8509962558746338, "num_tokens": 159859788.0, "step": 4206 }, { "epoch": 0.5351736420302761, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.253871440887451, "learning_rate": 1e-05, "loss": 0.512, "mean_token_accuracy": 0.8483665585517883, "num_tokens": 159897202.0, "step": 4207 }, { "epoch": 0.5353008523088666, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2647247314453125, "learning_rate": 1e-05, "loss": 0.5178, "mean_token_accuracy": 0.8461244106292725, "num_tokens": 159935840.0, "step": 4208 }, { "epoch": 0.5354280625874571, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1861727237701416, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8661340475082397, "num_tokens": 159975120.0, "step": 4209 }, { "epoch": 0.5355552728660475, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2177610397338867, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.863437294960022, "num_tokens": 160020935.0, "step": 4210 }, { "epoch": 0.5356824831446381, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1361875534057617, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8704975247383118, "num_tokens": 160065029.0, "step": 4211 }, { "epoch": 0.5358096934232286, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.219723701477051, "learning_rate": 1e-05, "loss": 0.5079, "mean_token_accuracy": 0.8469377756118774, "num_tokens": 160104865.0, "step": 4212 }, { "epoch": 0.5359369037018191, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.078634262084961, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8660128116607666, "num_tokens": 160145477.0, "step": 4213 }, { "epoch": 0.5360641139804097, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2511775493621826, "learning_rate": 1e-05, "loss": 0.5113, "mean_token_accuracy": 0.8490508794784546, "num_tokens": 160184906.0, "step": 4214 }, { "epoch": 0.5361913242590002, "ewc_loss": 0.051513671875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.022021770477295, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8533869981765747, "num_tokens": 160225011.0, "step": 4215 }, { "epoch": 0.5363185345375906, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3196306228637695, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8577414751052856, "num_tokens": 160263193.0, "step": 4216 }, { "epoch": 0.5364457448161811, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.949326753616333, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.882195234298706, "num_tokens": 160299552.0, "step": 4217 }, { "epoch": 0.5365729550947717, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4500515460968018, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8694546818733215, "num_tokens": 160334609.0, "step": 4218 }, { "epoch": 0.5367001653733622, "ewc_loss": 0.050537109375, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.8833088874816895, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8614373803138733, "num_tokens": 160374295.0, "step": 4219 }, { "epoch": 0.5368273756519527, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6501903533935547, "learning_rate": 1e-05, "loss": 0.5367, "mean_token_accuracy": 0.8405287861824036, "num_tokens": 160404420.0, "step": 4220 }, { "epoch": 0.5369545859305432, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.881941556930542, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.85483318567276, "num_tokens": 160439278.0, "step": 4221 }, { "epoch": 0.5370817962091337, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8790736198425293, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8738129734992981, "num_tokens": 160475225.0, "step": 4222 }, { "epoch": 0.5372090064877242, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.043461561203003, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8602392673492432, "num_tokens": 160516073.0, "step": 4223 }, { "epoch": 0.5373362167663147, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.030243873596191, "learning_rate": 1e-05, "loss": 0.5204, "mean_token_accuracy": 0.8523634672164917, "num_tokens": 160550938.0, "step": 4224 }, { "epoch": 0.5374634270449052, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3304271697998047, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8637852668762207, "num_tokens": 160590382.0, "step": 4225 }, { "epoch": 0.5375906373234958, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8992724418640137, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.869376540184021, "num_tokens": 160624830.0, "step": 4226 }, { "epoch": 0.5377178476020863, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.461087703704834, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8547006845474243, "num_tokens": 160663042.0, "step": 4227 }, { "epoch": 0.5378450578806767, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.68996524810791, "learning_rate": 1e-05, "loss": 0.518, "mean_token_accuracy": 0.847934365272522, "num_tokens": 160700502.0, "step": 4228 }, { "epoch": 0.5379722681592672, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4252212047576904, "learning_rate": 1e-05, "loss": 0.542, "mean_token_accuracy": 0.841504693031311, "num_tokens": 160736746.0, "step": 4229 }, { "epoch": 0.5380994784378578, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.666828155517578, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8662992119789124, "num_tokens": 160771430.0, "step": 4230 }, { "epoch": 0.5382266887164483, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3323974609375, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8560791015625, "num_tokens": 160807047.0, "step": 4231 }, { "epoch": 0.5383538989950388, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.438560962677002, "learning_rate": 1e-05, "loss": 0.545, "mean_token_accuracy": 0.8392167687416077, "num_tokens": 160844600.0, "step": 4232 }, { "epoch": 0.5384811092736294, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.222107172012329, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.860770046710968, "num_tokens": 160883132.0, "step": 4233 }, { "epoch": 0.5386083195522198, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3862087726593018, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8585477471351624, "num_tokens": 160922408.0, "step": 4234 }, { "epoch": 0.5387355298308103, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1544854640960693, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8611095547676086, "num_tokens": 160960563.0, "step": 4235 }, { "epoch": 0.5388627401094008, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00011539459228515625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.37195086479187, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8566526174545288, "num_tokens": 160995594.0, "step": 4236 }, { "epoch": 0.5389899503879914, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0944955348968506, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8535021543502808, "num_tokens": 161028341.0, "step": 4237 }, { "epoch": 0.5391171606665819, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00011587142944335938, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3919975757598877, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8600355386734009, "num_tokens": 161064549.0, "step": 4238 }, { "epoch": 0.5392443709451724, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.018249273300171, "learning_rate": 1e-05, "loss": 0.5243, "mean_token_accuracy": 0.8425889015197754, "num_tokens": 161105170.0, "step": 4239 }, { "epoch": 0.5393715812237628, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.489027738571167, "learning_rate": 1e-05, "loss": 0.5474, "mean_token_accuracy": 0.8381199240684509, "num_tokens": 161141018.0, "step": 4240 }, { "epoch": 0.5394987915023534, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.9773736000061035, "learning_rate": 1e-05, "loss": 0.5281, "mean_token_accuracy": 0.8414285778999329, "num_tokens": 161175363.0, "step": 4241 }, { "epoch": 0.5396260017809439, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6051032543182373, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8628897666931152, "num_tokens": 161210920.0, "step": 4242 }, { "epoch": 0.5397532120595344, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.967554807662964, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8699272871017456, "num_tokens": 161248743.0, "step": 4243 }, { "epoch": 0.5398804223381249, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7498223781585693, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8613094091415405, "num_tokens": 161282897.0, "step": 4244 }, { "epoch": 0.5400076326167155, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.072967529296875, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8581459522247314, "num_tokens": 161323103.0, "step": 4245 }, { "epoch": 0.5401348428953059, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8069400787353516, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8570608496665955, "num_tokens": 161359612.0, "step": 4246 }, { "epoch": 0.5402620531738964, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2243144512176514, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8690840005874634, "num_tokens": 161396961.0, "step": 4247 }, { "epoch": 0.540389263452487, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7065110206604004, "learning_rate": 1e-05, "loss": 0.5052, "mean_token_accuracy": 0.8523372411727905, "num_tokens": 161437380.0, "step": 4248 }, { "epoch": 0.5405164737310775, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.313394784927368, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.8544390201568604, "num_tokens": 161480237.0, "step": 4249 }, { "epoch": 0.540643684009668, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5215084552764893, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8724086284637451, "num_tokens": 161519447.0, "step": 4250 }, { "epoch": 0.5407708942882585, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.33492112159729, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8490074276924133, "num_tokens": 161556588.0, "step": 4251 }, { "epoch": 0.540898104566849, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.333845615386963, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8611760139465332, "num_tokens": 161591542.0, "step": 4252 }, { "epoch": 0.5410253148454395, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001163482666015625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.325950860977173, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8588078022003174, "num_tokens": 161630453.0, "step": 4253 }, { "epoch": 0.54115252512403, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1846487522125244, "learning_rate": 1e-05, "loss": 0.4009, "mean_token_accuracy": 0.8818100690841675, "num_tokens": 161660927.0, "step": 4254 }, { "epoch": 0.5412797354026205, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.312376022338867, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8597937822341919, "num_tokens": 161703284.0, "step": 4255 }, { "epoch": 0.5414069456812111, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.080766439437866, "learning_rate": 1e-05, "loss": 0.5167, "mean_token_accuracy": 0.8465769290924072, "num_tokens": 161739212.0, "step": 4256 }, { "epoch": 0.5415341559598016, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3554766178131104, "learning_rate": 1e-05, "loss": 0.5187, "mean_token_accuracy": 0.8555494546890259, "num_tokens": 161777635.0, "step": 4257 }, { "epoch": 0.5416613662383921, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0008695125579834, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8518558144569397, "num_tokens": 161816019.0, "step": 4258 }, { "epoch": 0.5417885765169825, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4212594032287598, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.859637975692749, "num_tokens": 161849690.0, "step": 4259 }, { "epoch": 0.5419157867955731, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9253127574920654, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8723939657211304, "num_tokens": 161889596.0, "step": 4260 }, { "epoch": 0.5420429970741636, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5568413734436035, "learning_rate": 1e-05, "loss": 0.5447, "mean_token_accuracy": 0.838520884513855, "num_tokens": 161929451.0, "step": 4261 }, { "epoch": 0.5421702073527541, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9103307723999023, "learning_rate": 1e-05, "loss": 0.5099, "mean_token_accuracy": 0.8480368256568909, "num_tokens": 161966422.0, "step": 4262 }, { "epoch": 0.5422974176313446, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.720961809158325, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8551469445228577, "num_tokens": 162005192.0, "step": 4263 }, { "epoch": 0.5424246279099352, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.9658091068267822, "learning_rate": 1e-05, "loss": 0.5068, "mean_token_accuracy": 0.8544293642044067, "num_tokens": 162051039.0, "step": 4264 }, { "epoch": 0.5425518381885256, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.884168863296509, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8625011444091797, "num_tokens": 162086869.0, "step": 4265 }, { "epoch": 0.5426790484671161, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.159532070159912, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.858696460723877, "num_tokens": 162122442.0, "step": 4266 }, { "epoch": 0.5428062587457066, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.899108648300171, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8605242371559143, "num_tokens": 162153057.0, "step": 4267 }, { "epoch": 0.5429334690242972, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.339684247970581, "learning_rate": 1e-05, "loss": 0.5516, "mean_token_accuracy": 0.8419452905654907, "num_tokens": 162192086.0, "step": 4268 }, { "epoch": 0.5430606793028877, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7456517219543457, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8671997785568237, "num_tokens": 162227206.0, "step": 4269 }, { "epoch": 0.5431878895814782, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3808481693267822, "learning_rate": 1e-05, "loss": 0.5122, "mean_token_accuracy": 0.8508878946304321, "num_tokens": 162263490.0, "step": 4270 }, { "epoch": 0.5433150998600687, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.546475648880005, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8694977760314941, "num_tokens": 162302750.0, "step": 4271 }, { "epoch": 0.5434423101386592, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3364059925079346, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8737053275108337, "num_tokens": 162339613.0, "step": 4272 }, { "epoch": 0.5435695204172497, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.387916088104248, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8522980213165283, "num_tokens": 162380163.0, "step": 4273 }, { "epoch": 0.5436967306958402, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4315145015716553, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.8533824682235718, "num_tokens": 162413306.0, "step": 4274 }, { "epoch": 0.5438239409744308, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2125887870788574, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8651684522628784, "num_tokens": 162446861.0, "step": 4275 }, { "epoch": 0.5439511512530213, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00011730194091796875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3027150630950928, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8510564565658569, "num_tokens": 162491414.0, "step": 4276 }, { "epoch": 0.5440783615316117, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.097407341003418, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.864002525806427, "num_tokens": 162535265.0, "step": 4277 }, { "epoch": 0.5442055718102022, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3134310245513916, "learning_rate": 1e-05, "loss": 0.5201, "mean_token_accuracy": 0.8458239436149597, "num_tokens": 162577198.0, "step": 4278 }, { "epoch": 0.5443327820887928, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0516116619110107, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8619931936264038, "num_tokens": 162609557.0, "step": 4279 }, { "epoch": 0.5444599923673833, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3334386348724365, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8623369336128235, "num_tokens": 162642943.0, "step": 4280 }, { "epoch": 0.5445872026459738, "ewc_loss": 0.052490234375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.999812602996826, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8597385287284851, "num_tokens": 162676444.0, "step": 4281 }, { "epoch": 0.5447144129245644, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3981168270111084, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8542948961257935, "num_tokens": 162715445.0, "step": 4282 }, { "epoch": 0.5448416232031548, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9412732124328613, "learning_rate": 1e-05, "loss": 0.5337, "mean_token_accuracy": 0.8416917324066162, "num_tokens": 162751306.0, "step": 4283 }, { "epoch": 0.5449688334817453, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5150797367095947, "learning_rate": 1e-05, "loss": 0.5775, "mean_token_accuracy": 0.8272817730903625, "num_tokens": 162793587.0, "step": 4284 }, { "epoch": 0.5450960437603358, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.909566879272461, "learning_rate": 1e-05, "loss": 0.5201, "mean_token_accuracy": 0.8419755697250366, "num_tokens": 162831000.0, "step": 4285 }, { "epoch": 0.5452232540389264, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.676276206970215, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8808212280273438, "num_tokens": 162863913.0, "step": 4286 }, { "epoch": 0.5453504643175169, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 2.9380135536193848, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8622719645500183, "num_tokens": 162899135.0, "step": 4287 }, { "epoch": 0.5454776745961074, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.85746169090271, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.869481086730957, "num_tokens": 162932534.0, "step": 4288 }, { "epoch": 0.5456048848746978, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.0862605571746826, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8600876331329346, "num_tokens": 162973199.0, "step": 4289 }, { "epoch": 0.5457320951532884, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9287731647491455, "learning_rate": 1e-05, "loss": 0.519, "mean_token_accuracy": 0.8540322780609131, "num_tokens": 163007018.0, "step": 4290 }, { "epoch": 0.5458593054318789, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2950780391693115, "learning_rate": 1e-05, "loss": 0.5097, "mean_token_accuracy": 0.8532007932662964, "num_tokens": 163044381.0, "step": 4291 }, { "epoch": 0.5459865157104694, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8247416019439697, "learning_rate": 1e-05, "loss": 0.5025, "mean_token_accuracy": 0.8571858406066895, "num_tokens": 163079606.0, "step": 4292 }, { "epoch": 0.5461137259890599, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.386671781539917, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.8531299233436584, "num_tokens": 163120421.0, "step": 4293 }, { "epoch": 0.5462409362676505, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.628159761428833, "learning_rate": 1e-05, "loss": 0.5474, "mean_token_accuracy": 0.8384674191474915, "num_tokens": 163164761.0, "step": 4294 }, { "epoch": 0.5463681465462409, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3701205253601074, "learning_rate": 1e-05, "loss": 0.6341, "mean_token_accuracy": 0.8142585158348083, "num_tokens": 163212432.0, "step": 4295 }, { "epoch": 0.5464953568248314, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4574437141418457, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8741787672042847, "num_tokens": 163251333.0, "step": 4296 }, { "epoch": 0.5466225671034219, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3270785808563232, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8555997014045715, "num_tokens": 163286428.0, "step": 4297 }, { "epoch": 0.5467497773820125, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.316326856613159, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8678683042526245, "num_tokens": 163318974.0, "step": 4298 }, { "epoch": 0.546876987660603, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.274101495742798, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8507803678512573, "num_tokens": 163358621.0, "step": 4299 }, { "epoch": 0.5470041979391935, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.210430383682251, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8644604086875916, "num_tokens": 163401062.0, "step": 4300 }, { "epoch": 0.5471314082177839, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2277939319610596, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8520971536636353, "num_tokens": 163436194.0, "step": 4301 }, { "epoch": 0.5472586184963745, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1619229316711426, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8506457209587097, "num_tokens": 163476403.0, "step": 4302 }, { "epoch": 0.547385828774965, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.197408437728882, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8530958890914917, "num_tokens": 163515822.0, "step": 4303 }, { "epoch": 0.5475130390535555, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1318936347961426, "learning_rate": 1e-05, "loss": 0.5143, "mean_token_accuracy": 0.8453758955001831, "num_tokens": 163553611.0, "step": 4304 }, { "epoch": 0.5476402493321461, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1780643463134766, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8636665344238281, "num_tokens": 163588591.0, "step": 4305 }, { "epoch": 0.5477674596107366, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.102287530899048, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.858177900314331, "num_tokens": 163630857.0, "step": 4306 }, { "epoch": 0.5478946698893271, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1826658248901367, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.847141444683075, "num_tokens": 163666684.0, "step": 4307 }, { "epoch": 0.5480218801679175, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.068516254425049, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.854503870010376, "num_tokens": 163707409.0, "step": 4308 }, { "epoch": 0.5481490904465081, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2062103748321533, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.861664891242981, "num_tokens": 163745190.0, "step": 4309 }, { "epoch": 0.5482763007250986, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0303597450256348, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8558419942855835, "num_tokens": 163781538.0, "step": 4310 }, { "epoch": 0.5484035110036891, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2583184242248535, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8551232218742371, "num_tokens": 163816412.0, "step": 4311 }, { "epoch": 0.5485307212822796, "ewc_loss": 0.051025390625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9689862728118896, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8647080659866333, "num_tokens": 163853202.0, "step": 4312 }, { "epoch": 0.5486579315608702, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.344271659851074, "learning_rate": 1e-05, "loss": 0.4985, "mean_token_accuracy": 0.8538957238197327, "num_tokens": 163893669.0, "step": 4313 }, { "epoch": 0.5487851418394606, "ewc_loss": 0.051025390625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.890223264694214, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8588224053382874, "num_tokens": 163932372.0, "step": 4314 }, { "epoch": 0.5489123521180511, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5196681022644043, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8693727850914001, "num_tokens": 163969431.0, "step": 4315 }, { "epoch": 0.5490395623966416, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.823063373565674, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8557949066162109, "num_tokens": 164006776.0, "step": 4316 }, { "epoch": 0.5491667726752322, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7877838611602783, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8708112835884094, "num_tokens": 164042178.0, "step": 4317 }, { "epoch": 0.5492939829538227, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.8611412048339844, "learning_rate": 1e-05, "loss": 0.5079, "mean_token_accuracy": 0.8491478562355042, "num_tokens": 164089028.0, "step": 4318 }, { "epoch": 0.5494211932324132, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.1005988121032715, "learning_rate": 1e-05, "loss": 0.5288, "mean_token_accuracy": 0.8563874363899231, "num_tokens": 164130311.0, "step": 4319 }, { "epoch": 0.5495484035110036, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.1993627548217773, "learning_rate": 1e-05, "loss": 0.4975, "mean_token_accuracy": 0.8548270463943481, "num_tokens": 164165398.0, "step": 4320 }, { "epoch": 0.5496756137895942, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.167572021484375, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8676893711090088, "num_tokens": 164202724.0, "step": 4321 }, { "epoch": 0.5498028240681847, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.544466733932495, "learning_rate": 1e-05, "loss": 0.5497, "mean_token_accuracy": 0.8405389189720154, "num_tokens": 164247666.0, "step": 4322 }, { "epoch": 0.5499300343467752, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9464871883392334, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8558408617973328, "num_tokens": 164292323.0, "step": 4323 }, { "epoch": 0.5500572446253658, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5717227458953857, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.872652530670166, "num_tokens": 164331160.0, "step": 4324 }, { "epoch": 0.5501844549039563, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7397096157073975, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8719594478607178, "num_tokens": 164365457.0, "step": 4325 }, { "epoch": 0.5503116651825467, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.470964193344116, "learning_rate": 1e-05, "loss": 0.5136, "mean_token_accuracy": 0.8505352735519409, "num_tokens": 164403731.0, "step": 4326 }, { "epoch": 0.5504388754611372, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.587721109390259, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8624798059463501, "num_tokens": 164436616.0, "step": 4327 }, { "epoch": 0.5505660857397278, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3341493606567383, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.860854983329773, "num_tokens": 164472453.0, "step": 4328 }, { "epoch": 0.5506932960183183, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.515958547592163, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8503192663192749, "num_tokens": 164506551.0, "step": 4329 }, { "epoch": 0.5508205062969088, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2158961296081543, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8713001012802124, "num_tokens": 164542638.0, "step": 4330 }, { "epoch": 0.5509477165754993, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.498199939727783, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8584756255149841, "num_tokens": 164580071.0, "step": 4331 }, { "epoch": 0.5510749268540898, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.120485544204712, "learning_rate": 1e-05, "loss": 0.5259, "mean_token_accuracy": 0.8436303734779358, "num_tokens": 164617894.0, "step": 4332 }, { "epoch": 0.5512021371326803, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5477852821350098, "learning_rate": 1e-05, "loss": 0.561, "mean_token_accuracy": 0.8345465064048767, "num_tokens": 164660509.0, "step": 4333 }, { "epoch": 0.5513293474112708, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.033414602279663, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8687421083450317, "num_tokens": 164699683.0, "step": 4334 }, { "epoch": 0.5514565576898613, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.678068161010742, "learning_rate": 1e-05, "loss": 0.4982, "mean_token_accuracy": 0.8556751012802124, "num_tokens": 164733134.0, "step": 4335 }, { "epoch": 0.5515837679684519, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0440969467163086, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8509094715118408, "num_tokens": 164771911.0, "step": 4336 }, { "epoch": 0.5517109782470424, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.804255485534668, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.8470766544342041, "num_tokens": 164810579.0, "step": 4337 }, { "epoch": 0.5518381885256328, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1535634994506836, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8663043975830078, "num_tokens": 164850215.0, "step": 4338 }, { "epoch": 0.5519653988042234, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.83012056350708, "learning_rate": 1e-05, "loss": 0.5036, "mean_token_accuracy": 0.8512899875640869, "num_tokens": 164889523.0, "step": 4339 }, { "epoch": 0.5520926090828139, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.317660093307495, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8724976778030396, "num_tokens": 164925946.0, "step": 4340 }, { "epoch": 0.5522198193614044, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6918187141418457, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8643583059310913, "num_tokens": 164961954.0, "step": 4341 }, { "epoch": 0.5523470296399949, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.371957302093506, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8604478240013123, "num_tokens": 165008013.0, "step": 4342 }, { "epoch": 0.5524742399185855, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4994072914123535, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.862519383430481, "num_tokens": 165037723.0, "step": 4343 }, { "epoch": 0.5526014501971759, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3778152465820312, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8620784878730774, "num_tokens": 165071432.0, "step": 4344 }, { "epoch": 0.5527286604757664, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2790868282318115, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8653982877731323, "num_tokens": 165118784.0, "step": 4345 }, { "epoch": 0.5528558707543569, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3540291786193848, "learning_rate": 1e-05, "loss": 0.54, "mean_token_accuracy": 0.8412377834320068, "num_tokens": 165163281.0, "step": 4346 }, { "epoch": 0.5529830810329475, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.146280288696289, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8502247333526611, "num_tokens": 165195615.0, "step": 4347 }, { "epoch": 0.553110291311538, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3581221103668213, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8544691205024719, "num_tokens": 165233710.0, "step": 4348 }, { "epoch": 0.5532375015901285, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0420587062835693, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8550839424133301, "num_tokens": 165275788.0, "step": 4349 }, { "epoch": 0.5533647118687189, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.390002965927124, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8659771680831909, "num_tokens": 165316627.0, "step": 4350 }, { "epoch": 0.5534919221473095, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.965635061264038, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8589357137680054, "num_tokens": 165356643.0, "step": 4351 }, { "epoch": 0.5536191324259, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.499540328979492, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8520034551620483, "num_tokens": 165396897.0, "step": 4352 }, { "epoch": 0.5537463427044905, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.914548397064209, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8626046776771545, "num_tokens": 165434409.0, "step": 4353 }, { "epoch": 0.553873552983081, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6709706783294678, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8577563762664795, "num_tokens": 165474674.0, "step": 4354 }, { "epoch": 0.5540007632616716, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00011777877807617188, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.952331304550171, "learning_rate": 1e-05, "loss": 0.5195, "mean_token_accuracy": 0.8455742597579956, "num_tokens": 165511865.0, "step": 4355 }, { "epoch": 0.554127973540262, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8371121883392334, "learning_rate": 1e-05, "loss": 0.5095, "mean_token_accuracy": 0.8558157682418823, "num_tokens": 165548589.0, "step": 4356 }, { "epoch": 0.5542551838188525, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.0781681537628174, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8623318076133728, "num_tokens": 165589553.0, "step": 4357 }, { "epoch": 0.554382394097443, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.921631097793579, "learning_rate": 1e-05, "loss": 0.5862, "mean_token_accuracy": 0.8328769207000732, "num_tokens": 165625499.0, "step": 4358 }, { "epoch": 0.5545096043760336, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.286558151245117, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8588982820510864, "num_tokens": 165663967.0, "step": 4359 }, { "epoch": 0.5546368146546241, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.811842441558838, "learning_rate": 1e-05, "loss": 0.5227, "mean_token_accuracy": 0.8521006107330322, "num_tokens": 165706382.0, "step": 4360 }, { "epoch": 0.5547640249332146, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.386503219604492, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8596961498260498, "num_tokens": 165742473.0, "step": 4361 }, { "epoch": 0.5548912352118052, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.643526554107666, "learning_rate": 1e-05, "loss": 0.4996, "mean_token_accuracy": 0.8511249423027039, "num_tokens": 165776329.0, "step": 4362 }, { "epoch": 0.5550184454903956, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.374373435974121, "learning_rate": 1e-05, "loss": 0.505, "mean_token_accuracy": 0.8534163236618042, "num_tokens": 165812887.0, "step": 4363 }, { "epoch": 0.5551456557689861, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4581458568573, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.867871105670929, "num_tokens": 165848933.0, "step": 4364 }, { "epoch": 0.5552728660475766, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3638246059417725, "learning_rate": 1e-05, "loss": 0.5327, "mean_token_accuracy": 0.8433284759521484, "num_tokens": 165887158.0, "step": 4365 }, { "epoch": 0.5554000763261672, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.300621271133423, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8560736179351807, "num_tokens": 165928214.0, "step": 4366 }, { "epoch": 0.5555272866047577, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2807748317718506, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8548770546913147, "num_tokens": 165964838.0, "step": 4367 }, { "epoch": 0.5556544968833482, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.222175121307373, "learning_rate": 1e-05, "loss": 0.5149, "mean_token_accuracy": 0.847784698009491, "num_tokens": 165996097.0, "step": 4368 }, { "epoch": 0.5557817071619386, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2459793090820312, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8555288910865784, "num_tokens": 166028660.0, "step": 4369 }, { "epoch": 0.5559089174405292, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.164161443710327, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8555761575698853, "num_tokens": 166065259.0, "step": 4370 }, { "epoch": 0.5560361277191197, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.205920934677124, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8522424697875977, "num_tokens": 166105757.0, "step": 4371 }, { "epoch": 0.5561633379977102, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.134516954421997, "learning_rate": 1e-05, "loss": 0.5165, "mean_token_accuracy": 0.8414815068244934, "num_tokens": 166142362.0, "step": 4372 }, { "epoch": 0.5562905482763008, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.18198299407959, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.857252299785614, "num_tokens": 166179437.0, "step": 4373 }, { "epoch": 0.5564177585548913, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.107790470123291, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8500617742538452, "num_tokens": 166218555.0, "step": 4374 }, { "epoch": 0.5565449688334817, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1906144618988037, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8664723634719849, "num_tokens": 166252297.0, "step": 4375 }, { "epoch": 0.5566721791120722, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0795254707336426, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8613830208778381, "num_tokens": 166289597.0, "step": 4376 }, { "epoch": 0.5567993893906628, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.2063066959381104, "learning_rate": 1e-05, "loss": 0.5104, "mean_token_accuracy": 0.8468446731567383, "num_tokens": 166328655.0, "step": 4377 }, { "epoch": 0.5569265996692533, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0356013774871826, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8546059131622314, "num_tokens": 166371825.0, "step": 4378 }, { "epoch": 0.5570538099478438, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2599306106567383, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8560099601745605, "num_tokens": 166407084.0, "step": 4379 }, { "epoch": 0.5571810202264343, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.98480224609375, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.863070547580719, "num_tokens": 166442272.0, "step": 4380 }, { "epoch": 0.5573082305050248, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3636415004730225, "learning_rate": 1e-05, "loss": 0.5109, "mean_token_accuracy": 0.8479295372962952, "num_tokens": 166475015.0, "step": 4381 }, { "epoch": 0.5574354407836153, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.8882503509521484, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8589106202125549, "num_tokens": 166512421.0, "step": 4382 }, { "epoch": 0.5575626510622058, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.537659168243408, "learning_rate": 1e-05, "loss": 0.5058, "mean_token_accuracy": 0.8489854335784912, "num_tokens": 166554248.0, "step": 4383 }, { "epoch": 0.5576898613407963, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.8154327869415283, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.873428225517273, "num_tokens": 166590664.0, "step": 4384 }, { "epoch": 0.5578170716193869, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8320364952087402, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8552829027175903, "num_tokens": 166625615.0, "step": 4385 }, { "epoch": 0.5579442818979774, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.8839125633239746, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.861292839050293, "num_tokens": 166662699.0, "step": 4386 }, { "epoch": 0.5580714921765678, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.141963958740234, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.863214373588562, "num_tokens": 166701218.0, "step": 4387 }, { "epoch": 0.5581987024551583, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.2425293922424316, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8546518087387085, "num_tokens": 166739699.0, "step": 4388 }, { "epoch": 0.5583259127337489, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.189474582672119, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8640690445899963, "num_tokens": 166782275.0, "step": 4389 }, { "epoch": 0.5584531230123394, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.5896518230438232, "learning_rate": 1e-05, "loss": 0.4973, "mean_token_accuracy": 0.86202073097229, "num_tokens": 166820535.0, "step": 4390 }, { "epoch": 0.5585803332909299, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9415667057037354, "learning_rate": 1e-05, "loss": 0.5371, "mean_token_accuracy": 0.8424530625343323, "num_tokens": 166865450.0, "step": 4391 }, { "epoch": 0.5587075435695205, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.599015474319458, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8611205816268921, "num_tokens": 166905838.0, "step": 4392 }, { "epoch": 0.5588347538481109, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7272286415100098, "learning_rate": 1e-05, "loss": 0.539, "mean_token_accuracy": 0.842821478843689, "num_tokens": 166943437.0, "step": 4393 }, { "epoch": 0.5589619641267014, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4809155464172363, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8623255491256714, "num_tokens": 166980863.0, "step": 4394 }, { "epoch": 0.5590891744052919, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.566070795059204, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8721094131469727, "num_tokens": 167017861.0, "step": 4395 }, { "epoch": 0.5592163846838825, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3490660190582275, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8684013485908508, "num_tokens": 167051779.0, "step": 4396 }, { "epoch": 0.559343594962473, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.464710235595703, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8572884202003479, "num_tokens": 167095205.0, "step": 4397 }, { "epoch": 0.5594708052410635, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2384142875671387, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8512551784515381, "num_tokens": 167130594.0, "step": 4398 }, { "epoch": 0.5595980155196539, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.440613031387329, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8570474982261658, "num_tokens": 167170816.0, "step": 4399 }, { "epoch": 0.5597252257982445, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.127776861190796, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8674992918968201, "num_tokens": 167205301.0, "step": 4400 }, { "epoch": 0.559852436076835, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.481733798980713, "learning_rate": 1e-05, "loss": 0.5241, "mean_token_accuracy": 0.8419384956359863, "num_tokens": 167240356.0, "step": 4401 }, { "epoch": 0.5599796463554255, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.046034097671509, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8633251190185547, "num_tokens": 167269853.0, "step": 4402 }, { "epoch": 0.560106856634016, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5732364654541016, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8611345887184143, "num_tokens": 167315398.0, "step": 4403 }, { "epoch": 0.5602340669126066, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.991126537322998, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8599458932876587, "num_tokens": 167352715.0, "step": 4404 }, { "epoch": 0.560361277191197, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.737114191055298, "learning_rate": 1e-05, "loss": 0.5436, "mean_token_accuracy": 0.8434770703315735, "num_tokens": 167391646.0, "step": 4405 }, { "epoch": 0.5604884874697875, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000118255615234375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.034388303756714, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8546537160873413, "num_tokens": 167431676.0, "step": 4406 }, { "epoch": 0.560615697748378, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8769006729125977, "learning_rate": 1e-05, "loss": 0.5157, "mean_token_accuracy": 0.8510510921478271, "num_tokens": 167469121.0, "step": 4407 }, { "epoch": 0.5607429080269686, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2135345935821533, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8617209196090698, "num_tokens": 167511721.0, "step": 4408 }, { "epoch": 0.5608701183055591, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8549351692199707, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.865706205368042, "num_tokens": 167546965.0, "step": 4409 }, { "epoch": 0.5609973285841496, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.388962984085083, "learning_rate": 1e-05, "loss": 0.5332, "mean_token_accuracy": 0.8449573516845703, "num_tokens": 167584968.0, "step": 4410 }, { "epoch": 0.5611245388627402, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6504905223846436, "learning_rate": 1e-05, "loss": 0.5385, "mean_token_accuracy": 0.8444964289665222, "num_tokens": 167623124.0, "step": 4411 }, { "epoch": 0.5612517491413306, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.407115936279297, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.8534737825393677, "num_tokens": 167664877.0, "step": 4412 }, { "epoch": 0.5613789594199211, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4313132762908936, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8508163690567017, "num_tokens": 167706154.0, "step": 4413 }, { "epoch": 0.5615061696985116, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.391730308532715, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8542481660842896, "num_tokens": 167740081.0, "step": 4414 }, { "epoch": 0.5616333799771022, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2350234985351562, "learning_rate": 1e-05, "loss": 0.5464, "mean_token_accuracy": 0.8382365107536316, "num_tokens": 167783324.0, "step": 4415 }, { "epoch": 0.5617605902556927, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.373915433883667, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8597217798233032, "num_tokens": 167823216.0, "step": 4416 }, { "epoch": 0.5618878005342832, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.111253499984741, "learning_rate": 1e-05, "loss": 0.5284, "mean_token_accuracy": 0.8391814231872559, "num_tokens": 167857452.0, "step": 4417 }, { "epoch": 0.5620150108128736, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3850789070129395, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8627285361289978, "num_tokens": 167896402.0, "step": 4418 }, { "epoch": 0.5621422210914642, "ewc_loss": 0.052001953125, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0119829177856445, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8590037822723389, "num_tokens": 167936441.0, "step": 4419 }, { "epoch": 0.5622694313700547, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00011920928955078125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.468029499053955, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8665719032287598, "num_tokens": 167970786.0, "step": 4420 }, { "epoch": 0.5623966416486452, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.9552221298217773, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8555417060852051, "num_tokens": 168011004.0, "step": 4421 }, { "epoch": 0.5625238519272358, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5890300273895264, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8552809953689575, "num_tokens": 168053418.0, "step": 4422 }, { "epoch": 0.5626510622058263, "ewc_loss": 0.052490234375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9559266567230225, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8633155822753906, "num_tokens": 168088156.0, "step": 4423 }, { "epoch": 0.5627782724844167, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.729315757751465, "learning_rate": 1e-05, "loss": 0.5228, "mean_token_accuracy": 0.8474917411804199, "num_tokens": 168132329.0, "step": 4424 }, { "epoch": 0.5629054827630072, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.00467586517334, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8615535497665405, "num_tokens": 168167882.0, "step": 4425 }, { "epoch": 0.5630326930415978, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8628153800964355, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8580020666122437, "num_tokens": 168203265.0, "step": 4426 }, { "epoch": 0.5631599033201883, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.1601524353027344, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8688411116600037, "num_tokens": 168241410.0, "step": 4427 }, { "epoch": 0.5632871135987788, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.85170578956604, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8627039194107056, "num_tokens": 168285412.0, "step": 4428 }, { "epoch": 0.5634143238773693, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3083996772766113, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.85075443983078, "num_tokens": 168324628.0, "step": 4429 }, { "epoch": 0.5635415341559598, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7377889156341553, "learning_rate": 1e-05, "loss": 0.5263, "mean_token_accuracy": 0.8508197069168091, "num_tokens": 168359759.0, "step": 4430 }, { "epoch": 0.5636687444345503, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3683247566223145, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8535361886024475, "num_tokens": 168393904.0, "step": 4431 }, { "epoch": 0.5637959547131408, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5512290000915527, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8519184589385986, "num_tokens": 168429718.0, "step": 4432 }, { "epoch": 0.5639231649917313, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.33788800239563, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8670704364776611, "num_tokens": 168462950.0, "step": 4433 }, { "epoch": 0.5640503752703219, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3795526027679443, "learning_rate": 1e-05, "loss": 0.4956, "mean_token_accuracy": 0.857039213180542, "num_tokens": 168504095.0, "step": 4434 }, { "epoch": 0.5641775855489124, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.294217109680176, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8531261682510376, "num_tokens": 168539931.0, "step": 4435 }, { "epoch": 0.5643047958275028, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.278371810913086, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8624658584594727, "num_tokens": 168573064.0, "step": 4436 }, { "epoch": 0.5644320061060933, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.243283987045288, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.856246829032898, "num_tokens": 168607997.0, "step": 4437 }, { "epoch": 0.5645592163846839, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.205024003982544, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8771877288818359, "num_tokens": 168647590.0, "step": 4438 }, { "epoch": 0.5646864266632744, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.209585666656494, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8612647652626038, "num_tokens": 168681347.0, "step": 4439 }, { "epoch": 0.5648136369418649, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1676900386810303, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8554216623306274, "num_tokens": 168722830.0, "step": 4440 }, { "epoch": 0.5649408472204555, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1770870685577393, "learning_rate": 1e-05, "loss": 0.5355, "mean_token_accuracy": 0.8409808278083801, "num_tokens": 168762381.0, "step": 4441 }, { "epoch": 0.5650680574990459, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1464405059814453, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8605331182479858, "num_tokens": 168802832.0, "step": 4442 }, { "epoch": 0.5651952677776364, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1634433269500732, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8554947376251221, "num_tokens": 168837780.0, "step": 4443 }, { "epoch": 0.5653224780562269, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.145566701889038, "learning_rate": 1e-05, "loss": 0.5262, "mean_token_accuracy": 0.8462618589401245, "num_tokens": 168871830.0, "step": 4444 }, { "epoch": 0.5654496883348175, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1426167488098145, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8626857995986938, "num_tokens": 168912477.0, "step": 4445 }, { "epoch": 0.565576898613408, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.1296303272247314, "learning_rate": 1e-05, "loss": 0.5099, "mean_token_accuracy": 0.8484073877334595, "num_tokens": 168948976.0, "step": 4446 }, { "epoch": 0.5657041088919985, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1319327354431152, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.856537401676178, "num_tokens": 168992454.0, "step": 4447 }, { "epoch": 0.5658313191705889, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.1447269916534424, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8533484935760498, "num_tokens": 169023135.0, "step": 4448 }, { "epoch": 0.5659585294491795, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.115906000137329, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.85789954662323, "num_tokens": 169062066.0, "step": 4449 }, { "epoch": 0.56608573972777, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.150247097015381, "learning_rate": 1e-05, "loss": 0.5408, "mean_token_accuracy": 0.837962806224823, "num_tokens": 169100317.0, "step": 4450 }, { "epoch": 0.5662129500063605, "ewc_loss": 0.052490234375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.1096925735473633, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8627962470054626, "num_tokens": 169128474.0, "step": 4451 }, { "epoch": 0.566340160284951, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1460390090942383, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8733882904052734, "num_tokens": 169164905.0, "step": 4452 }, { "epoch": 0.5664673705635416, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0933079719543457, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8531568050384521, "num_tokens": 169201153.0, "step": 4453 }, { "epoch": 0.566594580842132, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.146874189376831, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8709020614624023, "num_tokens": 169235325.0, "step": 4454 }, { "epoch": 0.5667217911207225, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001201629638671875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.081188201904297, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8753093481063843, "num_tokens": 169276690.0, "step": 4455 }, { "epoch": 0.566849001399313, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.169443368911743, "learning_rate": 1e-05, "loss": 0.5131, "mean_token_accuracy": 0.843338131904602, "num_tokens": 169315751.0, "step": 4456 }, { "epoch": 0.5669762116779036, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.05910062789917, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8584707975387573, "num_tokens": 169354626.0, "step": 4457 }, { "epoch": 0.5671034219564941, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2175683975219727, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8556576371192932, "num_tokens": 169390255.0, "step": 4458 }, { "epoch": 0.5672306322350846, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9949347972869873, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8482733964920044, "num_tokens": 169431912.0, "step": 4459 }, { "epoch": 0.5673578425136752, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.295295238494873, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8808952569961548, "num_tokens": 169467803.0, "step": 4460 }, { "epoch": 0.5674850527922656, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9162678718566895, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8475620746612549, "num_tokens": 169512252.0, "step": 4461 }, { "epoch": 0.5676122630708561, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4529058933258057, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8619043827056885, "num_tokens": 169547632.0, "step": 4462 }, { "epoch": 0.5677394733494466, "ewc_loss": 0.05078125, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.824178695678711, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8559447526931763, "num_tokens": 169589355.0, "step": 4463 }, { "epoch": 0.5678666836280372, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00012111663818359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7192275524139404, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8640076518058777, "num_tokens": 169627441.0, "step": 4464 }, { "epoch": 0.5679938939066277, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00012063980102539062, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.8076486587524414, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8580888509750366, "num_tokens": 169659752.0, "step": 4465 }, { "epoch": 0.5681211041852182, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00012111663818359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.095186233520508, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.873805046081543, "num_tokens": 169697197.0, "step": 4466 }, { "epoch": 0.5682483144638086, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00012111663818359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.0734102725982666, "learning_rate": 1e-05, "loss": 0.5495, "mean_token_accuracy": 0.8379005789756775, "num_tokens": 169732523.0, "step": 4467 }, { "epoch": 0.5683755247423992, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00012111663818359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.332393169403076, "learning_rate": 1e-05, "loss": 0.5201, "mean_token_accuracy": 0.8546874523162842, "num_tokens": 169771494.0, "step": 4468 }, { "epoch": 0.5685027350209897, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00012111663818359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.571556329727173, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.873399019241333, "num_tokens": 169809835.0, "step": 4469 }, { "epoch": 0.5686299452995802, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00012111663818359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.132153511047363, "learning_rate": 1e-05, "loss": 0.5196, "mean_token_accuracy": 0.8511524796485901, "num_tokens": 169853276.0, "step": 4470 }, { "epoch": 0.5687571555781707, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00012111663818359375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7295992374420166, "learning_rate": 1e-05, "loss": 0.5381, "mean_token_accuracy": 0.8410686254501343, "num_tokens": 169892106.0, "step": 4471 }, { "epoch": 0.5688843658567613, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8479018211364746, "learning_rate": 1e-05, "loss": 0.5163, "mean_token_accuracy": 0.8489360809326172, "num_tokens": 169933635.0, "step": 4472 }, { "epoch": 0.5690115761353517, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5911877155303955, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8557336926460266, "num_tokens": 169971694.0, "step": 4473 }, { "epoch": 0.5691387864139422, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6737756729125977, "learning_rate": 1e-05, "loss": 0.5407, "mean_token_accuracy": 0.8401271104812622, "num_tokens": 170013285.0, "step": 4474 }, { "epoch": 0.5692659966925327, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4279377460479736, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8576181530952454, "num_tokens": 170052831.0, "step": 4475 }, { "epoch": 0.5693932069711233, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5736072063446045, "learning_rate": 1e-05, "loss": 0.5322, "mean_token_accuracy": 0.8435431718826294, "num_tokens": 170089859.0, "step": 4476 }, { "epoch": 0.5695204172497138, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.293790102005005, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.851611316204071, "num_tokens": 170126494.0, "step": 4477 }, { "epoch": 0.5696476275283043, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5280377864837646, "learning_rate": 1e-05, "loss": 0.5103, "mean_token_accuracy": 0.8490958213806152, "num_tokens": 170167744.0, "step": 4478 }, { "epoch": 0.5697748378068948, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1670758724212646, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8556493520736694, "num_tokens": 170203988.0, "step": 4479 }, { "epoch": 0.5699020480854853, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.57033109664917, "learning_rate": 1e-05, "loss": 0.5559, "mean_token_accuracy": 0.8341635465621948, "num_tokens": 170243231.0, "step": 4480 }, { "epoch": 0.5700292583640758, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.081516742706299, "learning_rate": 1e-05, "loss": 0.5409, "mean_token_accuracy": 0.8377136588096619, "num_tokens": 170280637.0, "step": 4481 }, { "epoch": 0.5701564686426663, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.691549777984619, "learning_rate": 1e-05, "loss": 0.5262, "mean_token_accuracy": 0.8480193614959717, "num_tokens": 170317106.0, "step": 4482 }, { "epoch": 0.5702836789212569, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.051528215408325, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.8533585667610168, "num_tokens": 170353374.0, "step": 4483 }, { "epoch": 0.5704108891998474, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.8426003456115723, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8622419238090515, "num_tokens": 170386968.0, "step": 4484 }, { "epoch": 0.5705380994784378, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.160123109817505, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8622620701789856, "num_tokens": 170425386.0, "step": 4485 }, { "epoch": 0.5706653097570283, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8898942470550537, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8592206239700317, "num_tokens": 170467828.0, "step": 4486 }, { "epoch": 0.5707925200356189, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3504209518432617, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8546306490898132, "num_tokens": 170506333.0, "step": 4487 }, { "epoch": 0.5709197303142094, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.762864351272583, "learning_rate": 1e-05, "loss": 0.5249, "mean_token_accuracy": 0.8522838354110718, "num_tokens": 170542352.0, "step": 4488 }, { "epoch": 0.5710469405927999, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4373652935028076, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8713737726211548, "num_tokens": 170574251.0, "step": 4489 }, { "epoch": 0.5711741508713905, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5238184928894043, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8685867786407471, "num_tokens": 170610294.0, "step": 4490 }, { "epoch": 0.5713013611499809, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4095382690429688, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8620153069496155, "num_tokens": 170649245.0, "step": 4491 }, { "epoch": 0.5714285714285714, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3150906562805176, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8525491952896118, "num_tokens": 170692106.0, "step": 4492 }, { "epoch": 0.5715557817071619, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3745033740997314, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8638885617256165, "num_tokens": 170730319.0, "step": 4493 }, { "epoch": 0.5716829919857525, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.178178310394287, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8640320301055908, "num_tokens": 170763805.0, "step": 4494 }, { "epoch": 0.571810202264343, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3507192134857178, "learning_rate": 1e-05, "loss": 0.5378, "mean_token_accuracy": 0.8407259583473206, "num_tokens": 170804966.0, "step": 4495 }, { "epoch": 0.5719374125429335, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0627005100250244, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8613384962081909, "num_tokens": 170845652.0, "step": 4496 }, { "epoch": 0.5720646228215239, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3895251750946045, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8624635934829712, "num_tokens": 170883449.0, "step": 4497 }, { "epoch": 0.5721918331001145, "ewc_loss": 0.051513671875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9929986000061035, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.8485296964645386, "num_tokens": 170920187.0, "step": 4498 }, { "epoch": 0.572319043378705, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4817442893981934, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.8513891696929932, "num_tokens": 170952650.0, "step": 4499 }, { "epoch": 0.5724462536572955, "ewc_loss": 0.0517578125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9283671379089355, "learning_rate": 1e-05, "loss": 0.4949, "mean_token_accuracy": 0.8519642949104309, "num_tokens": 170989130.0, "step": 4500 }, { "epoch": 0.572573463935886, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6283912658691406, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8597869873046875, "num_tokens": 171030713.0, "step": 4501 }, { "epoch": 0.5727006742144766, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00012159347534179688, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9225358963012695, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.857819676399231, "num_tokens": 171066363.0, "step": 4502 }, { "epoch": 0.572827884493067, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8324079513549805, "learning_rate": 1e-05, "loss": 0.5295, "mean_token_accuracy": 0.8518328070640564, "num_tokens": 171108290.0, "step": 4503 }, { "epoch": 0.5729550947716575, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0546112060546875, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8823130130767822, "num_tokens": 171141990.0, "step": 4504 }, { "epoch": 0.573082305050248, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.946046829223633, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8589833974838257, "num_tokens": 171180032.0, "step": 4505 }, { "epoch": 0.5732095153288386, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.285236358642578, "learning_rate": 1e-05, "loss": 0.549, "mean_token_accuracy": 0.8396071195602417, "num_tokens": 171218918.0, "step": 4506 }, { "epoch": 0.5733367256074291, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8542938232421875, "learning_rate": 1e-05, "loss": 0.5143, "mean_token_accuracy": 0.8564318418502808, "num_tokens": 171261028.0, "step": 4507 }, { "epoch": 0.5734639358860196, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.39798641204834, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.8597913980484009, "num_tokens": 171301477.0, "step": 4508 }, { "epoch": 0.5735911461646102, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6740424633026123, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8698431253433228, "num_tokens": 171338937.0, "step": 4509 }, { "epoch": 0.5737183564432006, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4032022953033447, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.8549909591674805, "num_tokens": 171375058.0, "step": 4510 }, { "epoch": 0.5738455667217911, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4983835220336914, "learning_rate": 1e-05, "loss": 0.5154, "mean_token_accuracy": 0.8481436371803284, "num_tokens": 171410282.0, "step": 4511 }, { "epoch": 0.5739727770003816, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3540852069854736, "learning_rate": 1e-05, "loss": 0.4943, "mean_token_accuracy": 0.8542277216911316, "num_tokens": 171442864.0, "step": 4512 }, { "epoch": 0.5740999872789722, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3534655570983887, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8672006130218506, "num_tokens": 171476267.0, "step": 4513 }, { "epoch": 0.5742271975575627, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2820096015930176, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.854600727558136, "num_tokens": 171513173.0, "step": 4514 }, { "epoch": 0.5743544078361532, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2657854557037354, "learning_rate": 1e-05, "loss": 0.5678, "mean_token_accuracy": 0.8346309661865234, "num_tokens": 171555071.0, "step": 4515 }, { "epoch": 0.5744816181147436, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.21954607963562, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8561026453971863, "num_tokens": 171594263.0, "step": 4516 }, { "epoch": 0.5746088283933342, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2153055667877197, "learning_rate": 1e-05, "loss": 0.5097, "mean_token_accuracy": 0.8505586981773376, "num_tokens": 171635238.0, "step": 4517 }, { "epoch": 0.5747360386719247, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.185819387435913, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8512577414512634, "num_tokens": 171674537.0, "step": 4518 }, { "epoch": 0.5748632489505152, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1733837127685547, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8693118095397949, "num_tokens": 171718606.0, "step": 4519 }, { "epoch": 0.5749904592291057, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1553125381469727, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8606600165367126, "num_tokens": 171757092.0, "step": 4520 }, { "epoch": 0.5751176695076963, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1607439517974854, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8595780730247498, "num_tokens": 171799606.0, "step": 4521 }, { "epoch": 0.5752448797862867, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.1462771892547607, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8641709685325623, "num_tokens": 171834686.0, "step": 4522 }, { "epoch": 0.5753720900648772, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1647067070007324, "learning_rate": 1e-05, "loss": 0.5484, "mean_token_accuracy": 0.8393182754516602, "num_tokens": 171870443.0, "step": 4523 }, { "epoch": 0.5754993003434677, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.112339496612549, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8604333400726318, "num_tokens": 171907615.0, "step": 4524 }, { "epoch": 0.5756265106220583, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1784250736236572, "learning_rate": 1e-05, "loss": 0.513, "mean_token_accuracy": 0.8446900844573975, "num_tokens": 171943120.0, "step": 4525 }, { "epoch": 0.5757537209006488, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.079387903213501, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8628612756729126, "num_tokens": 171980637.0, "step": 4526 }, { "epoch": 0.5758809311792393, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.2063846588134766, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8584657907485962, "num_tokens": 172014232.0, "step": 4527 }, { "epoch": 0.5760081414578297, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.045646905899048, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8562681674957275, "num_tokens": 172049639.0, "step": 4528 }, { "epoch": 0.5761353517364203, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2477807998657227, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8583335876464844, "num_tokens": 172086482.0, "step": 4529 }, { "epoch": 0.5762625620150108, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9969441890716553, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8637827634811401, "num_tokens": 172117220.0, "step": 4530 }, { "epoch": 0.5763897722936013, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3498682975769043, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8606964349746704, "num_tokens": 172149925.0, "step": 4531 }, { "epoch": 0.5765169825721919, "ewc_loss": 0.051513671875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.9127166271209717, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8680195808410645, "num_tokens": 172186903.0, "step": 4532 }, { "epoch": 0.5766441928507824, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.5047805309295654, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8529877662658691, "num_tokens": 172226573.0, "step": 4533 }, { "epoch": 0.5767714031293728, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.844661235809326, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8617228269577026, "num_tokens": 172263639.0, "step": 4534 }, { "epoch": 0.5768986134079633, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7532904148101807, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8660029172897339, "num_tokens": 172297740.0, "step": 4535 }, { "epoch": 0.5770258236865539, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.8468754291534424, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.8483678102493286, "num_tokens": 172338243.0, "step": 4536 }, { "epoch": 0.5771530339651444, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.075020790100098, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8569607734680176, "num_tokens": 172383201.0, "step": 4537 }, { "epoch": 0.5772802442437349, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.1410629749298096, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8682569265365601, "num_tokens": 172420587.0, "step": 4538 }, { "epoch": 0.5774074545223254, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.219675540924072, "learning_rate": 1e-05, "loss": 0.5495, "mean_token_accuracy": 0.8415896892547607, "num_tokens": 172459194.0, "step": 4539 }, { "epoch": 0.5775346648009159, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.559497117996216, "learning_rate": 1e-05, "loss": 0.5444, "mean_token_accuracy": 0.8384876251220703, "num_tokens": 172492463.0, "step": 4540 }, { "epoch": 0.5776618750795064, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.991415023803711, "learning_rate": 1e-05, "loss": 0.5365, "mean_token_accuracy": 0.8449809551239014, "num_tokens": 172531675.0, "step": 4541 }, { "epoch": 0.5777890853580969, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5972375869750977, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8565332889556885, "num_tokens": 172571666.0, "step": 4542 }, { "epoch": 0.5779162956366874, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7774367332458496, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.857688844203949, "num_tokens": 172608132.0, "step": 4543 }, { "epoch": 0.578043505915278, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4873580932617188, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8592236638069153, "num_tokens": 172650881.0, "step": 4544 }, { "epoch": 0.5781707161938685, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6217901706695557, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8613691329956055, "num_tokens": 172684632.0, "step": 4545 }, { "epoch": 0.5782979264724589, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.36295223236084, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8645485639572144, "num_tokens": 172724804.0, "step": 4546 }, { "epoch": 0.5784251367510495, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5231006145477295, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8544996380805969, "num_tokens": 172762687.0, "step": 4547 }, { "epoch": 0.57855234702964, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.243295907974243, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8657204508781433, "num_tokens": 172802885.0, "step": 4548 }, { "epoch": 0.5786795573082305, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.502319574356079, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8598041534423828, "num_tokens": 172836158.0, "step": 4549 }, { "epoch": 0.578806767586821, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1495349407196045, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.857731282711029, "num_tokens": 172872319.0, "step": 4550 }, { "epoch": 0.5789339778654116, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5425286293029785, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8631707429885864, "num_tokens": 172906803.0, "step": 4551 }, { "epoch": 0.579061188144002, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.060518741607666, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8593999147415161, "num_tokens": 172944095.0, "step": 4552 }, { "epoch": 0.5791883984225925, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6426825523376465, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8631647825241089, "num_tokens": 172980026.0, "step": 4553 }, { "epoch": 0.579315608701183, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0331852436065674, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.8497565984725952, "num_tokens": 173019558.0, "step": 4554 }, { "epoch": 0.5794428189797736, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.789156198501587, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.861023485660553, "num_tokens": 173059067.0, "step": 4555 }, { "epoch": 0.5795700292583641, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.126678466796875, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8664365410804749, "num_tokens": 173096277.0, "step": 4556 }, { "epoch": 0.5796972395369546, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.853055715560913, "learning_rate": 1e-05, "loss": 0.5349, "mean_token_accuracy": 0.8454977869987488, "num_tokens": 173134746.0, "step": 4557 }, { "epoch": 0.5798244498155452, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3085951805114746, "learning_rate": 1e-05, "loss": 0.5547, "mean_token_accuracy": 0.8369655609130859, "num_tokens": 173164786.0, "step": 4558 }, { "epoch": 0.5799516600941356, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7301185131073, "learning_rate": 1e-05, "loss": 0.5176, "mean_token_accuracy": 0.8504445552825928, "num_tokens": 173202948.0, "step": 4559 }, { "epoch": 0.5800788703727261, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3885247707366943, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8509387373924255, "num_tokens": 173235253.0, "step": 4560 }, { "epoch": 0.5802060806513166, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.509014844894409, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8526743650436401, "num_tokens": 173279878.0, "step": 4561 }, { "epoch": 0.5803332909299072, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.379804849624634, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8595275282859802, "num_tokens": 173316719.0, "step": 4562 }, { "epoch": 0.5804605012084977, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.3296799659729004, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8631147146224976, "num_tokens": 173347077.0, "step": 4563 }, { "epoch": 0.5805877114870882, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3534128665924072, "learning_rate": 1e-05, "loss": 0.5191, "mean_token_accuracy": 0.8483104109764099, "num_tokens": 173386891.0, "step": 4564 }, { "epoch": 0.5807149217656786, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1569182872772217, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8585492372512817, "num_tokens": 173426919.0, "step": 4565 }, { "epoch": 0.5808421320442692, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3648793697357178, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.86406010389328, "num_tokens": 173461380.0, "step": 4566 }, { "epoch": 0.5809693423228597, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0422914028167725, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8585139513015747, "num_tokens": 173504680.0, "step": 4567 }, { "epoch": 0.5810965526014502, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4200057983398438, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8582499027252197, "num_tokens": 173541229.0, "step": 4568 }, { "epoch": 0.5812237628800407, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.973895788192749, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8539226055145264, "num_tokens": 173573805.0, "step": 4569 }, { "epoch": 0.5813509731586313, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5716657638549805, "learning_rate": 1e-05, "loss": 0.528, "mean_token_accuracy": 0.8469735383987427, "num_tokens": 173613899.0, "step": 4570 }, { "epoch": 0.5814781834372217, "ewc_loss": 0.052490234375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.9141409397125244, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8604532480239868, "num_tokens": 173650590.0, "step": 4571 }, { "epoch": 0.5816053937158122, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.74279522895813, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8593010306358337, "num_tokens": 173684067.0, "step": 4572 }, { "epoch": 0.5817326039944027, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.9605047702789307, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8622409701347351, "num_tokens": 173721846.0, "step": 4573 }, { "epoch": 0.5818598142729933, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.941819190979004, "learning_rate": 1e-05, "loss": 0.5432, "mean_token_accuracy": 0.840551495552063, "num_tokens": 173754616.0, "step": 4574 }, { "epoch": 0.5819870245515838, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.1846628189086914, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8532720804214478, "num_tokens": 173790208.0, "step": 4575 }, { "epoch": 0.5821142348301743, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9373064041137695, "learning_rate": 1e-05, "loss": 0.5398, "mean_token_accuracy": 0.8480300307273865, "num_tokens": 173833132.0, "step": 4576 }, { "epoch": 0.5822414451087647, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3872008323669434, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8495489358901978, "num_tokens": 173870526.0, "step": 4577 }, { "epoch": 0.5823686553873553, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7819907665252686, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.852653443813324, "num_tokens": 173901561.0, "step": 4578 }, { "epoch": 0.5824958656659458, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.4347617626190186, "learning_rate": 1e-05, "loss": 0.5298, "mean_token_accuracy": 0.848041296005249, "num_tokens": 173941197.0, "step": 4579 }, { "epoch": 0.5826230759445363, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.574794054031372, "learning_rate": 1e-05, "loss": 0.5021, "mean_token_accuracy": 0.8546332716941833, "num_tokens": 173981755.0, "step": 4580 }, { "epoch": 0.5827502862231269, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3888099193573, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8519959449768066, "num_tokens": 174015831.0, "step": 4581 }, { "epoch": 0.5828774965017174, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4088339805603027, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8634070158004761, "num_tokens": 174051623.0, "step": 4582 }, { "epoch": 0.5830047067803078, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.321866512298584, "learning_rate": 1e-05, "loss": 0.5284, "mean_token_accuracy": 0.8438838124275208, "num_tokens": 174086422.0, "step": 4583 }, { "epoch": 0.5831319170588983, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.289588212966919, "learning_rate": 1e-05, "loss": 0.5045, "mean_token_accuracy": 0.8517988920211792, "num_tokens": 174128557.0, "step": 4584 }, { "epoch": 0.5832591273374889, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2609825134277344, "learning_rate": 1e-05, "loss": 0.5134, "mean_token_accuracy": 0.8481468558311462, "num_tokens": 174169704.0, "step": 4585 }, { "epoch": 0.5833863376160794, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2134387493133545, "learning_rate": 1e-05, "loss": 0.5195, "mean_token_accuracy": 0.8452574610710144, "num_tokens": 174213975.0, "step": 4586 }, { "epoch": 0.5835135478946699, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.216146230697632, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8546261191368103, "num_tokens": 174254425.0, "step": 4587 }, { "epoch": 0.5836407581732604, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.197521209716797, "learning_rate": 1e-05, "loss": 0.5146, "mean_token_accuracy": 0.8494110107421875, "num_tokens": 174292411.0, "step": 4588 }, { "epoch": 0.5837679684518509, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1710729598999023, "learning_rate": 1e-05, "loss": 0.5419, "mean_token_accuracy": 0.8390560150146484, "num_tokens": 174335103.0, "step": 4589 }, { "epoch": 0.5838951787304414, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1750411987304688, "learning_rate": 1e-05, "loss": 0.5243, "mean_token_accuracy": 0.8473299145698547, "num_tokens": 174377984.0, "step": 4590 }, { "epoch": 0.5840223890090319, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.151036024093628, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8545678853988647, "num_tokens": 174408211.0, "step": 4591 }, { "epoch": 0.5841495992876224, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1703567504882812, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8416743278503418, "num_tokens": 174444137.0, "step": 4592 }, { "epoch": 0.584276809566213, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.112940549850464, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8486644625663757, "num_tokens": 174488382.0, "step": 4593 }, { "epoch": 0.5844040198448035, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1731581687927246, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8710162043571472, "num_tokens": 174524436.0, "step": 4594 }, { "epoch": 0.5845312301233939, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0904765129089355, "learning_rate": 1e-05, "loss": 0.5303, "mean_token_accuracy": 0.8446499705314636, "num_tokens": 174563830.0, "step": 4595 }, { "epoch": 0.5846584404019844, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.206167221069336, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8556405305862427, "num_tokens": 174598600.0, "step": 4596 }, { "epoch": 0.584785650680575, "ewc_loss": 0.052490234375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.054901599884033, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8643466234207153, "num_tokens": 174631234.0, "step": 4597 }, { "epoch": 0.5849128609591655, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.243370771408081, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8547207713127136, "num_tokens": 174673933.0, "step": 4598 }, { "epoch": 0.585040071237756, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.993640899658203, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8578937649726868, "num_tokens": 174715801.0, "step": 4599 }, { "epoch": 0.5851672815163466, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.31791090965271, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8555184602737427, "num_tokens": 174757309.0, "step": 4600 }, { "epoch": 0.585294491794937, "ewc_loss": 0.052001953125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.9383201599121094, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8719867467880249, "num_tokens": 174787942.0, "step": 4601 }, { "epoch": 0.5854217020735275, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.457894802093506, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8589810729026794, "num_tokens": 174827483.0, "step": 4602 }, { "epoch": 0.585548912352118, "ewc_loss": 0.051513671875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 3.933906555175781e-05, "grad_norm": 2.8559494018554688, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8689942359924316, "num_tokens": 174862847.0, "step": 4603 }, { "epoch": 0.5856761226307086, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6903436183929443, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8476340770721436, "num_tokens": 174903327.0, "step": 4604 }, { "epoch": 0.5858033329092991, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.8380320072174072, "learning_rate": 1e-05, "loss": 0.5243, "mean_token_accuracy": 0.842540979385376, "num_tokens": 174939278.0, "step": 4605 }, { "epoch": 0.5859305431878896, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.024196147918701, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8556773066520691, "num_tokens": 174972671.0, "step": 4606 }, { "epoch": 0.5860577534664801, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.059293746948242, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8485985398292542, "num_tokens": 175009753.0, "step": 4607 }, { "epoch": 0.5861849637450706, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.220902442932129, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8726131319999695, "num_tokens": 175044271.0, "step": 4608 }, { "epoch": 0.5863121740236611, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.4679739475250244, "learning_rate": 1e-05, "loss": 0.5256, "mean_token_accuracy": 0.851294994354248, "num_tokens": 175086132.0, "step": 4609 }, { "epoch": 0.5864393843022516, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.080113887786865, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8703973293304443, "num_tokens": 175114902.0, "step": 4610 }, { "epoch": 0.5865665945808421, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.64215087890625, "learning_rate": 1e-05, "loss": 0.5092, "mean_token_accuracy": 0.8499757051467896, "num_tokens": 175152274.0, "step": 4611 }, { "epoch": 0.5866938048594327, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.805004358291626, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8620981574058533, "num_tokens": 175198449.0, "step": 4612 }, { "epoch": 0.5868210151380232, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5361826419830322, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8702657222747803, "num_tokens": 175236851.0, "step": 4613 }, { "epoch": 0.5869482254166136, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.642364978790283, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8586834669113159, "num_tokens": 175275694.0, "step": 4614 }, { "epoch": 0.5870754356952042, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.402884006500244, "learning_rate": 1e-05, "loss": 0.5404, "mean_token_accuracy": 0.8424849510192871, "num_tokens": 175321484.0, "step": 4615 }, { "epoch": 0.5872026459737947, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.529066562652588, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8603217005729675, "num_tokens": 175358387.0, "step": 4616 }, { "epoch": 0.5873298562523852, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.27026104927063, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8490695357322693, "num_tokens": 175402363.0, "step": 4617 }, { "epoch": 0.5874570665309757, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.486981153488159, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8539305925369263, "num_tokens": 175440655.0, "step": 4618 }, { "epoch": 0.5875842768095663, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1697864532470703, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8555482029914856, "num_tokens": 175476041.0, "step": 4619 }, { "epoch": 0.5877114870881567, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.505467176437378, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8761990070343018, "num_tokens": 175514673.0, "step": 4620 }, { "epoch": 0.5878386973667472, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0647165775299072, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.867748498916626, "num_tokens": 175555396.0, "step": 4621 }, { "epoch": 0.5879659076453377, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.6001925468444824, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8735108971595764, "num_tokens": 175597588.0, "step": 4622 }, { "epoch": 0.5880931179239283, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.02472186088562, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8641839623451233, "num_tokens": 175636911.0, "step": 4623 }, { "epoch": 0.5882203282025188, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7585017681121826, "learning_rate": 1e-05, "loss": 0.5391, "mean_token_accuracy": 0.8413812518119812, "num_tokens": 175671778.0, "step": 4624 }, { "epoch": 0.5883475384811093, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0747885704040527, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8755131363868713, "num_tokens": 175711218.0, "step": 4625 }, { "epoch": 0.5884747487596997, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8781001567840576, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.8585894107818604, "num_tokens": 175744630.0, "step": 4626 }, { "epoch": 0.5886019590382903, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2559702396392822, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8592442274093628, "num_tokens": 175783342.0, "step": 4627 }, { "epoch": 0.5887291693168808, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8126933574676514, "learning_rate": 1e-05, "loss": 0.5152, "mean_token_accuracy": 0.8467978835105896, "num_tokens": 175820361.0, "step": 4628 }, { "epoch": 0.5888563795954713, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3856329917907715, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8531191349029541, "num_tokens": 175860388.0, "step": 4629 }, { "epoch": 0.5889835898740619, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.61859393119812, "learning_rate": 1e-05, "loss": 0.5104, "mean_token_accuracy": 0.8564984798431396, "num_tokens": 175895725.0, "step": 4630 }, { "epoch": 0.5891108001526524, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.416684627532959, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.862790584564209, "num_tokens": 175933535.0, "step": 4631 }, { "epoch": 0.5892380104312428, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.395549774169922, "learning_rate": 1e-05, "loss": 0.5122, "mean_token_accuracy": 0.8491311073303223, "num_tokens": 175968681.0, "step": 4632 }, { "epoch": 0.5893652207098333, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3913679122924805, "learning_rate": 1e-05, "loss": 0.5006, "mean_token_accuracy": 0.8512395620346069, "num_tokens": 176006088.0, "step": 4633 }, { "epoch": 0.5894924309884239, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1980834007263184, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.872696042060852, "num_tokens": 176040855.0, "step": 4634 }, { "epoch": 0.5896196412670144, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3853721618652344, "learning_rate": 1e-05, "loss": 0.4751, "mean_token_accuracy": 0.8586163520812988, "num_tokens": 176080015.0, "step": 4635 }, { "epoch": 0.5897468515456049, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0757856369018555, "learning_rate": 1e-05, "loss": 0.5011, "mean_token_accuracy": 0.8471282124519348, "num_tokens": 176120266.0, "step": 4636 }, { "epoch": 0.5898740618241954, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4618265628814697, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8632338047027588, "num_tokens": 176150436.0, "step": 4637 }, { "epoch": 0.5900012721027859, "ewc_loss": 0.052001953125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.9860787391662598, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8479833006858826, "num_tokens": 176185340.0, "step": 4638 }, { "epoch": 0.5901284823813764, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.545452833175659, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8516464233398438, "num_tokens": 176225372.0, "step": 4639 }, { "epoch": 0.5902556926599669, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.932011127471924, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8578265905380249, "num_tokens": 176267114.0, "step": 4640 }, { "epoch": 0.5903829029385574, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7150204181671143, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8722929954528809, "num_tokens": 176302937.0, "step": 4641 }, { "epoch": 0.590510113217148, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0324599742889404, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8687982559204102, "num_tokens": 176340307.0, "step": 4642 }, { "epoch": 0.5906373234957385, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.806053638458252, "learning_rate": 1e-05, "loss": 0.5238, "mean_token_accuracy": 0.85001540184021, "num_tokens": 176385420.0, "step": 4643 }, { "epoch": 0.5907645337743289, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1165082454681396, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8541039824485779, "num_tokens": 176424975.0, "step": 4644 }, { "epoch": 0.5908917440529194, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.846381902694702, "learning_rate": 1e-05, "loss": 0.5484, "mean_token_accuracy": 0.8404857516288757, "num_tokens": 176465471.0, "step": 4645 }, { "epoch": 0.59101895433151, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.273420572280884, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8615627884864807, "num_tokens": 176493717.0, "step": 4646 }, { "epoch": 0.5911461646101005, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7420499324798584, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8560730814933777, "num_tokens": 176533636.0, "step": 4647 }, { "epoch": 0.591273374888691, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.32749605178833, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8556874990463257, "num_tokens": 176573248.0, "step": 4648 }, { "epoch": 0.5914005851672816, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.59321665763855, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8702612519264221, "num_tokens": 176614607.0, "step": 4649 }, { "epoch": 0.591527795445872, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.314201593399048, "learning_rate": 1e-05, "loss": 0.5033, "mean_token_accuracy": 0.8522995710372925, "num_tokens": 176657482.0, "step": 4650 }, { "epoch": 0.5916550057244625, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.46028208732605, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8674298524856567, "num_tokens": 176694448.0, "step": 4651 }, { "epoch": 0.591782216003053, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2829067707061768, "learning_rate": 1e-05, "loss": 0.5132, "mean_token_accuracy": 0.8534821271896362, "num_tokens": 176726899.0, "step": 4652 }, { "epoch": 0.5919094262816436, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.355935573577881, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8524817228317261, "num_tokens": 176767495.0, "step": 4653 }, { "epoch": 0.5920366365602341, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2203633785247803, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8544560670852661, "num_tokens": 176802804.0, "step": 4654 }, { "epoch": 0.5921638468388246, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2930562496185303, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8674578070640564, "num_tokens": 176847224.0, "step": 4655 }, { "epoch": 0.592291057117415, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1685166358947754, "learning_rate": 1e-05, "loss": 0.4949, "mean_token_accuracy": 0.8554703593254089, "num_tokens": 176884423.0, "step": 4656 }, { "epoch": 0.5924182673960056, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2942042350769043, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8668221831321716, "num_tokens": 176917776.0, "step": 4657 }, { "epoch": 0.5925454776745961, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1047847270965576, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.866473913192749, "num_tokens": 176951387.0, "step": 4658 }, { "epoch": 0.5926726879531866, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3240702152252197, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.85074782371521, "num_tokens": 176986533.0, "step": 4659 }, { "epoch": 0.5927998982317771, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0394020080566406, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.869225025177002, "num_tokens": 177021541.0, "step": 4660 }, { "epoch": 0.5929271085103677, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3844683170318604, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.860458493232727, "num_tokens": 177062933.0, "step": 4661 }, { "epoch": 0.5930543187889582, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.987497568130493, "learning_rate": 1e-05, "loss": 0.5313, "mean_token_accuracy": 0.8388587236404419, "num_tokens": 177102549.0, "step": 4662 }, { "epoch": 0.5931815290675486, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.490940570831299, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8569172620773315, "num_tokens": 177139618.0, "step": 4663 }, { "epoch": 0.5933087393461391, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.945383310317993, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8504902720451355, "num_tokens": 177171912.0, "step": 4664 }, { "epoch": 0.5934359496247297, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6441826820373535, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8584451675415039, "num_tokens": 177214279.0, "step": 4665 }, { "epoch": 0.5935631599033202, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.968047618865967, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.8548942804336548, "num_tokens": 177247277.0, "step": 4666 }, { "epoch": 0.5936903701819107, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7845308780670166, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8569101095199585, "num_tokens": 177287097.0, "step": 4667 }, { "epoch": 0.5938175804605013, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.0707874298095703, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8702277541160583, "num_tokens": 177324400.0, "step": 4668 }, { "epoch": 0.5939447907390917, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8481996059417725, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8652981519699097, "num_tokens": 177362962.0, "step": 4669 }, { "epoch": 0.5940720010176822, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2376480102539062, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8566583395004272, "num_tokens": 177402708.0, "step": 4670 }, { "epoch": 0.5941992112962727, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7572264671325684, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8678780794143677, "num_tokens": 177433957.0, "step": 4671 }, { "epoch": 0.5943264215748633, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001220703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3413822650909424, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8624306917190552, "num_tokens": 177465265.0, "step": 4672 }, { "epoch": 0.5944536318534538, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.583966016769409, "learning_rate": 1e-05, "loss": 0.5067, "mean_token_accuracy": 0.8498934507369995, "num_tokens": 177495448.0, "step": 4673 }, { "epoch": 0.5945808421320443, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3285484313964844, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8634134531021118, "num_tokens": 177531582.0, "step": 4674 }, { "epoch": 0.5947080524106347, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.408851146697998, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8564471006393433, "num_tokens": 177565039.0, "step": 4675 }, { "epoch": 0.5948352626892253, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3067798614501953, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8575516939163208, "num_tokens": 177601872.0, "step": 4676 }, { "epoch": 0.5949624729678158, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2620129585266113, "learning_rate": 1e-05, "loss": 0.5228, "mean_token_accuracy": 0.8448586463928223, "num_tokens": 177645693.0, "step": 4677 }, { "epoch": 0.5950896832464063, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2841663360595703, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8632858991622925, "num_tokens": 177685334.0, "step": 4678 }, { "epoch": 0.5952168935249968, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.177565336227417, "learning_rate": 1e-05, "loss": 0.6115, "mean_token_accuracy": 0.819193422794342, "num_tokens": 177727851.0, "step": 4679 }, { "epoch": 0.5953441038035874, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2716355323791504, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.854543149471283, "num_tokens": 177769115.0, "step": 4680 }, { "epoch": 0.5954713140821778, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1060664653778076, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8519400358200073, "num_tokens": 177809094.0, "step": 4681 }, { "epoch": 0.5955985243607683, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.302312135696411, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8588898777961731, "num_tokens": 177842409.0, "step": 4682 }, { "epoch": 0.5957257346393589, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.049186944961548, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8524239659309387, "num_tokens": 177882525.0, "step": 4683 }, { "epoch": 0.5958529449179494, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3781261444091797, "learning_rate": 1e-05, "loss": 0.5391, "mean_token_accuracy": 0.8422743082046509, "num_tokens": 177924293.0, "step": 4684 }, { "epoch": 0.5959801551965399, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.965585708618164, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8607465028762817, "num_tokens": 177964716.0, "step": 4685 }, { "epoch": 0.5961073654751304, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4656968116760254, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8780355453491211, "num_tokens": 177999029.0, "step": 4686 }, { "epoch": 0.5962345757537209, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.909987211227417, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8701331615447998, "num_tokens": 178033872.0, "step": 4687 }, { "epoch": 0.5963617860323114, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6371986865997314, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8622332215309143, "num_tokens": 178077835.0, "step": 4688 }, { "epoch": 0.5964889963109019, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.8890175819396973, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8657124042510986, "num_tokens": 178119205.0, "step": 4689 }, { "epoch": 0.5966162065894924, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8683788776397705, "learning_rate": 1e-05, "loss": 0.528, "mean_token_accuracy": 0.847609281539917, "num_tokens": 178157923.0, "step": 4690 }, { "epoch": 0.596743416868083, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.019913911819458, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.8734573125839233, "num_tokens": 178200465.0, "step": 4691 }, { "epoch": 0.5968706271466735, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.026332855224609, "learning_rate": 1e-05, "loss": 0.5801, "mean_token_accuracy": 0.8291240334510803, "num_tokens": 178234767.0, "step": 4692 }, { "epoch": 0.5969978374252639, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3133597373962402, "learning_rate": 1e-05, "loss": 0.5195, "mean_token_accuracy": 0.8487445116043091, "num_tokens": 178269569.0, "step": 4693 }, { "epoch": 0.5971250477038544, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.952104330062866, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.8575036525726318, "num_tokens": 178300789.0, "step": 4694 }, { "epoch": 0.597252257982445, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.456681728363037, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8668058514595032, "num_tokens": 178333341.0, "step": 4695 }, { "epoch": 0.5973794682610355, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7496261596679688, "learning_rate": 1e-05, "loss": 0.5361, "mean_token_accuracy": 0.8442909717559814, "num_tokens": 178368743.0, "step": 4696 }, { "epoch": 0.597506678539626, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4404397010803223, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.8783971071243286, "num_tokens": 178407432.0, "step": 4697 }, { "epoch": 0.5976338888182166, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5748791694641113, "learning_rate": 1e-05, "loss": 0.5045, "mean_token_accuracy": 0.8547377586364746, "num_tokens": 178440245.0, "step": 4698 }, { "epoch": 0.597761099096807, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.353407382965088, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8620108366012573, "num_tokens": 178484446.0, "step": 4699 }, { "epoch": 0.5978883093753975, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.432711601257324, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8600564002990723, "num_tokens": 178522390.0, "step": 4700 }, { "epoch": 0.598015519653988, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2751355171203613, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.864030122756958, "num_tokens": 178561671.0, "step": 4701 }, { "epoch": 0.5981427299325786, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.358388900756836, "learning_rate": 1e-05, "loss": 0.5452, "mean_token_accuracy": 0.8399390578269958, "num_tokens": 178605289.0, "step": 4702 }, { "epoch": 0.5982699402111691, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1940629482269287, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.866187572479248, "num_tokens": 178647947.0, "step": 4703 }, { "epoch": 0.5983971504897596, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3236119747161865, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8710402846336365, "num_tokens": 178686779.0, "step": 4704 }, { "epoch": 0.59852436076835, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1220703125, "learning_rate": 1e-05, "loss": 0.5116, "mean_token_accuracy": 0.8481390476226807, "num_tokens": 178730517.0, "step": 4705 }, { "epoch": 0.5986515710469406, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3684768676757812, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8570138216018677, "num_tokens": 178762521.0, "step": 4706 }, { "epoch": 0.5987787813255311, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0463814735412598, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8499257564544678, "num_tokens": 178804816.0, "step": 4707 }, { "epoch": 0.5989059916041216, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4314253330230713, "learning_rate": 1e-05, "loss": 0.541, "mean_token_accuracy": 0.8449593782424927, "num_tokens": 178845413.0, "step": 4708 }, { "epoch": 0.5990332018827121, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9660050868988037, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8566340208053589, "num_tokens": 178886887.0, "step": 4709 }, { "epoch": 0.5991604121613027, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5653936862945557, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8625096082687378, "num_tokens": 178922067.0, "step": 4710 }, { "epoch": 0.5992876224398932, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.964897394180298, "learning_rate": 1e-05, "loss": 0.5443, "mean_token_accuracy": 0.8369663953781128, "num_tokens": 178955645.0, "step": 4711 }, { "epoch": 0.5994148327184836, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7284622192382812, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8597598075866699, "num_tokens": 178992389.0, "step": 4712 }, { "epoch": 0.5995420429970741, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.004866361618042, "learning_rate": 1e-05, "loss": 0.5051, "mean_token_accuracy": 0.8470646142959595, "num_tokens": 179031140.0, "step": 4713 }, { "epoch": 0.5996692532756647, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8567118644714355, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8745622038841248, "num_tokens": 179065486.0, "step": 4714 }, { "epoch": 0.5997964635542552, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1768548488616943, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8550639152526855, "num_tokens": 179104957.0, "step": 4715 }, { "epoch": 0.5999236738328457, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.842303991317749, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.858372688293457, "num_tokens": 179146022.0, "step": 4716 }, { "epoch": 0.6000508841114363, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3246922492980957, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8664618134498596, "num_tokens": 179184691.0, "step": 4717 }, { "epoch": 0.6001780943900267, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6864986419677734, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8602504730224609, "num_tokens": 179223734.0, "step": 4718 }, { "epoch": 0.6003053046686172, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.364060401916504, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8656586408615112, "num_tokens": 179265997.0, "step": 4719 }, { "epoch": 0.6004325149472077, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.49784779548645, "learning_rate": 1e-05, "loss": 0.5002, "mean_token_accuracy": 0.8523419499397278, "num_tokens": 179302734.0, "step": 4720 }, { "epoch": 0.6005597252257983, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3572804927825928, "learning_rate": 1e-05, "loss": 0.5071, "mean_token_accuracy": 0.852393388748169, "num_tokens": 179343779.0, "step": 4721 }, { "epoch": 0.6006869355043888, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.309113025665283, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8656351566314697, "num_tokens": 179384713.0, "step": 4722 }, { "epoch": 0.6008141457829793, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3419454097747803, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8645843267440796, "num_tokens": 179418005.0, "step": 4723 }, { "epoch": 0.6009413560615697, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1833808422088623, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8511403203010559, "num_tokens": 179459437.0, "step": 4724 }, { "epoch": 0.6010685663401603, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3399481773376465, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.861331045627594, "num_tokens": 179490534.0, "step": 4725 }, { "epoch": 0.6011957766187508, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.085847854614258, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8569295406341553, "num_tokens": 179533141.0, "step": 4726 }, { "epoch": 0.6013229868973413, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3760194778442383, "learning_rate": 1e-05, "loss": 0.5127, "mean_token_accuracy": 0.84857577085495, "num_tokens": 179565438.0, "step": 4727 }, { "epoch": 0.6014501971759318, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.006962537765503, "learning_rate": 1e-05, "loss": 0.5322, "mean_token_accuracy": 0.8405855894088745, "num_tokens": 179605184.0, "step": 4728 }, { "epoch": 0.6015774074545224, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.449632167816162, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8584622144699097, "num_tokens": 179643355.0, "step": 4729 }, { "epoch": 0.6017046177331128, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.945075035095215, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8634297847747803, "num_tokens": 179682471.0, "step": 4730 }, { "epoch": 0.6018318280117033, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5901737213134766, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8486725091934204, "num_tokens": 179722290.0, "step": 4731 }, { "epoch": 0.6019590382902938, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.9279701709747314, "learning_rate": 1e-05, "loss": 0.5182, "mean_token_accuracy": 0.8441118001937866, "num_tokens": 179762409.0, "step": 4732 }, { "epoch": 0.6020862485688844, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.77327561378479, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.8467099666595459, "num_tokens": 179802561.0, "step": 4733 }, { "epoch": 0.6022134588474749, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 2.9987032413482666, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8600789308547974, "num_tokens": 179842923.0, "step": 4734 }, { "epoch": 0.6023406691260654, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9174153804779053, "learning_rate": 1e-05, "loss": 0.5082, "mean_token_accuracy": 0.8548569679260254, "num_tokens": 179882430.0, "step": 4735 }, { "epoch": 0.6024678794046558, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.1905288696289062, "learning_rate": 1e-05, "loss": 0.5123, "mean_token_accuracy": 0.8504797220230103, "num_tokens": 179923974.0, "step": 4736 }, { "epoch": 0.6025950896832464, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9080922603607178, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8667943477630615, "num_tokens": 179966468.0, "step": 4737 }, { "epoch": 0.6027222999618369, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.354365587234497, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8718245029449463, "num_tokens": 180005356.0, "step": 4738 }, { "epoch": 0.6028495102404274, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7662458419799805, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8562272787094116, "num_tokens": 180044765.0, "step": 4739 }, { "epoch": 0.602976720519018, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.409355401992798, "learning_rate": 1e-05, "loss": 0.5187, "mean_token_accuracy": 0.8497012853622437, "num_tokens": 180082311.0, "step": 4740 }, { "epoch": 0.6031039307976085, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5683865547180176, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8625013828277588, "num_tokens": 180122889.0, "step": 4741 }, { "epoch": 0.6032311410761989, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3659629821777344, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8636439442634583, "num_tokens": 180157425.0, "step": 4742 }, { "epoch": 0.6033583513547894, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4096240997314453, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8645104169845581, "num_tokens": 180196099.0, "step": 4743 }, { "epoch": 0.60348556163338, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3026812076568604, "learning_rate": 1e-05, "loss": 0.5382, "mean_token_accuracy": 0.8404898643493652, "num_tokens": 180238036.0, "step": 4744 }, { "epoch": 0.6036127719119705, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3034346103668213, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8599687218666077, "num_tokens": 180276193.0, "step": 4745 }, { "epoch": 0.603739982190561, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2450170516967773, "learning_rate": 1e-05, "loss": 0.4941, "mean_token_accuracy": 0.8552765846252441, "num_tokens": 180313470.0, "step": 4746 }, { "epoch": 0.6038671924691515, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2450804710388184, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8614314794540405, "num_tokens": 180354588.0, "step": 4747 }, { "epoch": 0.603994402747742, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1933655738830566, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8578898310661316, "num_tokens": 180395792.0, "step": 4748 }, { "epoch": 0.6041216130263325, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.231199264526367, "learning_rate": 1e-05, "loss": 0.5056, "mean_token_accuracy": 0.8494173288345337, "num_tokens": 180436823.0, "step": 4749 }, { "epoch": 0.604248823304923, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1423795223236084, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8671467304229736, "num_tokens": 180472643.0, "step": 4750 }, { "epoch": 0.6043760335835135, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2288029193878174, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8491898775100708, "num_tokens": 180511788.0, "step": 4751 }, { "epoch": 0.6045032438621041, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00012302398681640625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.097916841506958, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8712782859802246, "num_tokens": 180547871.0, "step": 4752 }, { "epoch": 0.6046304541406946, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2784719467163086, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8469357490539551, "num_tokens": 180580521.0, "step": 4753 }, { "epoch": 0.604757664419285, "ewc_loss": 0.052490234375, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.034625768661499, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8524901866912842, "num_tokens": 180617726.0, "step": 4754 }, { "epoch": 0.6048848746978756, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3395979404449463, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8567718863487244, "num_tokens": 180657705.0, "step": 4755 }, { "epoch": 0.6050120849764661, "ewc_loss": 0.051513671875, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 3.910064697265625e-05, "grad_norm": 2.985555410385132, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.866526186466217, "num_tokens": 180693582.0, "step": 4756 }, { "epoch": 0.6051392952550566, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001239776611328125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.445894956588745, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8511135578155518, "num_tokens": 180731159.0, "step": 4757 }, { "epoch": 0.6052665055336471, "ewc_loss": 0.05126953125, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 3.886222839355469e-05, "grad_norm": 2.9127163887023926, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8482012152671814, "num_tokens": 180769901.0, "step": 4758 }, { "epoch": 0.6053937158122377, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.616708755493164, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8613663911819458, "num_tokens": 180811763.0, "step": 4759 }, { "epoch": 0.6055209260908282, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.879983425140381, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8610767126083374, "num_tokens": 180846073.0, "step": 4760 }, { "epoch": 0.6056481363694186, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8666114807128906, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8661952614784241, "num_tokens": 180883609.0, "step": 4761 }, { "epoch": 0.6057753466480091, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.0018277168273926, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8558968305587769, "num_tokens": 180921492.0, "step": 4762 }, { "epoch": 0.6059025569265997, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.062028408050537, "learning_rate": 1e-05, "loss": 0.5396, "mean_token_accuracy": 0.8446863889694214, "num_tokens": 180960298.0, "step": 4763 }, { "epoch": 0.6060297672051902, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.294736385345459, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8510808348655701, "num_tokens": 180997002.0, "step": 4764 }, { "epoch": 0.6061569774837807, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.000810623168945, "learning_rate": 1e-05, "loss": 0.5571, "mean_token_accuracy": 0.8393781185150146, "num_tokens": 181033575.0, "step": 4765 }, { "epoch": 0.6062841877623713, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.484173059463501, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.857900083065033, "num_tokens": 181068584.0, "step": 4766 }, { "epoch": 0.6064113980409617, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.80155086517334, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8628978729248047, "num_tokens": 181103017.0, "step": 4767 }, { "epoch": 0.6065386083195522, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.482609272003174, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8584354519844055, "num_tokens": 181137590.0, "step": 4768 }, { "epoch": 0.6066658185981427, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.59438157081604, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8577583432197571, "num_tokens": 181177581.0, "step": 4769 }, { "epoch": 0.6067930288767333, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3938775062561035, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8676625490188599, "num_tokens": 181212582.0, "step": 4770 }, { "epoch": 0.6069202391553238, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.453524112701416, "learning_rate": 1e-05, "loss": 0.5119, "mean_token_accuracy": 0.8486615419387817, "num_tokens": 181246373.0, "step": 4771 }, { "epoch": 0.6070474494339143, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3058621883392334, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8721433281898499, "num_tokens": 181280657.0, "step": 4772 }, { "epoch": 0.6071746597125047, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00012493133544921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3448116779327393, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8672230243682861, "num_tokens": 181322165.0, "step": 4773 }, { "epoch": 0.6073018699910953, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.223500967025757, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8560408353805542, "num_tokens": 181358612.0, "step": 4774 }, { "epoch": 0.6074290802696858, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3149566650390625, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.871332049369812, "num_tokens": 181392928.0, "step": 4775 }, { "epoch": 0.6075562905482763, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.153578996658325, "learning_rate": 1e-05, "loss": 0.5427, "mean_token_accuracy": 0.8410829305648804, "num_tokens": 181431127.0, "step": 4776 }, { "epoch": 0.6076835008268668, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3146462440490723, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8591828346252441, "num_tokens": 181469830.0, "step": 4777 }, { "epoch": 0.6078107111054574, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.082111120223999, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8574185371398926, "num_tokens": 181505590.0, "step": 4778 }, { "epoch": 0.6079379213840478, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3611772060394287, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8728610873222351, "num_tokens": 181541675.0, "step": 4779 }, { "epoch": 0.6080651316626383, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0074541568756104, "learning_rate": 1e-05, "loss": 0.5126, "mean_token_accuracy": 0.844991147518158, "num_tokens": 181579890.0, "step": 4780 }, { "epoch": 0.6081923419412288, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4713146686553955, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8563317060470581, "num_tokens": 181613821.0, "step": 4781 }, { "epoch": 0.6083195522198194, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9323439598083496, "learning_rate": 1e-05, "loss": 0.5179, "mean_token_accuracy": 0.8447796106338501, "num_tokens": 181656899.0, "step": 4782 }, { "epoch": 0.6084467624984099, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.659215211868286, "learning_rate": 1e-05, "loss": 0.528, "mean_token_accuracy": 0.8418419361114502, "num_tokens": 181690044.0, "step": 4783 }, { "epoch": 0.6085739727770004, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.929985761642456, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8627904653549194, "num_tokens": 181725038.0, "step": 4784 }, { "epoch": 0.6087011830555908, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8622472286224365, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8596013784408569, "num_tokens": 181767914.0, "step": 4785 }, { "epoch": 0.6088283933341814, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.0723204612731934, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8688095808029175, "num_tokens": 181798955.0, "step": 4786 }, { "epoch": 0.6089556036127719, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.981910467147827, "learning_rate": 1e-05, "loss": 0.5399, "mean_token_accuracy": 0.8449094295501709, "num_tokens": 181831967.0, "step": 4787 }, { "epoch": 0.6090828138913624, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3211607933044434, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8577017188072205, "num_tokens": 181865717.0, "step": 4788 }, { "epoch": 0.609210024169953, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.869774103164673, "learning_rate": 1e-05, "loss": 0.5495, "mean_token_accuracy": 0.8414541482925415, "num_tokens": 181902847.0, "step": 4789 }, { "epoch": 0.6093372344485435, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4364538192749023, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8701672554016113, "num_tokens": 181937129.0, "step": 4790 }, { "epoch": 0.6094644447271339, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.660810708999634, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8575705289840698, "num_tokens": 181979576.0, "step": 4791 }, { "epoch": 0.6095916550057244, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4046874046325684, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8569720983505249, "num_tokens": 182021510.0, "step": 4792 }, { "epoch": 0.609718865284315, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.477574586868286, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8584810495376587, "num_tokens": 182061178.0, "step": 4793 }, { "epoch": 0.6098460755629055, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.351205348968506, "learning_rate": 1e-05, "loss": 0.5388, "mean_token_accuracy": 0.8441975116729736, "num_tokens": 182100575.0, "step": 4794 }, { "epoch": 0.609973285841496, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.339362859725952, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8561339378356934, "num_tokens": 182136142.0, "step": 4795 }, { "epoch": 0.6101004961200865, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2960801124572754, "learning_rate": 1e-05, "loss": 0.5167, "mean_token_accuracy": 0.8488908410072327, "num_tokens": 182173108.0, "step": 4796 }, { "epoch": 0.610227706398677, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.268066167831421, "learning_rate": 1e-05, "loss": 0.5211, "mean_token_accuracy": 0.8461923599243164, "num_tokens": 182209931.0, "step": 4797 }, { "epoch": 0.6103549166772675, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2480368614196777, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8615034818649292, "num_tokens": 182244128.0, "step": 4798 }, { "epoch": 0.610482126955858, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.20906662940979, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8568300008773804, "num_tokens": 182278112.0, "step": 4799 }, { "epoch": 0.6106093372344485, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00012683868408203125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.203509569168091, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8617745637893677, "num_tokens": 182313753.0, "step": 4800 }, { "epoch": 0.6107365475130391, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1720738410949707, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8782016038894653, "num_tokens": 182348758.0, "step": 4801 }, { "epoch": 0.6108637577916296, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2006514072418213, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8575884103775024, "num_tokens": 182381084.0, "step": 4802 }, { "epoch": 0.61099096807022, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.142072916030884, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8595406413078308, "num_tokens": 182420530.0, "step": 4803 }, { "epoch": 0.6111181783488105, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.193807363510132, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8570365905761719, "num_tokens": 182462286.0, "step": 4804 }, { "epoch": 0.6112453886274011, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.109954357147217, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8725911974906921, "num_tokens": 182497209.0, "step": 4805 }, { "epoch": 0.6113725989059916, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.205875873565674, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8699625134468079, "num_tokens": 182530609.0, "step": 4806 }, { "epoch": 0.6114998091845821, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0895705223083496, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8653241395950317, "num_tokens": 182566698.0, "step": 4807 }, { "epoch": 0.6116270194631727, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2274158000946045, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8642118573188782, "num_tokens": 182609524.0, "step": 4808 }, { "epoch": 0.6117542297417632, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.03788685798645, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8612500429153442, "num_tokens": 182649339.0, "step": 4809 }, { "epoch": 0.6118814400203536, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2996442317962646, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8594000935554504, "num_tokens": 182684142.0, "step": 4810 }, { "epoch": 0.6120086502989441, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9726359844207764, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8556195497512817, "num_tokens": 182722857.0, "step": 4811 }, { "epoch": 0.6121358605775347, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4179747104644775, "learning_rate": 1e-05, "loss": 0.4941, "mean_token_accuracy": 0.8518495559692383, "num_tokens": 182760196.0, "step": 4812 }, { "epoch": 0.6122630708561252, "ewc_loss": 0.05224609375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.8860299587249756, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8540483713150024, "num_tokens": 182795807.0, "step": 4813 }, { "epoch": 0.6123902811347157, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6184918880462646, "learning_rate": 1e-05, "loss": 0.5143, "mean_token_accuracy": 0.850348949432373, "num_tokens": 182837956.0, "step": 4814 }, { "epoch": 0.6125174914133062, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.8332600593566895, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8638500571250916, "num_tokens": 182873374.0, "step": 4815 }, { "epoch": 0.6126447016918967, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.925858974456787, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8712705373764038, "num_tokens": 182908018.0, "step": 4816 }, { "epoch": 0.6127719119704872, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 2.951481819152832, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.856475830078125, "num_tokens": 182950314.0, "step": 4817 }, { "epoch": 0.6128991222490777, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.21622896194458, "learning_rate": 1e-05, "loss": 0.5116, "mean_token_accuracy": 0.8571281433105469, "num_tokens": 182988434.0, "step": 4818 }, { "epoch": 0.6130263325276682, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.3624351024627686, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8650034666061401, "num_tokens": 183026355.0, "step": 4819 }, { "epoch": 0.6131535428062588, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.215906143188477, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8568823337554932, "num_tokens": 183063031.0, "step": 4820 }, { "epoch": 0.6132807530848493, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.679131269454956, "learning_rate": 1e-05, "loss": 0.5494, "mean_token_accuracy": 0.8463265299797058, "num_tokens": 183102082.0, "step": 4821 }, { "epoch": 0.6134079633634397, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.890901565551758, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8572156429290771, "num_tokens": 183137961.0, "step": 4822 }, { "epoch": 0.6135351736420303, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6078274250030518, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8709335327148438, "num_tokens": 183173779.0, "step": 4823 }, { "epoch": 0.6136623839206208, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6947338581085205, "learning_rate": 1e-05, "loss": 0.5132, "mean_token_accuracy": 0.8500105738639832, "num_tokens": 183215044.0, "step": 4824 }, { "epoch": 0.6137895941992113, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.467498540878296, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8651280403137207, "num_tokens": 183248793.0, "step": 4825 }, { "epoch": 0.6139168044778018, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5471768379211426, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8566774129867554, "num_tokens": 183291180.0, "step": 4826 }, { "epoch": 0.6140440147563924, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.332308292388916, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8580970764160156, "num_tokens": 183326214.0, "step": 4827 }, { "epoch": 0.6141712250349828, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001277923583984375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4793567657470703, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8623119592666626, "num_tokens": 183362403.0, "step": 4828 }, { "epoch": 0.6142984353135733, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2026383876800537, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8645790219306946, "num_tokens": 183407419.0, "step": 4829 }, { "epoch": 0.6144256455921638, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.484591245651245, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8581499457359314, "num_tokens": 183443140.0, "step": 4830 }, { "epoch": 0.6145528558707544, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.108555555343628, "learning_rate": 1e-05, "loss": 0.5371, "mean_token_accuracy": 0.8426839113235474, "num_tokens": 183485986.0, "step": 4831 }, { "epoch": 0.6146800661493449, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.567962408065796, "learning_rate": 1e-05, "loss": 0.547, "mean_token_accuracy": 0.8426022529602051, "num_tokens": 183525302.0, "step": 4832 }, { "epoch": 0.6148072764279354, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.030024290084839, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8518592715263367, "num_tokens": 183561575.0, "step": 4833 }, { "epoch": 0.6149344867065258, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.71752667427063, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8650511503219604, "num_tokens": 183596465.0, "step": 4834 }, { "epoch": 0.6150616969851164, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.033155679702759, "learning_rate": 1e-05, "loss": 0.5323, "mean_token_accuracy": 0.8359948396682739, "num_tokens": 183631733.0, "step": 4835 }, { "epoch": 0.6151889072637069, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8739612102508545, "learning_rate": 1e-05, "loss": 0.5108, "mean_token_accuracy": 0.8548550605773926, "num_tokens": 183676138.0, "step": 4836 }, { "epoch": 0.6153161175422974, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1731650829315186, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8578791618347168, "num_tokens": 183711004.0, "step": 4837 }, { "epoch": 0.615443327820888, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.918531656265259, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8796284794807434, "num_tokens": 183744945.0, "step": 4838 }, { "epoch": 0.6155705380994785, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3793184757232666, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.868000864982605, "num_tokens": 183779119.0, "step": 4839 }, { "epoch": 0.6156977483780689, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7462029457092285, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8661787509918213, "num_tokens": 183814136.0, "step": 4840 }, { "epoch": 0.6158249586566594, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.441364288330078, "learning_rate": 1e-05, "loss": 0.4981, "mean_token_accuracy": 0.8528268337249756, "num_tokens": 183850957.0, "step": 4841 }, { "epoch": 0.61595216893525, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.512382745742798, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8691058158874512, "num_tokens": 183890049.0, "step": 4842 }, { "epoch": 0.6160793792138405, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.411555290222168, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8585944175720215, "num_tokens": 183933055.0, "step": 4843 }, { "epoch": 0.616206589492431, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.31413197517395, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8579128980636597, "num_tokens": 183975059.0, "step": 4844 }, { "epoch": 0.6163337997710215, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3984413146972656, "learning_rate": 1e-05, "loss": 0.5085, "mean_token_accuracy": 0.8514879941940308, "num_tokens": 184015130.0, "step": 4845 }, { "epoch": 0.616461010049612, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1823642253875732, "learning_rate": 1e-05, "loss": 0.5214, "mean_token_accuracy": 0.8469003438949585, "num_tokens": 184048416.0, "step": 4846 }, { "epoch": 0.6165882203282025, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4016644954681396, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8486766815185547, "num_tokens": 184081591.0, "step": 4847 }, { "epoch": 0.616715430606793, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.058002471923828, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8785459399223328, "num_tokens": 184118078.0, "step": 4848 }, { "epoch": 0.6168426408853835, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4505324363708496, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8628436326980591, "num_tokens": 184154860.0, "step": 4849 }, { "epoch": 0.6169698511639741, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9871346950531006, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.869970440864563, "num_tokens": 184186425.0, "step": 4850 }, { "epoch": 0.6170970614425646, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5669589042663574, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.85190349817276, "num_tokens": 184224891.0, "step": 4851 }, { "epoch": 0.617224271721155, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.9348814487457275, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8705410957336426, "num_tokens": 184266648.0, "step": 4852 }, { "epoch": 0.6173514819997455, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7604241371154785, "learning_rate": 1e-05, "loss": 0.5496, "mean_token_accuracy": 0.8469445109367371, "num_tokens": 184301453.0, "step": 4853 }, { "epoch": 0.6174786922783361, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.002545118331909, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8568918704986572, "num_tokens": 184344271.0, "step": 4854 }, { "epoch": 0.6176059025569266, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.909869432449341, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8643596768379211, "num_tokens": 184378212.0, "step": 4855 }, { "epoch": 0.6177331128355171, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.184523820877075, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8647419214248657, "num_tokens": 184414177.0, "step": 4856 }, { "epoch": 0.6178603231141077, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9168291091918945, "learning_rate": 1e-05, "loss": 0.5314, "mean_token_accuracy": 0.8478556871414185, "num_tokens": 184455088.0, "step": 4857 }, { "epoch": 0.6179875333926982, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.365056037902832, "learning_rate": 1e-05, "loss": 0.5324, "mean_token_accuracy": 0.8480042219161987, "num_tokens": 184496574.0, "step": 4858 }, { "epoch": 0.6181147436712886, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7732748985290527, "learning_rate": 1e-05, "loss": 0.519, "mean_token_accuracy": 0.8525278568267822, "num_tokens": 184542303.0, "step": 4859 }, { "epoch": 0.6182419539498791, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3990046977996826, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.8579521179199219, "num_tokens": 184583337.0, "step": 4860 }, { "epoch": 0.6183691642284697, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6044933795928955, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.8539687395095825, "num_tokens": 184622424.0, "step": 4861 }, { "epoch": 0.6184963745070602, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.376121759414673, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8435680270195007, "num_tokens": 184660137.0, "step": 4862 }, { "epoch": 0.6186235847856507, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4428889751434326, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8558450937271118, "num_tokens": 184698353.0, "step": 4863 }, { "epoch": 0.6187507950642412, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3073551654815674, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8504303097724915, "num_tokens": 184738734.0, "step": 4864 }, { "epoch": 0.6188780053428317, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.351726770401001, "learning_rate": 1e-05, "loss": 0.5025, "mean_token_accuracy": 0.8492395877838135, "num_tokens": 184774525.0, "step": 4865 }, { "epoch": 0.6190052156214222, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2343876361846924, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8628453612327576, "num_tokens": 184809028.0, "step": 4866 }, { "epoch": 0.6191324259000127, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2912158966064453, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8771315217018127, "num_tokens": 184845479.0, "step": 4867 }, { "epoch": 0.6192596361786032, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1785919666290283, "learning_rate": 1e-05, "loss": 0.5315, "mean_token_accuracy": 0.8425933122634888, "num_tokens": 184886596.0, "step": 4868 }, { "epoch": 0.6193868464571938, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.28997802734375, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8545635342597961, "num_tokens": 184923763.0, "step": 4869 }, { "epoch": 0.6195140567357843, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.115788221359253, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8642619848251343, "num_tokens": 184959882.0, "step": 4870 }, { "epoch": 0.6196412670143747, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.304651975631714, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8736395239830017, "num_tokens": 184999063.0, "step": 4871 }, { "epoch": 0.6197684772929652, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0490949153900146, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.873760461807251, "num_tokens": 185040016.0, "step": 4872 }, { "epoch": 0.6198956875715558, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3728744983673096, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8561892509460449, "num_tokens": 185075085.0, "step": 4873 }, { "epoch": 0.6200228978501463, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9996235370635986, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8575303554534912, "num_tokens": 185111621.0, "step": 4874 }, { "epoch": 0.6201501081287368, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.486813545227051, "learning_rate": 1e-05, "loss": 0.5078, "mean_token_accuracy": 0.8519454002380371, "num_tokens": 185149681.0, "step": 4875 }, { "epoch": 0.6202773184073274, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.93684458732605, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8654483556747437, "num_tokens": 185186449.0, "step": 4876 }, { "epoch": 0.6204045286859178, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6459009647369385, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8642463684082031, "num_tokens": 185219324.0, "step": 4877 }, { "epoch": 0.6205317389645083, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.946068048477173, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8558872938156128, "num_tokens": 185256049.0, "step": 4878 }, { "epoch": 0.6206589492430988, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8305907249450684, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8629989624023438, "num_tokens": 185293304.0, "step": 4879 }, { "epoch": 0.6207861595216894, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.0560996532440186, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8618495464324951, "num_tokens": 185334007.0, "step": 4880 }, { "epoch": 0.6209133698002799, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9491679668426514, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8513561487197876, "num_tokens": 185372932.0, "step": 4881 }, { "epoch": 0.6210405800788704, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.273019790649414, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8655941486358643, "num_tokens": 185410267.0, "step": 4882 }, { "epoch": 0.6211677903574608, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00012874603271484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8750810623168945, "learning_rate": 1e-05, "loss": 0.5179, "mean_token_accuracy": 0.8519942760467529, "num_tokens": 185445838.0, "step": 4883 }, { "epoch": 0.6212950006360514, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.4139063358306885, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8621349334716797, "num_tokens": 185486002.0, "step": 4884 }, { "epoch": 0.6214222109146419, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6672632694244385, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.853722870349884, "num_tokens": 185524501.0, "step": 4885 }, { "epoch": 0.6215494211932324, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4130806922912598, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.855556309223175, "num_tokens": 185562105.0, "step": 4886 }, { "epoch": 0.621676631471823, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.464343309402466, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8673655986785889, "num_tokens": 185602821.0, "step": 4887 }, { "epoch": 0.6218038417504135, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.366060972213745, "learning_rate": 1e-05, "loss": 0.5128, "mean_token_accuracy": 0.849744439125061, "num_tokens": 185643188.0, "step": 4888 }, { "epoch": 0.6219310520290039, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.314451217651367, "learning_rate": 1e-05, "loss": 0.5324, "mean_token_accuracy": 0.8434380888938904, "num_tokens": 185685770.0, "step": 4889 }, { "epoch": 0.6220582623075944, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3341658115386963, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8614060878753662, "num_tokens": 185718981.0, "step": 4890 }, { "epoch": 0.622185472586185, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2092103958129883, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.862080991268158, "num_tokens": 185752184.0, "step": 4891 }, { "epoch": 0.6223126828647755, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3079640865325928, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8535240888595581, "num_tokens": 185791405.0, "step": 4892 }, { "epoch": 0.622439893143366, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.13539457321167, "learning_rate": 1e-05, "loss": 0.5414, "mean_token_accuracy": 0.8379454016685486, "num_tokens": 185832227.0, "step": 4893 }, { "epoch": 0.6225671034219565, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.3079702854156494, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8623173832893372, "num_tokens": 185867175.0, "step": 4894 }, { "epoch": 0.622694313700547, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0832395553588867, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8552287817001343, "num_tokens": 185907582.0, "step": 4895 }, { "epoch": 0.6228215239791375, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3350062370300293, "learning_rate": 1e-05, "loss": 0.5072, "mean_token_accuracy": 0.8465043306350708, "num_tokens": 185944625.0, "step": 4896 }, { "epoch": 0.622948734257728, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.027726888656616, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8707543611526489, "num_tokens": 185975579.0, "step": 4897 }, { "epoch": 0.6230759445363185, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3879597187042236, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8606537580490112, "num_tokens": 186016948.0, "step": 4898 }, { "epoch": 0.6232031548149091, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.958556652069092, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8511339426040649, "num_tokens": 186060700.0, "step": 4899 }, { "epoch": 0.6233303650934996, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.513105630874634, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8723611831665039, "num_tokens": 186099624.0, "step": 4900 }, { "epoch": 0.62345757537209, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9014861583709717, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8570034503936768, "num_tokens": 186142360.0, "step": 4901 }, { "epoch": 0.6235847856506805, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.712620496749878, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8690764307975769, "num_tokens": 186182642.0, "step": 4902 }, { "epoch": 0.6237119959292711, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.932534694671631, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8548461198806763, "num_tokens": 186218867.0, "step": 4903 }, { "epoch": 0.6238392062078616, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9258670806884766, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8706064224243164, "num_tokens": 186257424.0, "step": 4904 }, { "epoch": 0.6239664164864521, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.0926997661590576, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.859214186668396, "num_tokens": 186298588.0, "step": 4905 }, { "epoch": 0.6240936267650427, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.040280818939209, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8665489554405212, "num_tokens": 186334638.0, "step": 4906 }, { "epoch": 0.6242208370436332, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3788530826568604, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.851752758026123, "num_tokens": 186368303.0, "step": 4907 }, { "epoch": 0.6243480473222236, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9096803665161133, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8723989129066467, "num_tokens": 186406133.0, "step": 4908 }, { "epoch": 0.6244752576008141, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.470289945602417, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.8766874074935913, "num_tokens": 186440763.0, "step": 4909 }, { "epoch": 0.6246024678794047, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7183024883270264, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8578566908836365, "num_tokens": 186479662.0, "step": 4910 }, { "epoch": 0.6247296781579952, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.429581642150879, "learning_rate": 1e-05, "loss": 0.5301, "mean_token_accuracy": 0.847497820854187, "num_tokens": 186518944.0, "step": 4911 }, { "epoch": 0.6248568884365857, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5374183654785156, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8686017990112305, "num_tokens": 186555902.0, "step": 4912 }, { "epoch": 0.6249840987151762, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.346188545227051, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8758653998374939, "num_tokens": 186591962.0, "step": 4913 }, { "epoch": 0.6251113089937667, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.421187162399292, "learning_rate": 1e-05, "loss": 0.5119, "mean_token_accuracy": 0.851649284362793, "num_tokens": 186631841.0, "step": 4914 }, { "epoch": 0.6252385192723572, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2516770362854004, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8685073852539062, "num_tokens": 186677963.0, "step": 4915 }, { "epoch": 0.6253657295509477, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3707141876220703, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8574607968330383, "num_tokens": 186715658.0, "step": 4916 }, { "epoch": 0.6254929398295382, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1909725666046143, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8661278486251831, "num_tokens": 186749405.0, "step": 4917 }, { "epoch": 0.6256201501081288, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00012969970703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3545403480529785, "learning_rate": 1e-05, "loss": 0.5068, "mean_token_accuracy": 0.8495609760284424, "num_tokens": 186785667.0, "step": 4918 }, { "epoch": 0.6257473603867193, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1212644577026367, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8719955682754517, "num_tokens": 186817425.0, "step": 4919 }, { "epoch": 0.6258745706653097, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3729870319366455, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8677819967269897, "num_tokens": 186859693.0, "step": 4920 }, { "epoch": 0.6260017809439002, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.052912473678589, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8664886951446533, "num_tokens": 186897774.0, "step": 4921 }, { "epoch": 0.6261289912224908, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.449747323989868, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8600653409957886, "num_tokens": 186935530.0, "step": 4922 }, { "epoch": 0.6262562015010813, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9989495277404785, "learning_rate": 1e-05, "loss": 0.5111, "mean_token_accuracy": 0.8465307354927063, "num_tokens": 186970778.0, "step": 4923 }, { "epoch": 0.6263834117796718, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.575190544128418, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8587853908538818, "num_tokens": 187007338.0, "step": 4924 }, { "epoch": 0.6265106220582624, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.9608280658721924, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8593600988388062, "num_tokens": 187042614.0, "step": 4925 }, { "epoch": 0.6266378323368528, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.740507125854492, "learning_rate": 1e-05, "loss": 0.5066, "mean_token_accuracy": 0.853705883026123, "num_tokens": 187081753.0, "step": 4926 }, { "epoch": 0.6267650426154433, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.023658275604248, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8516852259635925, "num_tokens": 187116758.0, "step": 4927 }, { "epoch": 0.6268922528940338, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8674685955047607, "learning_rate": 1e-05, "loss": 0.5208, "mean_token_accuracy": 0.8502010107040405, "num_tokens": 187152296.0, "step": 4928 }, { "epoch": 0.6270194631726244, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.186946153640747, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8518620133399963, "num_tokens": 187187619.0, "step": 4929 }, { "epoch": 0.6271466734512149, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8408172130584717, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8585668802261353, "num_tokens": 187227872.0, "step": 4930 }, { "epoch": 0.6272738837298054, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3242974281311035, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8579977750778198, "num_tokens": 187265720.0, "step": 4931 }, { "epoch": 0.6274010940083958, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.698090076446533, "learning_rate": 1e-05, "loss": 0.575, "mean_token_accuracy": 0.8348921537399292, "num_tokens": 187307119.0, "step": 4932 }, { "epoch": 0.6275283042869864, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.374370574951172, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8619773387908936, "num_tokens": 187342804.0, "step": 4933 }, { "epoch": 0.6276555145655769, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.49015212059021, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8589158058166504, "num_tokens": 187381540.0, "step": 4934 }, { "epoch": 0.6277827248441674, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3564162254333496, "learning_rate": 1e-05, "loss": 0.5525, "mean_token_accuracy": 0.8419790863990784, "num_tokens": 187424952.0, "step": 4935 }, { "epoch": 0.627909935122758, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3229615688323975, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8488849401473999, "num_tokens": 187463668.0, "step": 4936 }, { "epoch": 0.6280371454013485, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3340320587158203, "learning_rate": 1e-05, "loss": 0.5188, "mean_token_accuracy": 0.8550866842269897, "num_tokens": 187505962.0, "step": 4937 }, { "epoch": 0.6281643556799389, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1976637840270996, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8524146676063538, "num_tokens": 187546342.0, "step": 4938 }, { "epoch": 0.6282915659585294, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3238027095794678, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8572772741317749, "num_tokens": 187587378.0, "step": 4939 }, { "epoch": 0.62841877623712, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1181604862213135, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8607345819473267, "num_tokens": 187621972.0, "step": 4940 }, { "epoch": 0.6285459865157105, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.375204563140869, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.8471950888633728, "num_tokens": 187657205.0, "step": 4941 }, { "epoch": 0.628673196794301, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0294861793518066, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8512316942214966, "num_tokens": 187698201.0, "step": 4942 }, { "epoch": 0.6288004070728915, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.428972005844116, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8686903715133667, "num_tokens": 187734883.0, "step": 4943 }, { "epoch": 0.628927617351482, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9583899974823, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8724319934844971, "num_tokens": 187777459.0, "step": 4944 }, { "epoch": 0.6290548276300725, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5589568614959717, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.852026104927063, "num_tokens": 187819661.0, "step": 4945 }, { "epoch": 0.629182037908663, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.926516532897949, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8648048639297485, "num_tokens": 187857685.0, "step": 4946 }, { "epoch": 0.6293092481872535, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7515039443969727, "learning_rate": 1e-05, "loss": 0.5655, "mean_token_accuracy": 0.8370568752288818, "num_tokens": 187896614.0, "step": 4947 }, { "epoch": 0.6294364584658441, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.00248384475708, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8502461910247803, "num_tokens": 187925480.0, "step": 4948 }, { "epoch": 0.6295636687444346, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.890484094619751, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.8600244522094727, "num_tokens": 187965045.0, "step": 4949 }, { "epoch": 0.629690879023025, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.1327033042907715, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.857431173324585, "num_tokens": 188001633.0, "step": 4950 }, { "epoch": 0.6298180893016155, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.947462558746338, "learning_rate": 1e-05, "loss": 0.5124, "mean_token_accuracy": 0.8560264110565186, "num_tokens": 188039931.0, "step": 4951 }, { "epoch": 0.6299452995802061, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.325934410095215, "learning_rate": 1e-05, "loss": 0.5469, "mean_token_accuracy": 0.8418890833854675, "num_tokens": 188085353.0, "step": 4952 }, { "epoch": 0.6300725098587966, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8302345275878906, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8780209422111511, "num_tokens": 188118966.0, "step": 4953 }, { "epoch": 0.6301997201373871, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4112603664398193, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8819100856781006, "num_tokens": 188153680.0, "step": 4954 }, { "epoch": 0.6303269304159776, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6501946449279785, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.872402548789978, "num_tokens": 188189549.0, "step": 4955 }, { "epoch": 0.6304541406945681, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00013065338134765625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3804001808166504, "learning_rate": 1e-05, "loss": 0.5087, "mean_token_accuracy": 0.850863516330719, "num_tokens": 188230469.0, "step": 4956 }, { "epoch": 0.6305813509731586, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.486039400100708, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.859838604927063, "num_tokens": 188265074.0, "step": 4957 }, { "epoch": 0.6307085612517491, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3306686878204346, "learning_rate": 1e-05, "loss": 0.5065, "mean_token_accuracy": 0.8512527346611023, "num_tokens": 188302759.0, "step": 4958 }, { "epoch": 0.6308357715303397, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3832106590270996, "learning_rate": 1e-05, "loss": 0.5775, "mean_token_accuracy": 0.8305023908615112, "num_tokens": 188337627.0, "step": 4959 }, { "epoch": 0.6309629818089302, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2820305824279785, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8580633997917175, "num_tokens": 188365094.0, "step": 4960 }, { "epoch": 0.6310901920875207, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2884438037872314, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8616121411323547, "num_tokens": 188403244.0, "step": 4961 }, { "epoch": 0.6312174023661112, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.211041212081909, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8669261932373047, "num_tokens": 188442785.0, "step": 4962 }, { "epoch": 0.6313446126447017, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2597436904907227, "learning_rate": 1e-05, "loss": 0.5082, "mean_token_accuracy": 0.8508061170578003, "num_tokens": 188479298.0, "step": 4963 }, { "epoch": 0.6314718229232922, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1639719009399414, "learning_rate": 1e-05, "loss": 0.53, "mean_token_accuracy": 0.8416683673858643, "num_tokens": 188519304.0, "step": 4964 }, { "epoch": 0.6315990332018827, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.256540298461914, "learning_rate": 1e-05, "loss": 0.5228, "mean_token_accuracy": 0.8414901494979858, "num_tokens": 188559372.0, "step": 4965 }, { "epoch": 0.6317262434804732, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1144320964813232, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8731123208999634, "num_tokens": 188595337.0, "step": 4966 }, { "epoch": 0.6318534537590638, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2807185649871826, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8608899116516113, "num_tokens": 188632988.0, "step": 4967 }, { "epoch": 0.6319806640376543, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.053288698196411, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8716579675674438, "num_tokens": 188667205.0, "step": 4968 }, { "epoch": 0.6321078743162447, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.350207805633545, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.8536743521690369, "num_tokens": 188707717.0, "step": 4969 }, { "epoch": 0.6322350845948352, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9840807914733887, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8664481043815613, "num_tokens": 188742811.0, "step": 4970 }, { "epoch": 0.6323622948734258, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4689228534698486, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.858363151550293, "num_tokens": 188781004.0, "step": 4971 }, { "epoch": 0.6324895051520163, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9360971450805664, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8548256158828735, "num_tokens": 188813923.0, "step": 4972 }, { "epoch": 0.6326167154306068, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.629443645477295, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8616881370544434, "num_tokens": 188855375.0, "step": 4973 }, { "epoch": 0.6327439257091974, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.8975601196289062, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8747996091842651, "num_tokens": 188892807.0, "step": 4974 }, { "epoch": 0.6328711359877878, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.871807336807251, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8661158084869385, "num_tokens": 188932942.0, "step": 4975 }, { "epoch": 0.6329983462663783, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.0319361686706543, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8581690192222595, "num_tokens": 188968206.0, "step": 4976 }, { "epoch": 0.6331255565449688, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.051766872406006, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8517036437988281, "num_tokens": 189001315.0, "step": 4977 }, { "epoch": 0.6332527668235594, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.311039924621582, "learning_rate": 1e-05, "loss": 0.5417, "mean_token_accuracy": 0.8406504392623901, "num_tokens": 189039729.0, "step": 4978 }, { "epoch": 0.6333799771021499, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.967366933822632, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8752807378768921, "num_tokens": 189077455.0, "step": 4979 }, { "epoch": 0.6335071873807404, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.466858148574829, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8498330116271973, "num_tokens": 189118417.0, "step": 4980 }, { "epoch": 0.6336343976593308, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7649545669555664, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8707642555236816, "num_tokens": 189155581.0, "step": 4981 }, { "epoch": 0.6337616079379214, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.467571258544922, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.85837721824646, "num_tokens": 189191192.0, "step": 4982 }, { "epoch": 0.6338888182165119, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5688796043395996, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8649483323097229, "num_tokens": 189230475.0, "step": 4983 }, { "epoch": 0.6340160284951024, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3949832916259766, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8590741753578186, "num_tokens": 189265949.0, "step": 4984 }, { "epoch": 0.6341432387736929, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.420576333999634, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8567253351211548, "num_tokens": 189302914.0, "step": 4985 }, { "epoch": 0.6342704490522835, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.29817533493042, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8697985410690308, "num_tokens": 189347741.0, "step": 4986 }, { "epoch": 0.6343976593308739, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.327993154525757, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8681141138076782, "num_tokens": 189389672.0, "step": 4987 }, { "epoch": 0.6345248696094644, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2371442317962646, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8693469762802124, "num_tokens": 189425418.0, "step": 4988 }, { "epoch": 0.6346520798880549, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.297624349594116, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.869222104549408, "num_tokens": 189457042.0, "step": 4989 }, { "epoch": 0.6347792901666455, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.174062728881836, "learning_rate": 1e-05, "loss": 0.5126, "mean_token_accuracy": 0.848691999912262, "num_tokens": 189497135.0, "step": 4990 }, { "epoch": 0.634906500445236, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2860615253448486, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8784424066543579, "num_tokens": 189534829.0, "step": 4991 }, { "epoch": 0.6350337107238265, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.1104743480682373, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8535363674163818, "num_tokens": 189570535.0, "step": 4992 }, { "epoch": 0.635160921002417, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.310267210006714, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8667072057723999, "num_tokens": 189606448.0, "step": 4993 }, { "epoch": 0.6352881312810075, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0401554107666016, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8608798980712891, "num_tokens": 189650880.0, "step": 4994 }, { "epoch": 0.635415341559598, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3826818466186523, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8446804285049438, "num_tokens": 189690866.0, "step": 4995 }, { "epoch": 0.6355425518381885, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9877190589904785, "learning_rate": 1e-05, "loss": 0.561, "mean_token_accuracy": 0.8308819532394409, "num_tokens": 189726891.0, "step": 4996 }, { "epoch": 0.6356697621167791, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.502612590789795, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8602586984634399, "num_tokens": 189763893.0, "step": 4997 }, { "epoch": 0.6357969723953696, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.922670364379883, "learning_rate": 1e-05, "loss": 0.523, "mean_token_accuracy": 0.843359112739563, "num_tokens": 189804122.0, "step": 4998 }, { "epoch": 0.63592418267396, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.698451519012451, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8564493060112, "num_tokens": 189836355.0, "step": 4999 }, { "epoch": 0.6360513929525505, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.9306750297546387, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8624185919761658, "num_tokens": 189871839.0, "step": 5000 }, { "epoch": 0.6361786032311411, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9058279991149902, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8696320056915283, "num_tokens": 189915631.0, "step": 5001 }, { "epoch": 0.6363058135097316, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.08687424659729, "learning_rate": 1e-05, "loss": 0.4981, "mean_token_accuracy": 0.8521115183830261, "num_tokens": 189953635.0, "step": 5002 }, { "epoch": 0.6364330237883221, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.021792411804199, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8618238568305969, "num_tokens": 189999338.0, "step": 5003 }, { "epoch": 0.6365602340669126, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.342653512954712, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8600858449935913, "num_tokens": 190045804.0, "step": 5004 }, { "epoch": 0.6366874443455031, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.917092800140381, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8552051782608032, "num_tokens": 190083350.0, "step": 5005 }, { "epoch": 0.6368146546240936, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.48003888130188, "learning_rate": 1e-05, "loss": 0.5126, "mean_token_accuracy": 0.8513972759246826, "num_tokens": 190121910.0, "step": 5006 }, { "epoch": 0.6369418649026841, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6936635971069336, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8617026209831238, "num_tokens": 190163857.0, "step": 5007 }, { "epoch": 0.6370690751812746, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.446981430053711, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8526724576950073, "num_tokens": 190204337.0, "step": 5008 }, { "epoch": 0.6371962854598652, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5065059661865234, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8714816570281982, "num_tokens": 190240646.0, "step": 5009 }, { "epoch": 0.6373234957384557, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.384291887283325, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8549591898918152, "num_tokens": 190273507.0, "step": 5010 }, { "epoch": 0.6374507060170462, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.365104913711548, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8589730858802795, "num_tokens": 190306876.0, "step": 5011 }, { "epoch": 0.6375779162956366, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3210408687591553, "learning_rate": 1e-05, "loss": 0.5042, "mean_token_accuracy": 0.8485947847366333, "num_tokens": 190343126.0, "step": 5012 }, { "epoch": 0.6377051265742272, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.265986680984497, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.849446713924408, "num_tokens": 190382089.0, "step": 5013 }, { "epoch": 0.6378323368528177, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.258408784866333, "learning_rate": 1e-05, "loss": 0.5126, "mean_token_accuracy": 0.8461329936981201, "num_tokens": 190419577.0, "step": 5014 }, { "epoch": 0.6379595471314082, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2156004905700684, "learning_rate": 1e-05, "loss": 0.5317, "mean_token_accuracy": 0.8460389375686646, "num_tokens": 190459725.0, "step": 5015 }, { "epoch": 0.6380867574099988, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2259514331817627, "learning_rate": 1e-05, "loss": 0.5244, "mean_token_accuracy": 0.847637414932251, "num_tokens": 190493575.0, "step": 5016 }, { "epoch": 0.6382139676885893, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.177198886871338, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8608482480049133, "num_tokens": 190530326.0, "step": 5017 }, { "epoch": 0.6383411779671797, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.205284833908081, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8569629192352295, "num_tokens": 190565727.0, "step": 5018 }, { "epoch": 0.6384683882457702, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1725223064422607, "learning_rate": 1e-05, "loss": 0.515, "mean_token_accuracy": 0.8477540016174316, "num_tokens": 190600163.0, "step": 5019 }, { "epoch": 0.6385955985243608, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.184283494949341, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8499279618263245, "num_tokens": 190640801.0, "step": 5020 }, { "epoch": 0.6387228088029513, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.140061140060425, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8781481981277466, "num_tokens": 190679886.0, "step": 5021 }, { "epoch": 0.6388500190815418, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1711502075195312, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8571356534957886, "num_tokens": 190726246.0, "step": 5022 }, { "epoch": 0.6389772293601323, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1221139430999756, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8582763671875, "num_tokens": 190765496.0, "step": 5023 }, { "epoch": 0.6391044396387228, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.184248924255371, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8539882898330688, "num_tokens": 190808770.0, "step": 5024 }, { "epoch": 0.6392316499173133, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.113945960998535, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8678299784660339, "num_tokens": 190844242.0, "step": 5025 }, { "epoch": 0.6393588601959038, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2023768424987793, "learning_rate": 1e-05, "loss": 0.5389, "mean_token_accuracy": 0.839277982711792, "num_tokens": 190887658.0, "step": 5026 }, { "epoch": 0.6394860704744944, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0807955265045166, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8621013164520264, "num_tokens": 190922322.0, "step": 5027 }, { "epoch": 0.6396132807530849, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.236085891723633, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8530162572860718, "num_tokens": 190959085.0, "step": 5028 }, { "epoch": 0.6397404910316754, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.0410709381103516, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.872412919998169, "num_tokens": 190991646.0, "step": 5029 }, { "epoch": 0.6398677013102658, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.285661220550537, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8592332005500793, "num_tokens": 191039489.0, "step": 5030 }, { "epoch": 0.6399949115888564, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.981645107269287, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8540299534797668, "num_tokens": 191078166.0, "step": 5031 }, { "epoch": 0.6401221218674469, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3972578048706055, "learning_rate": 1e-05, "loss": 0.5389, "mean_token_accuracy": 0.8376458883285522, "num_tokens": 191115981.0, "step": 5032 }, { "epoch": 0.6402493321460374, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.8974931240081787, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8595324158668518, "num_tokens": 191153989.0, "step": 5033 }, { "epoch": 0.6403765424246279, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.603179931640625, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8594678640365601, "num_tokens": 191192799.0, "step": 5034 }, { "epoch": 0.6405037527032185, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.807694435119629, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.8572799563407898, "num_tokens": 191231923.0, "step": 5035 }, { "epoch": 0.6406309629818089, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9489269256591797, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.857912003993988, "num_tokens": 191271650.0, "step": 5036 }, { "epoch": 0.6407581732603994, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 2.942721366882324, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8605888485908508, "num_tokens": 191300212.0, "step": 5037 }, { "epoch": 0.6408853835389899, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.267919063568115, "learning_rate": 1e-05, "loss": 0.5083, "mean_token_accuracy": 0.8538175821304321, "num_tokens": 191336163.0, "step": 5038 }, { "epoch": 0.6410125938175805, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.367635488510132, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8563293218612671, "num_tokens": 191374749.0, "step": 5039 }, { "epoch": 0.641139804096171, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.244387626647949, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8603806495666504, "num_tokens": 191413342.0, "step": 5040 }, { "epoch": 0.6412670143747615, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.6966824531555176, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8646224737167358, "num_tokens": 191450332.0, "step": 5041 }, { "epoch": 0.6413942246533519, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.96905779838562, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8619160652160645, "num_tokens": 191486969.0, "step": 5042 }, { "epoch": 0.6415214349319425, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.656248092651367, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8619965314865112, "num_tokens": 191525074.0, "step": 5043 }, { "epoch": 0.641648645210533, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7545669078826904, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8572185039520264, "num_tokens": 191564324.0, "step": 5044 }, { "epoch": 0.6417758554891235, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.506430149078369, "learning_rate": 1e-05, "loss": 0.5118, "mean_token_accuracy": 0.8519809246063232, "num_tokens": 191604081.0, "step": 5045 }, { "epoch": 0.641903065767714, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.628216028213501, "learning_rate": 1e-05, "loss": 0.5174, "mean_token_accuracy": 0.8444858193397522, "num_tokens": 191639811.0, "step": 5046 }, { "epoch": 0.6420302760463046, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.356234550476074, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.866473913192749, "num_tokens": 191680941.0, "step": 5047 }, { "epoch": 0.642157486324895, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5614163875579834, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8519834876060486, "num_tokens": 191718039.0, "step": 5048 }, { "epoch": 0.6422846966034855, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2294504642486572, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8588583469390869, "num_tokens": 191759837.0, "step": 5049 }, { "epoch": 0.6424119068820761, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.571611166000366, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8540382385253906, "num_tokens": 191801791.0, "step": 5050 }, { "epoch": 0.6425391171606666, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.125986337661743, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8679158687591553, "num_tokens": 191838193.0, "step": 5051 }, { "epoch": 0.6426663274392571, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6722655296325684, "learning_rate": 1e-05, "loss": 0.5734, "mean_token_accuracy": 0.8312703371047974, "num_tokens": 191874811.0, "step": 5052 }, { "epoch": 0.6427935377178476, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.075679302215576, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.8520673513412476, "num_tokens": 191921289.0, "step": 5053 }, { "epoch": 0.6429207479964381, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.81723952293396, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8605769276618958, "num_tokens": 191961796.0, "step": 5054 }, { "epoch": 0.6430479582750286, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1369690895080566, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8769671320915222, "num_tokens": 192000853.0, "step": 5055 }, { "epoch": 0.6431751685536191, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.9372270107269287, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8616833686828613, "num_tokens": 192029126.0, "step": 5056 }, { "epoch": 0.6433023788322096, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.330735206604004, "learning_rate": 1e-05, "loss": 0.5034, "mean_token_accuracy": 0.85235995054245, "num_tokens": 192070472.0, "step": 5057 }, { "epoch": 0.6434295891108002, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8349084854125977, "learning_rate": 1e-05, "loss": 0.5314, "mean_token_accuracy": 0.8476096391677856, "num_tokens": 192107844.0, "step": 5058 }, { "epoch": 0.6435567993893907, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4586994647979736, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8601433038711548, "num_tokens": 192140637.0, "step": 5059 }, { "epoch": 0.6436840096679812, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.603830099105835, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8665796518325806, "num_tokens": 192176177.0, "step": 5060 }, { "epoch": 0.6438112199465716, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4409801959991455, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.863554835319519, "num_tokens": 192218131.0, "step": 5061 }, { "epoch": 0.6439384302251622, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3839316368103027, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8627623319625854, "num_tokens": 192258720.0, "step": 5062 }, { "epoch": 0.6440656405037527, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.410820960998535, "learning_rate": 1e-05, "loss": 0.526, "mean_token_accuracy": 0.8477622866630554, "num_tokens": 192303197.0, "step": 5063 }, { "epoch": 0.6441928507823432, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2333147525787354, "learning_rate": 1e-05, "loss": 0.5573, "mean_token_accuracy": 0.8356140851974487, "num_tokens": 192340882.0, "step": 5064 }, { "epoch": 0.6443200610609338, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.380068778991699, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8625190854072571, "num_tokens": 192383596.0, "step": 5065 }, { "epoch": 0.6444472713395243, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.10882306098938, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.858104944229126, "num_tokens": 192423861.0, "step": 5066 }, { "epoch": 0.6445744816181147, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.4084525108337402, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8655755519866943, "num_tokens": 192456564.0, "step": 5067 }, { "epoch": 0.6447016918967052, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0248329639434814, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8552805781364441, "num_tokens": 192496917.0, "step": 5068 }, { "epoch": 0.6448289021752958, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.483200788497925, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8566740155220032, "num_tokens": 192539548.0, "step": 5069 }, { "epoch": 0.6449561124538863, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.972008466720581, "learning_rate": 1e-05, "loss": 0.5199, "mean_token_accuracy": 0.8471728563308716, "num_tokens": 192577430.0, "step": 5070 }, { "epoch": 0.6450833227324768, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.618971586227417, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8696725964546204, "num_tokens": 192617343.0, "step": 5071 }, { "epoch": 0.6452105330110673, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.965083122253418, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.86113041639328, "num_tokens": 192662274.0, "step": 5072 }, { "epoch": 0.6453377432896578, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7861902713775635, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8659263849258423, "num_tokens": 192703305.0, "step": 5073 }, { "epoch": 0.6454649535682483, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0403339862823486, "learning_rate": 1e-05, "loss": 0.4961, "mean_token_accuracy": 0.8542660474777222, "num_tokens": 192738325.0, "step": 5074 }, { "epoch": 0.6455921638468388, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9210891723632812, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8724683523178101, "num_tokens": 192778285.0, "step": 5075 }, { "epoch": 0.6457193741254293, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2090628147125244, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8630790710449219, "num_tokens": 192822822.0, "step": 5076 }, { "epoch": 0.6458465844040199, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9178192615509033, "learning_rate": 1e-05, "loss": 0.5168, "mean_token_accuracy": 0.8555939197540283, "num_tokens": 192857312.0, "step": 5077 }, { "epoch": 0.6459737946826104, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3865857124328613, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8656907081604004, "num_tokens": 192892363.0, "step": 5078 }, { "epoch": 0.6461010049612008, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7550957202911377, "learning_rate": 1e-05, "loss": 0.5129, "mean_token_accuracy": 0.8491249084472656, "num_tokens": 192931774.0, "step": 5079 }, { "epoch": 0.6462282152397913, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4367480278015137, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8652859926223755, "num_tokens": 192964463.0, "step": 5080 }, { "epoch": 0.6463554255183819, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5498709678649902, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.870078444480896, "num_tokens": 192999807.0, "step": 5081 }, { "epoch": 0.6464826357969724, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.378058910369873, "learning_rate": 1e-05, "loss": 0.4973, "mean_token_accuracy": 0.8533877730369568, "num_tokens": 193041408.0, "step": 5082 }, { "epoch": 0.6466098460755629, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4083845615386963, "learning_rate": 1e-05, "loss": 0.5649, "mean_token_accuracy": 0.8329830169677734, "num_tokens": 193078104.0, "step": 5083 }, { "epoch": 0.6467370563541535, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.317702054977417, "learning_rate": 1e-05, "loss": 0.5301, "mean_token_accuracy": 0.8416121006011963, "num_tokens": 193118821.0, "step": 5084 }, { "epoch": 0.6468642666327439, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.301985025405884, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8544332981109619, "num_tokens": 193152633.0, "step": 5085 }, { "epoch": 0.6469914769113344, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.257723569869995, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8622313141822815, "num_tokens": 193191398.0, "step": 5086 }, { "epoch": 0.6471186871899249, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2422571182250977, "learning_rate": 1e-05, "loss": 0.5084, "mean_token_accuracy": 0.8501176238059998, "num_tokens": 193229612.0, "step": 5087 }, { "epoch": 0.6472458974685155, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.21321439743042, "learning_rate": 1e-05, "loss": 0.5154, "mean_token_accuracy": 0.8441578149795532, "num_tokens": 193271993.0, "step": 5088 }, { "epoch": 0.647373107747106, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2327938079833984, "learning_rate": 1e-05, "loss": 0.5159, "mean_token_accuracy": 0.849480926990509, "num_tokens": 193308146.0, "step": 5089 }, { "epoch": 0.6475003180256965, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1572511196136475, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8634290099143982, "num_tokens": 193348448.0, "step": 5090 }, { "epoch": 0.6476275283042869, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2434377670288086, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8593089580535889, "num_tokens": 193379279.0, "step": 5091 }, { "epoch": 0.6477547385828775, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.1197879314422607, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.863850474357605, "num_tokens": 193415999.0, "step": 5092 }, { "epoch": 0.647881948861468, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2688379287719727, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8623532056808472, "num_tokens": 193450287.0, "step": 5093 }, { "epoch": 0.6480091591400585, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.06508469581604, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8622176647186279, "num_tokens": 193485980.0, "step": 5094 }, { "epoch": 0.648136369418649, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3354482650756836, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8533512949943542, "num_tokens": 193525179.0, "step": 5095 }, { "epoch": 0.6482635796972396, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.0061194896698, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.856741726398468, "num_tokens": 193561580.0, "step": 5096 }, { "epoch": 0.64839078997583, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4453768730163574, "learning_rate": 1e-05, "loss": 0.5261, "mean_token_accuracy": 0.8485943675041199, "num_tokens": 193598218.0, "step": 5097 }, { "epoch": 0.6485180002544205, "ewc_loss": 0.052734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.932278871536255, "learning_rate": 1e-05, "loss": 0.5145, "mean_token_accuracy": 0.8475898504257202, "num_tokens": 193635256.0, "step": 5098 }, { "epoch": 0.648645210533011, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.624084949493408, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.849359929561615, "num_tokens": 193672957.0, "step": 5099 }, { "epoch": 0.6487724208116016, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.8770227432250977, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8681726455688477, "num_tokens": 193710575.0, "step": 5100 }, { "epoch": 0.6488996310901921, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.884798288345337, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.867444634437561, "num_tokens": 193749006.0, "step": 5101 }, { "epoch": 0.6490268413687826, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 2.987344980239868, "learning_rate": 1e-05, "loss": 0.5432, "mean_token_accuracy": 0.8419955968856812, "num_tokens": 193788975.0, "step": 5102 }, { "epoch": 0.649154051647373, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.12603235244751, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8578822612762451, "num_tokens": 193822898.0, "step": 5103 }, { "epoch": 0.6492812619259636, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.310225486755371, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8724383115768433, "num_tokens": 193862439.0, "step": 5104 }, { "epoch": 0.6494084722045541, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.089447975158691, "learning_rate": 1e-05, "loss": 0.5136, "mean_token_accuracy": 0.8571423292160034, "num_tokens": 193906721.0, "step": 5105 }, { "epoch": 0.6495356824831446, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5574018955230713, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8621769547462463, "num_tokens": 193940521.0, "step": 5106 }, { "epoch": 0.6496628927617352, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.865394353866577, "learning_rate": 1e-05, "loss": 0.5124, "mean_token_accuracy": 0.8497264385223389, "num_tokens": 193979340.0, "step": 5107 }, { "epoch": 0.6497901030403257, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.543158531188965, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8584601879119873, "num_tokens": 194018562.0, "step": 5108 }, { "epoch": 0.6499173133189162, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6637182235717773, "learning_rate": 1e-05, "loss": 0.558, "mean_token_accuracy": 0.8311071395874023, "num_tokens": 194056804.0, "step": 5109 }, { "epoch": 0.6500445235975066, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.441985607147217, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8601722717285156, "num_tokens": 194095194.0, "step": 5110 }, { "epoch": 0.6501717338760972, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5085012912750244, "learning_rate": 1e-05, "loss": 0.4941, "mean_token_accuracy": 0.856772780418396, "num_tokens": 194135061.0, "step": 5111 }, { "epoch": 0.6502989441546877, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3326215744018555, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.862612247467041, "num_tokens": 194169017.0, "step": 5112 }, { "epoch": 0.6504261544332782, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.417963743209839, "learning_rate": 1e-05, "loss": 0.5046, "mean_token_accuracy": 0.8495577573776245, "num_tokens": 194213424.0, "step": 5113 }, { "epoch": 0.6505533647118688, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2286221981048584, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8685725927352905, "num_tokens": 194247465.0, "step": 5114 }, { "epoch": 0.6506805749904593, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.388679027557373, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8654052019119263, "num_tokens": 194282575.0, "step": 5115 }, { "epoch": 0.6508077852690497, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.160085678100586, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8598418831825256, "num_tokens": 194311822.0, "step": 5116 }, { "epoch": 0.6509349955476402, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4118130207061768, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8479225635528564, "num_tokens": 194346990.0, "step": 5117 }, { "epoch": 0.6510622058262308, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0549368858337402, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8605653047561646, "num_tokens": 194384836.0, "step": 5118 }, { "epoch": 0.6511894161048213, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5002124309539795, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8706260919570923, "num_tokens": 194418080.0, "step": 5119 }, { "epoch": 0.6513166263834118, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.9897236824035645, "learning_rate": 1e-05, "loss": 0.5188, "mean_token_accuracy": 0.8470802903175354, "num_tokens": 194460111.0, "step": 5120 }, { "epoch": 0.6514438366620023, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6515960693359375, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8736501932144165, "num_tokens": 194498157.0, "step": 5121 }, { "epoch": 0.6515710469405928, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.9757235050201416, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8605148792266846, "num_tokens": 194538928.0, "step": 5122 }, { "epoch": 0.6516982572191833, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8437273502349854, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8610641956329346, "num_tokens": 194582276.0, "step": 5123 }, { "epoch": 0.6518254674977738, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.0940866470336914, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8591972589492798, "num_tokens": 194620197.0, "step": 5124 }, { "epoch": 0.6519526777763643, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9456489086151123, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8656433820724487, "num_tokens": 194654634.0, "step": 5125 }, { "epoch": 0.6520798880549549, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.311363697052002, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8534557819366455, "num_tokens": 194696400.0, "step": 5126 }, { "epoch": 0.6522070983335454, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001316070556640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8336563110351562, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8653501272201538, "num_tokens": 194738326.0, "step": 5127 }, { "epoch": 0.6523343086121358, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.4269134998321533, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8624062538146973, "num_tokens": 194778550.0, "step": 5128 }, { "epoch": 0.6524615188907263, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6338675022125244, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8554118871688843, "num_tokens": 194815124.0, "step": 5129 }, { "epoch": 0.6525887291693169, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4321541786193848, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8602835536003113, "num_tokens": 194851372.0, "step": 5130 }, { "epoch": 0.6527159394479074, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4164998531341553, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8704315423965454, "num_tokens": 194886705.0, "step": 5131 }, { "epoch": 0.6528431497264979, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.381330728530884, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8594592809677124, "num_tokens": 194930372.0, "step": 5132 }, { "epoch": 0.6529703600050885, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2555830478668213, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8617227077484131, "num_tokens": 194971869.0, "step": 5133 }, { "epoch": 0.6530975702836789, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.36720609664917, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8525087237358093, "num_tokens": 195011271.0, "step": 5134 }, { "epoch": 0.6532247805622694, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1578335762023926, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8607832193374634, "num_tokens": 195046328.0, "step": 5135 }, { "epoch": 0.6533519908408599, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.359025478363037, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8647481799125671, "num_tokens": 195086424.0, "step": 5136 }, { "epoch": 0.6534792011194505, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.089855670928955, "learning_rate": 1e-05, "loss": 0.5022, "mean_token_accuracy": 0.8524397015571594, "num_tokens": 195121751.0, "step": 5137 }, { "epoch": 0.653606411398041, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.400486469268799, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8613449335098267, "num_tokens": 195158231.0, "step": 5138 }, { "epoch": 0.6537336216766315, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0160064697265625, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.857681930065155, "num_tokens": 195195009.0, "step": 5139 }, { "epoch": 0.6538608319552219, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.485002040863037, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8690611720085144, "num_tokens": 195232892.0, "step": 5140 }, { "epoch": 0.6539880422338125, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.9436001777648926, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8580886125564575, "num_tokens": 195283233.0, "step": 5141 }, { "epoch": 0.654115252512403, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.644793748855591, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8780804872512817, "num_tokens": 195319402.0, "step": 5142 }, { "epoch": 0.6542424627909935, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9549055099487305, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.864705502986908, "num_tokens": 195358330.0, "step": 5143 }, { "epoch": 0.654369673069584, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.819850206375122, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8728188276290894, "num_tokens": 195398074.0, "step": 5144 }, { "epoch": 0.6544968833481746, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0462045669555664, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8731876015663147, "num_tokens": 195435684.0, "step": 5145 }, { "epoch": 0.654624093626765, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9547059535980225, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8706607222557068, "num_tokens": 195475126.0, "step": 5146 }, { "epoch": 0.6547513039053555, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2525904178619385, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8559184670448303, "num_tokens": 195516398.0, "step": 5147 }, { "epoch": 0.654878514183946, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.912100076675415, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8631740808486938, "num_tokens": 195558777.0, "step": 5148 }, { "epoch": 0.6550057244625366, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4014179706573486, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8540439009666443, "num_tokens": 195598663.0, "step": 5149 }, { "epoch": 0.6551329347411271, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7588460445404053, "learning_rate": 1e-05, "loss": 0.516, "mean_token_accuracy": 0.8455373644828796, "num_tokens": 195628594.0, "step": 5150 }, { "epoch": 0.6552601450197176, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.43070650100708, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8529589176177979, "num_tokens": 195662484.0, "step": 5151 }, { "epoch": 0.655387355298308, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.558396577835083, "learning_rate": 1e-05, "loss": 0.5143, "mean_token_accuracy": 0.8500690460205078, "num_tokens": 195697753.0, "step": 5152 }, { "epoch": 0.6555145655768986, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3607795238494873, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8706258535385132, "num_tokens": 195740931.0, "step": 5153 }, { "epoch": 0.6556417758554891, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.407597303390503, "learning_rate": 1e-05, "loss": 0.5258, "mean_token_accuracy": 0.8461266160011292, "num_tokens": 195783056.0, "step": 5154 }, { "epoch": 0.6557689861340796, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3197171688079834, "learning_rate": 1e-05, "loss": 0.5232, "mean_token_accuracy": 0.8475363254547119, "num_tokens": 195819515.0, "step": 5155 }, { "epoch": 0.6558961964126702, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3096160888671875, "learning_rate": 1e-05, "loss": 0.5271, "mean_token_accuracy": 0.8450988531112671, "num_tokens": 195861039.0, "step": 5156 }, { "epoch": 0.6560234066912607, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.262821912765503, "learning_rate": 1e-05, "loss": 0.5296, "mean_token_accuracy": 0.843230664730072, "num_tokens": 195897818.0, "step": 5157 }, { "epoch": 0.6561506169698512, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.251523733139038, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8746287822723389, "num_tokens": 195930752.0, "step": 5158 }, { "epoch": 0.6562778272484416, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2193076610565186, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.854263424873352, "num_tokens": 195965036.0, "step": 5159 }, { "epoch": 0.6564050375270322, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2339792251586914, "learning_rate": 1e-05, "loss": 0.4997, "mean_token_accuracy": 0.8580451607704163, "num_tokens": 195999293.0, "step": 5160 }, { "epoch": 0.6565322478056227, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1725099086761475, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8491872549057007, "num_tokens": 196039066.0, "step": 5161 }, { "epoch": 0.6566594580842132, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2243082523345947, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8507943749427795, "num_tokens": 196077805.0, "step": 5162 }, { "epoch": 0.6567866683628037, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1308300495147705, "learning_rate": 1e-05, "loss": 0.5402, "mean_token_accuracy": 0.8374909162521362, "num_tokens": 196117460.0, "step": 5163 }, { "epoch": 0.6569138786413943, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2409138679504395, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8604583144187927, "num_tokens": 196154855.0, "step": 5164 }, { "epoch": 0.6570410889199847, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0836188793182373, "learning_rate": 1e-05, "loss": 0.5255, "mean_token_accuracy": 0.8431771993637085, "num_tokens": 196197360.0, "step": 5165 }, { "epoch": 0.6571682991985752, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2843072414398193, "learning_rate": 1e-05, "loss": 0.5508, "mean_token_accuracy": 0.8371521234512329, "num_tokens": 196237419.0, "step": 5166 }, { "epoch": 0.6572955094771658, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0351357460021973, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8485564589500427, "num_tokens": 196275533.0, "step": 5167 }, { "epoch": 0.6574227197557563, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3522493839263916, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8652729988098145, "num_tokens": 196312269.0, "step": 5168 }, { "epoch": 0.6575499300343468, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.9713973999023438, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8533203601837158, "num_tokens": 196349338.0, "step": 5169 }, { "epoch": 0.6576771403129373, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4849424362182617, "learning_rate": 1e-05, "loss": 0.5297, "mean_token_accuracy": 0.8462738990783691, "num_tokens": 196396000.0, "step": 5170 }, { "epoch": 0.6578043505915278, "ewc_loss": 0.052978515625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9081337451934814, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8517948389053345, "num_tokens": 196428119.0, "step": 5171 }, { "epoch": 0.6579315608701183, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.68723726272583, "learning_rate": 1e-05, "loss": 0.5372, "mean_token_accuracy": 0.8448049426078796, "num_tokens": 196470466.0, "step": 5172 }, { "epoch": 0.6580587711487088, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.878842353820801, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8590427041053772, "num_tokens": 196508883.0, "step": 5173 }, { "epoch": 0.6581859814272993, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9621315002441406, "learning_rate": 1e-05, "loss": 0.5231, "mean_token_accuracy": 0.8459237813949585, "num_tokens": 196552762.0, "step": 5174 }, { "epoch": 0.6583131917058899, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.0623795986175537, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8787208795547485, "num_tokens": 196581394.0, "step": 5175 }, { "epoch": 0.6584404019844804, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.133834362030029, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8644028902053833, "num_tokens": 196623237.0, "step": 5176 }, { "epoch": 0.6585676122630708, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00013256072998046875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3854737281799316, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8709379434585571, "num_tokens": 196659219.0, "step": 5177 }, { "epoch": 0.6586948225416613, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.045476913452148, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8534016013145447, "num_tokens": 196700102.0, "step": 5178 }, { "epoch": 0.6588220328202519, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5476746559143066, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8856116533279419, "num_tokens": 196740569.0, "step": 5179 }, { "epoch": 0.6589492430988424, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.832214117050171, "learning_rate": 1e-05, "loss": 0.5088, "mean_token_accuracy": 0.8533463478088379, "num_tokens": 196778634.0, "step": 5180 }, { "epoch": 0.6590764533774329, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5462682247161865, "learning_rate": 1e-05, "loss": 0.5207, "mean_token_accuracy": 0.8485739827156067, "num_tokens": 196805021.0, "step": 5181 }, { "epoch": 0.6592036636560235, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6294970512390137, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8618270754814148, "num_tokens": 196839676.0, "step": 5182 }, { "epoch": 0.6593308739346139, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4088191986083984, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8656736612319946, "num_tokens": 196877564.0, "step": 5183 }, { "epoch": 0.6594580842132044, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.507690668106079, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.8513054847717285, "num_tokens": 196913819.0, "step": 5184 }, { "epoch": 0.6595852944917949, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3078083992004395, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8448343276977539, "num_tokens": 196951019.0, "step": 5185 }, { "epoch": 0.6597125047703855, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.422424793243408, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8576545715332031, "num_tokens": 196995112.0, "step": 5186 }, { "epoch": 0.659839715048976, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.200813055038452, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.863544225692749, "num_tokens": 197035334.0, "step": 5187 }, { "epoch": 0.6599669253275665, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4219918251037598, "learning_rate": 1e-05, "loss": 0.5131, "mean_token_accuracy": 0.8497275710105896, "num_tokens": 197074234.0, "step": 5188 }, { "epoch": 0.6600941356061569, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.118056535720825, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8528266549110413, "num_tokens": 197110977.0, "step": 5189 }, { "epoch": 0.6602213458847475, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.479548215866089, "learning_rate": 1e-05, "loss": 0.5212, "mean_token_accuracy": 0.848007321357727, "num_tokens": 197148634.0, "step": 5190 }, { "epoch": 0.660348556163338, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0339038372039795, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8632550835609436, "num_tokens": 197186361.0, "step": 5191 }, { "epoch": 0.6604757664419285, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.593775987625122, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8566877841949463, "num_tokens": 197222642.0, "step": 5192 }, { "epoch": 0.660602976720519, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9843599796295166, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8735038042068481, "num_tokens": 197263320.0, "step": 5193 }, { "epoch": 0.6607301869991096, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7916383743286133, "learning_rate": 1e-05, "loss": 0.5199, "mean_token_accuracy": 0.8475499153137207, "num_tokens": 197294041.0, "step": 5194 }, { "epoch": 0.6608573972777, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.073026657104492, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8640487194061279, "num_tokens": 197324878.0, "step": 5195 }, { "epoch": 0.6609846075562905, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8898394107818604, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8690210580825806, "num_tokens": 197368318.0, "step": 5196 }, { "epoch": 0.661111817834881, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2318286895751953, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8623005747795105, "num_tokens": 197407665.0, "step": 5197 }, { "epoch": 0.6612390281134716, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.860747814178467, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8592497110366821, "num_tokens": 197445812.0, "step": 5198 }, { "epoch": 0.6613662383920621, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.381632089614868, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8677332401275635, "num_tokens": 197482733.0, "step": 5199 }, { "epoch": 0.6614934486706526, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.684314012527466, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.858393132686615, "num_tokens": 197520874.0, "step": 5200 }, { "epoch": 0.661620658949243, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.413325309753418, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8556027412414551, "num_tokens": 197563422.0, "step": 5201 }, { "epoch": 0.6617478692278336, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.476534366607666, "learning_rate": 1e-05, "loss": 0.4985, "mean_token_accuracy": 0.8512563705444336, "num_tokens": 197603632.0, "step": 5202 }, { "epoch": 0.6618750795064241, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.395597219467163, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.861137330532074, "num_tokens": 197642090.0, "step": 5203 }, { "epoch": 0.6620022897850146, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2997348308563232, "learning_rate": 1e-05, "loss": 0.573, "mean_token_accuracy": 0.8310040235519409, "num_tokens": 197679294.0, "step": 5204 }, { "epoch": 0.6621295000636052, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.375171422958374, "learning_rate": 1e-05, "loss": 0.558, "mean_token_accuracy": 0.8389859199523926, "num_tokens": 197721254.0, "step": 5205 }, { "epoch": 0.6622567103421957, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1593897342681885, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8767192363739014, "num_tokens": 197764656.0, "step": 5206 }, { "epoch": 0.6623839206207861, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.376528263092041, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8614726066589355, "num_tokens": 197805900.0, "step": 5207 }, { "epoch": 0.6625111308993766, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0826199054718018, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8627205491065979, "num_tokens": 197842752.0, "step": 5208 }, { "epoch": 0.6626383411779672, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4170539379119873, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8685360550880432, "num_tokens": 197883531.0, "step": 5209 }, { "epoch": 0.6627655514565577, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.008634328842163, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8630958795547485, "num_tokens": 197920875.0, "step": 5210 }, { "epoch": 0.6628927617351482, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5251667499542236, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8614435195922852, "num_tokens": 197955194.0, "step": 5211 }, { "epoch": 0.6630199720137387, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.9510180950164795, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8707829713821411, "num_tokens": 197991730.0, "step": 5212 }, { "epoch": 0.6631471822923293, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6728875637054443, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8643335103988647, "num_tokens": 198030032.0, "step": 5213 }, { "epoch": 0.6632743925709197, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.9625961780548096, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8630691766738892, "num_tokens": 198068274.0, "step": 5214 }, { "epoch": 0.6634016028495102, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8555564880371094, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8549489974975586, "num_tokens": 198105326.0, "step": 5215 }, { "epoch": 0.6635288131281007, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.0819785594940186, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8727868795394897, "num_tokens": 198147485.0, "step": 5216 }, { "epoch": 0.6636560234066913, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9567205905914307, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8646076917648315, "num_tokens": 198184928.0, "step": 5217 }, { "epoch": 0.6637832336852818, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3046159744262695, "learning_rate": 1e-05, "loss": 0.4824, "mean_token_accuracy": 0.8620794415473938, "num_tokens": 198224641.0, "step": 5218 }, { "epoch": 0.6639104439638723, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8584513664245605, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8705234527587891, "num_tokens": 198263821.0, "step": 5219 }, { "epoch": 0.6640376542424627, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4082696437835693, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8622210621833801, "num_tokens": 198300315.0, "step": 5220 }, { "epoch": 0.6641648645210533, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.683758497238159, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8646031618118286, "num_tokens": 198336543.0, "step": 5221 }, { "epoch": 0.6642920747996438, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.412773609161377, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8624598979949951, "num_tokens": 198374728.0, "step": 5222 }, { "epoch": 0.6644192850782343, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4968156814575195, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8630788922309875, "num_tokens": 198414504.0, "step": 5223 }, { "epoch": 0.6645464953568249, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3503386974334717, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8718599081039429, "num_tokens": 198449976.0, "step": 5224 }, { "epoch": 0.6646737056354154, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3554775714874268, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.865453839302063, "num_tokens": 198489522.0, "step": 5225 }, { "epoch": 0.6648009159140058, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.297520399093628, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8696651458740234, "num_tokens": 198524460.0, "step": 5226 }, { "epoch": 0.6649281261925963, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2637441158294678, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8654877543449402, "num_tokens": 198569139.0, "step": 5227 }, { "epoch": 0.6650553364711869, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.259495258331299, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8785074353218079, "num_tokens": 198597166.0, "step": 5228 }, { "epoch": 0.6651825467497774, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.230340003967285, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8651046752929688, "num_tokens": 198631646.0, "step": 5229 }, { "epoch": 0.6653097570283679, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2002952098846436, "learning_rate": 1e-05, "loss": 0.5142, "mean_token_accuracy": 0.8491789698600769, "num_tokens": 198675743.0, "step": 5230 }, { "epoch": 0.6654369673069584, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2121379375457764, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8572779893875122, "num_tokens": 198715826.0, "step": 5231 }, { "epoch": 0.6655641775855489, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.158895969390869, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8595044016838074, "num_tokens": 198757320.0, "step": 5232 }, { "epoch": 0.6656913878641394, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.230055809020996, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8681811690330505, "num_tokens": 198791110.0, "step": 5233 }, { "epoch": 0.6658185981427299, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1305572986602783, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8664050698280334, "num_tokens": 198828506.0, "step": 5234 }, { "epoch": 0.6659458084213205, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2520265579223633, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8446060419082642, "num_tokens": 198870759.0, "step": 5235 }, { "epoch": 0.666073018699911, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0792007446289062, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8698185086250305, "num_tokens": 198908178.0, "step": 5236 }, { "epoch": 0.6662002289785015, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3029439449310303, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8604036569595337, "num_tokens": 198942847.0, "step": 5237 }, { "epoch": 0.6663274392570919, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0222296714782715, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8588171005249023, "num_tokens": 198981444.0, "step": 5238 }, { "epoch": 0.6664546495356825, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3930344581604004, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8684829473495483, "num_tokens": 199019159.0, "step": 5239 }, { "epoch": 0.666581859814273, "ewc_loss": 0.053466796875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.939220428466797, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.852912187576294, "num_tokens": 199056696.0, "step": 5240 }, { "epoch": 0.6667090700928635, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5623302459716797, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8716123104095459, "num_tokens": 199092358.0, "step": 5241 }, { "epoch": 0.666836280371454, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.8794820308685303, "learning_rate": 1e-05, "loss": 0.563, "mean_token_accuracy": 0.831303060054779, "num_tokens": 199135527.0, "step": 5242 }, { "epoch": 0.6669634906500446, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.816599130630493, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8620736002922058, "num_tokens": 199171886.0, "step": 5243 }, { "epoch": 0.667090700928635, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 2.9175941944122314, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8556299209594727, "num_tokens": 199209807.0, "step": 5244 }, { "epoch": 0.6672179112072255, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.113762378692627, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.8568090796470642, "num_tokens": 199244878.0, "step": 5245 }, { "epoch": 0.667345121485816, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2241621017456055, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8558269739151001, "num_tokens": 199279445.0, "step": 5246 }, { "epoch": 0.6674723317644066, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.169995307922363, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8764610290527344, "num_tokens": 199315504.0, "step": 5247 }, { "epoch": 0.6675995420429971, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5422136783599854, "learning_rate": 1e-05, "loss": 0.5363, "mean_token_accuracy": 0.8429700136184692, "num_tokens": 199354345.0, "step": 5248 }, { "epoch": 0.6677267523215876, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9645190238952637, "learning_rate": 1e-05, "loss": 0.5448, "mean_token_accuracy": 0.8424952030181885, "num_tokens": 199401530.0, "step": 5249 }, { "epoch": 0.667853962600178, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.586703062057495, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8551398515701294, "num_tokens": 199440428.0, "step": 5250 }, { "epoch": 0.6679811728787686, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.754056453704834, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8767331838607788, "num_tokens": 199478875.0, "step": 5251 }, { "epoch": 0.6681083831573591, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.491068124771118, "learning_rate": 1e-05, "loss": 0.5583, "mean_token_accuracy": 0.8392201066017151, "num_tokens": 199517287.0, "step": 5252 }, { "epoch": 0.6682355934359496, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.613327741622925, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8557072877883911, "num_tokens": 199553723.0, "step": 5253 }, { "epoch": 0.6683628037145402, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3692824840545654, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8605684638023376, "num_tokens": 199592059.0, "step": 5254 }, { "epoch": 0.6684900139931307, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.512707471847534, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8552556037902832, "num_tokens": 199633005.0, "step": 5255 }, { "epoch": 0.6686172242717211, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2526869773864746, "learning_rate": 1e-05, "loss": 0.5652, "mean_token_accuracy": 0.831877589225769, "num_tokens": 199677507.0, "step": 5256 }, { "epoch": 0.6687444345503116, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4965505599975586, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8610711097717285, "num_tokens": 199716040.0, "step": 5257 }, { "epoch": 0.6688716448289022, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.150353193283081, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8643999695777893, "num_tokens": 199750438.0, "step": 5258 }, { "epoch": 0.6689988551074927, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.540571451187134, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8593587875366211, "num_tokens": 199786889.0, "step": 5259 }, { "epoch": 0.6691260653860832, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.066004753112793, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8627142906188965, "num_tokens": 199821673.0, "step": 5260 }, { "epoch": 0.6692532756646737, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.669877529144287, "learning_rate": 1e-05, "loss": 0.5249, "mean_token_accuracy": 0.8463897705078125, "num_tokens": 199854313.0, "step": 5261 }, { "epoch": 0.6693804859432643, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.043501853942871, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8632853031158447, "num_tokens": 199892853.0, "step": 5262 }, { "epoch": 0.6695076962218547, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8127810955047607, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8581091165542603, "num_tokens": 199931754.0, "step": 5263 }, { "epoch": 0.6696349065004452, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.126011610031128, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.871833860874176, "num_tokens": 199965570.0, "step": 5264 }, { "epoch": 0.6697621167790357, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8795225620269775, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8731149435043335, "num_tokens": 200004931.0, "step": 5265 }, { "epoch": 0.6698893270576263, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.309645414352417, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8700270652770996, "num_tokens": 200040315.0, "step": 5266 }, { "epoch": 0.6700165373362168, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7801272869110107, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8549625277519226, "num_tokens": 200081206.0, "step": 5267 }, { "epoch": 0.6701437476148073, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.405590057373047, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8645330667495728, "num_tokens": 200119975.0, "step": 5268 }, { "epoch": 0.6702709578933977, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.587528944015503, "learning_rate": 1e-05, "loss": 0.5492, "mean_token_accuracy": 0.8384503722190857, "num_tokens": 200154956.0, "step": 5269 }, { "epoch": 0.6703981681719883, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4097766876220703, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8513913154602051, "num_tokens": 200194499.0, "step": 5270 }, { "epoch": 0.6705253784505788, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3819262981414795, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8643755316734314, "num_tokens": 200227142.0, "step": 5271 }, { "epoch": 0.6706525887291693, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3994882106781006, "learning_rate": 1e-05, "loss": 0.5335, "mean_token_accuracy": 0.8459957838058472, "num_tokens": 200262123.0, "step": 5272 }, { "epoch": 0.6707797990077599, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2139101028442383, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8547350168228149, "num_tokens": 200299648.0, "step": 5273 }, { "epoch": 0.6709070092863504, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.400308132171631, "learning_rate": 1e-05, "loss": 0.5313, "mean_token_accuracy": 0.8419699668884277, "num_tokens": 200337191.0, "step": 5274 }, { "epoch": 0.6710342195649408, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.090099334716797, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8663274049758911, "num_tokens": 200371487.0, "step": 5275 }, { "epoch": 0.6711614298435313, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4337058067321777, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8626766800880432, "num_tokens": 200412634.0, "step": 5276 }, { "epoch": 0.6712886401221219, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.0103700160980225, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8642094135284424, "num_tokens": 200449663.0, "step": 5277 }, { "epoch": 0.6714158504007124, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.539205551147461, "learning_rate": 1e-05, "loss": 0.5356, "mean_token_accuracy": 0.8455624580383301, "num_tokens": 200494962.0, "step": 5278 }, { "epoch": 0.6715430606793029, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.9598641395568848, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8730858564376831, "num_tokens": 200531101.0, "step": 5279 }, { "epoch": 0.6716702709578934, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.70353627204895, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8665574789047241, "num_tokens": 200568725.0, "step": 5280 }, { "epoch": 0.6717974812364839, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.9781970977783203, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8547805547714233, "num_tokens": 200607319.0, "step": 5281 }, { "epoch": 0.6719246915150744, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.862778663635254, "learning_rate": 1e-05, "loss": 0.4946, "mean_token_accuracy": 0.8570031523704529, "num_tokens": 200654791.0, "step": 5282 }, { "epoch": 0.6720519017936649, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.11590313911438, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.862910270690918, "num_tokens": 200694175.0, "step": 5283 }, { "epoch": 0.6721791120722554, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9508543014526367, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8599435091018677, "num_tokens": 200726675.0, "step": 5284 }, { "epoch": 0.672306322350846, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3242433071136475, "learning_rate": 1e-05, "loss": 0.5178, "mean_token_accuracy": 0.8491644859313965, "num_tokens": 200759358.0, "step": 5285 }, { "epoch": 0.6724335326294365, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8453528881073, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8612223267555237, "num_tokens": 200797735.0, "step": 5286 }, { "epoch": 0.6725607429080269, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4024698734283447, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8560318350791931, "num_tokens": 200835446.0, "step": 5287 }, { "epoch": 0.6726879531866174, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6663646697998047, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8647251129150391, "num_tokens": 200881020.0, "step": 5288 }, { "epoch": 0.672815163465208, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3885273933410645, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8625053763389587, "num_tokens": 200921549.0, "step": 5289 }, { "epoch": 0.6729423737437985, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5170376300811768, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.8482375741004944, "num_tokens": 200953991.0, "step": 5290 }, { "epoch": 0.673069584022389, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.342697858810425, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8624547123908997, "num_tokens": 200990240.0, "step": 5291 }, { "epoch": 0.6731967943009796, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3915421962738037, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.87409508228302, "num_tokens": 201022433.0, "step": 5292 }, { "epoch": 0.67332400457957, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.283188819885254, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8600945472717285, "num_tokens": 201055801.0, "step": 5293 }, { "epoch": 0.6734512148581605, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.310516357421875, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8510050177574158, "num_tokens": 201102141.0, "step": 5294 }, { "epoch": 0.673578425136751, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2347640991210938, "learning_rate": 1e-05, "loss": 0.5447, "mean_token_accuracy": 0.848915696144104, "num_tokens": 201139834.0, "step": 5295 }, { "epoch": 0.6737056354153416, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.288764715194702, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8694018125534058, "num_tokens": 201169650.0, "step": 5296 }, { "epoch": 0.6738328456939321, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1645376682281494, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8629981279373169, "num_tokens": 201208716.0, "step": 5297 }, { "epoch": 0.6739600559725226, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2780370712280273, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8617486357688904, "num_tokens": 201253976.0, "step": 5298 }, { "epoch": 0.674087266251113, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.126385450363159, "learning_rate": 1e-05, "loss": 0.5026, "mean_token_accuracy": 0.8518041968345642, "num_tokens": 201291478.0, "step": 5299 }, { "epoch": 0.6742144765297036, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3099288940429688, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8569868803024292, "num_tokens": 201328824.0, "step": 5300 }, { "epoch": 0.6743416868082941, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.075859785079956, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8459939956665039, "num_tokens": 201363859.0, "step": 5301 }, { "epoch": 0.6744688970868846, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.368870735168457, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8673619031906128, "num_tokens": 201402635.0, "step": 5302 }, { "epoch": 0.6745961073654752, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0029075145721436, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8519077301025391, "num_tokens": 201444203.0, "step": 5303 }, { "epoch": 0.6747233176440657, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4845526218414307, "learning_rate": 1e-05, "loss": 0.5367, "mean_token_accuracy": 0.8439759016036987, "num_tokens": 201486397.0, "step": 5304 }, { "epoch": 0.6748505279226561, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.935121774673462, "learning_rate": 1e-05, "loss": 0.4924, "mean_token_accuracy": 0.8524245619773865, "num_tokens": 201528262.0, "step": 5305 }, { "epoch": 0.6749777382012466, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6681370735168457, "learning_rate": 1e-05, "loss": 0.501, "mean_token_accuracy": 0.8586120009422302, "num_tokens": 201566274.0, "step": 5306 }, { "epoch": 0.6751049484798372, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.925150156021118, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8538967370986938, "num_tokens": 201610196.0, "step": 5307 }, { "epoch": 0.6752321587584277, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8944413661956787, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.8538475632667542, "num_tokens": 201650144.0, "step": 5308 }, { "epoch": 0.6753593690370182, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.058587074279785, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8591741919517517, "num_tokens": 201689550.0, "step": 5309 }, { "epoch": 0.6754865793156087, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.038919925689697, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8734576106071472, "num_tokens": 201729807.0, "step": 5310 }, { "epoch": 0.6756137895941993, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.3446271419525146, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8637765645980835, "num_tokens": 201759540.0, "step": 5311 }, { "epoch": 0.6757409998727897, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9457850456237793, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.8549071550369263, "num_tokens": 201800067.0, "step": 5312 }, { "epoch": 0.6758682101513802, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.475407361984253, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8615343570709229, "num_tokens": 201832688.0, "step": 5313 }, { "epoch": 0.6759954204299707, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.741182804107666, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8549590110778809, "num_tokens": 201876090.0, "step": 5314 }, { "epoch": 0.6761226307085613, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.446654796600342, "learning_rate": 1e-05, "loss": 0.5349, "mean_token_accuracy": 0.8433116674423218, "num_tokens": 201914339.0, "step": 5315 }, { "epoch": 0.6762498409871518, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.575343608856201, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.869286060333252, "num_tokens": 201945746.0, "step": 5316 }, { "epoch": 0.6763770512657423, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.378293514251709, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8645555973052979, "num_tokens": 201978773.0, "step": 5317 }, { "epoch": 0.6765042615443327, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4331555366516113, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8589937686920166, "num_tokens": 202019838.0, "step": 5318 }, { "epoch": 0.6766314718229233, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.288001775741577, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8610848188400269, "num_tokens": 202060825.0, "step": 5319 }, { "epoch": 0.6767586821015138, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.367474317550659, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8760216236114502, "num_tokens": 202092656.0, "step": 5320 }, { "epoch": 0.6768858923801043, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2187228202819824, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8440490961074829, "num_tokens": 202126982.0, "step": 5321 }, { "epoch": 0.6770131026586949, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3524794578552246, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.862892210483551, "num_tokens": 202157550.0, "step": 5322 }, { "epoch": 0.6771403129372854, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.143651008605957, "learning_rate": 1e-05, "loss": 0.5104, "mean_token_accuracy": 0.8460839986801147, "num_tokens": 202192168.0, "step": 5323 }, { "epoch": 0.6772675232158758, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3637773990631104, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8699951767921448, "num_tokens": 202226887.0, "step": 5324 }, { "epoch": 0.6773947334944663, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0653116703033447, "learning_rate": 1e-05, "loss": 0.5049, "mean_token_accuracy": 0.8462272882461548, "num_tokens": 202269631.0, "step": 5325 }, { "epoch": 0.6775219437730569, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4411656856536865, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8577455878257751, "num_tokens": 202306951.0, "step": 5326 }, { "epoch": 0.6776491540516474, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.994560718536377, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8696158528327942, "num_tokens": 202342922.0, "step": 5327 }, { "epoch": 0.6777763643302379, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.569671392440796, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8580954074859619, "num_tokens": 202382125.0, "step": 5328 }, { "epoch": 0.6779035746088284, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9552953243255615, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.860845685005188, "num_tokens": 202424914.0, "step": 5329 }, { "epoch": 0.6780307848874189, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7620444297790527, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8597628474235535, "num_tokens": 202460828.0, "step": 5330 }, { "epoch": 0.6781579951660094, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0126731395721436, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8644255995750427, "num_tokens": 202492484.0, "step": 5331 }, { "epoch": 0.6782852054445999, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.936122179031372, "learning_rate": 1e-05, "loss": 0.5811, "mean_token_accuracy": 0.8306995630264282, "num_tokens": 202529079.0, "step": 5332 }, { "epoch": 0.6784124157231904, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.210348129272461, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8463890552520752, "num_tokens": 202563024.0, "step": 5333 }, { "epoch": 0.678539626001781, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9241206645965576, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.860522449016571, "num_tokens": 202601264.0, "step": 5334 }, { "epoch": 0.6786668362803715, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.3786754608154297, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8542587757110596, "num_tokens": 202642933.0, "step": 5335 }, { "epoch": 0.6787940465589619, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.768014669418335, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8680797219276428, "num_tokens": 202674851.0, "step": 5336 }, { "epoch": 0.6789212568375524, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.4298083782196045, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8705390095710754, "num_tokens": 202711117.0, "step": 5337 }, { "epoch": 0.679048467116143, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.562021493911743, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8688473105430603, "num_tokens": 202749025.0, "step": 5338 }, { "epoch": 0.6791756773947335, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3950860500335693, "learning_rate": 1e-05, "loss": 0.5372, "mean_token_accuracy": 0.8421785831451416, "num_tokens": 202790733.0, "step": 5339 }, { "epoch": 0.679302887673324, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.402303695678711, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8549371957778931, "num_tokens": 202825753.0, "step": 5340 }, { "epoch": 0.6794300979519146, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3476266860961914, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8741150498390198, "num_tokens": 202862069.0, "step": 5341 }, { "epoch": 0.679557308230505, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2673416137695312, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8620291948318481, "num_tokens": 202904526.0, "step": 5342 }, { "epoch": 0.6796845185090955, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3193249702453613, "learning_rate": 1e-05, "loss": 0.5201, "mean_token_accuracy": 0.8474083542823792, "num_tokens": 202942342.0, "step": 5343 }, { "epoch": 0.679811728787686, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.197394847869873, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8517559766769409, "num_tokens": 202974644.0, "step": 5344 }, { "epoch": 0.6799389390662766, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2973456382751465, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8733264207839966, "num_tokens": 203011728.0, "step": 5345 }, { "epoch": 0.6800661493448671, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.130124807357788, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8553481101989746, "num_tokens": 203051768.0, "step": 5346 }, { "epoch": 0.6801933596234576, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.308074474334717, "learning_rate": 1e-05, "loss": 0.5225, "mean_token_accuracy": 0.8498027920722961, "num_tokens": 203094943.0, "step": 5347 }, { "epoch": 0.680320569902048, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.090766191482544, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8618062734603882, "num_tokens": 203130587.0, "step": 5348 }, { "epoch": 0.6804477801806386, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.340650796890259, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8556166291236877, "num_tokens": 203170030.0, "step": 5349 }, { "epoch": 0.6805749904592291, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.0336813926696777, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8553933501243591, "num_tokens": 203209382.0, "step": 5350 }, { "epoch": 0.6807022007378196, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4227664470672607, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8576145172119141, "num_tokens": 203247750.0, "step": 5351 }, { "epoch": 0.6808294110164101, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9645113945007324, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8634070158004761, "num_tokens": 203284443.0, "step": 5352 }, { "epoch": 0.6809566212950007, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5612525939941406, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8694854974746704, "num_tokens": 203324525.0, "step": 5353 }, { "epoch": 0.6810838315735911, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9031410217285156, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8631676435470581, "num_tokens": 203369633.0, "step": 5354 }, { "epoch": 0.6812110418521816, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7832188606262207, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8579468727111816, "num_tokens": 203409549.0, "step": 5355 }, { "epoch": 0.6813382521307721, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.947922468185425, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8511680960655212, "num_tokens": 203446856.0, "step": 5356 }, { "epoch": 0.6814654624093627, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.029151916503906, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.882535457611084, "num_tokens": 203478575.0, "step": 5357 }, { "epoch": 0.6815926726879532, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.1743624210357666, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8722712993621826, "num_tokens": 203515617.0, "step": 5358 }, { "epoch": 0.6817198829665437, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.114100456237793, "learning_rate": 1e-05, "loss": 0.5202, "mean_token_accuracy": 0.853090763092041, "num_tokens": 203548636.0, "step": 5359 }, { "epoch": 0.6818470932451343, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.475186586380005, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.867383599281311, "num_tokens": 203581358.0, "step": 5360 }, { "epoch": 0.6819743035237247, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9292826652526855, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8768114447593689, "num_tokens": 203615052.0, "step": 5361 }, { "epoch": 0.6821015138023152, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.535569667816162, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.8550287485122681, "num_tokens": 203655743.0, "step": 5362 }, { "epoch": 0.6822287240809057, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7291109561920166, "learning_rate": 1e-05, "loss": 0.5229, "mean_token_accuracy": 0.8477615714073181, "num_tokens": 203689370.0, "step": 5363 }, { "epoch": 0.6823559343594963, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.471306324005127, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8557271957397461, "num_tokens": 203723594.0, "step": 5364 }, { "epoch": 0.6824831446380868, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.548607587814331, "learning_rate": 1e-05, "loss": 0.5293, "mean_token_accuracy": 0.844383955001831, "num_tokens": 203761746.0, "step": 5365 }, { "epoch": 0.6826103549166773, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.370063304901123, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8597085475921631, "num_tokens": 203801904.0, "step": 5366 }, { "epoch": 0.6827375651952677, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.442944288253784, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8617209196090698, "num_tokens": 203837775.0, "step": 5367 }, { "epoch": 0.6828647754738583, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.268434524536133, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8776216506958008, "num_tokens": 203876456.0, "step": 5368 }, { "epoch": 0.6829919857524488, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3908045291900635, "learning_rate": 1e-05, "loss": 0.4997, "mean_token_accuracy": 0.8562486171722412, "num_tokens": 203917946.0, "step": 5369 }, { "epoch": 0.6831191960310393, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1748275756835938, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8635191321372986, "num_tokens": 203959544.0, "step": 5370 }, { "epoch": 0.6832464063096299, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.399784564971924, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8506138920783997, "num_tokens": 204000882.0, "step": 5371 }, { "epoch": 0.6833736165882204, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.103956937789917, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8650287389755249, "num_tokens": 204037715.0, "step": 5372 }, { "epoch": 0.6835008268668108, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.468473434448242, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8558741807937622, "num_tokens": 204073145.0, "step": 5373 }, { "epoch": 0.6836280371454013, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.020772695541382, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8544800281524658, "num_tokens": 204112781.0, "step": 5374 }, { "epoch": 0.6837552474239919, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5840904712677, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8583085536956787, "num_tokens": 204151128.0, "step": 5375 }, { "epoch": 0.6838824577025824, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.989535093307495, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8665762543678284, "num_tokens": 204188666.0, "step": 5376 }, { "epoch": 0.6840096679811729, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.748649835586548, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8798916339874268, "num_tokens": 204225410.0, "step": 5377 }, { "epoch": 0.6841368782597634, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.037836790084839, "learning_rate": 1e-05, "loss": 0.543, "mean_token_accuracy": 0.8423948287963867, "num_tokens": 204271506.0, "step": 5378 }, { "epoch": 0.6842640885383539, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8857924938201904, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.8572947978973389, "num_tokens": 204310569.0, "step": 5379 }, { "epoch": 0.6843912988169444, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.190967559814453, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8627082109451294, "num_tokens": 204349614.0, "step": 5380 }, { "epoch": 0.6845185090955349, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9008240699768066, "learning_rate": 1e-05, "loss": 0.5318, "mean_token_accuracy": 0.8467550277709961, "num_tokens": 204381985.0, "step": 5381 }, { "epoch": 0.6846457193741254, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3719894886016846, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8537601232528687, "num_tokens": 204418433.0, "step": 5382 }, { "epoch": 0.684772929652716, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7288153171539307, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8589234352111816, "num_tokens": 204457544.0, "step": 5383 }, { "epoch": 0.6849001399313065, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4157192707061768, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8617867231369019, "num_tokens": 204500626.0, "step": 5384 }, { "epoch": 0.6850273502098969, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.518683671951294, "learning_rate": 1e-05, "loss": 0.5146, "mean_token_accuracy": 0.8473504185676575, "num_tokens": 204540861.0, "step": 5385 }, { "epoch": 0.6851545604884874, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.3977344036102295, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8717771172523499, "num_tokens": 204581555.0, "step": 5386 }, { "epoch": 0.685281770767078, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.337845802307129, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8627785444259644, "num_tokens": 204614931.0, "step": 5387 }, { "epoch": 0.6854089810456685, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3854563236236572, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8581767082214355, "num_tokens": 204649678.0, "step": 5388 }, { "epoch": 0.685536191324259, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2019028663635254, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8533717393875122, "num_tokens": 204690617.0, "step": 5389 }, { "epoch": 0.6856634016028496, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.384803295135498, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8533127903938293, "num_tokens": 204725685.0, "step": 5390 }, { "epoch": 0.68579061188144, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.102937698364258, "learning_rate": 1e-05, "loss": 0.539, "mean_token_accuracy": 0.8393908143043518, "num_tokens": 204768265.0, "step": 5391 }, { "epoch": 0.6859178221600305, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.414907932281494, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8625295162200928, "num_tokens": 204806211.0, "step": 5392 }, { "epoch": 0.686045032438621, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.034087896347046, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8601609468460083, "num_tokens": 204844982.0, "step": 5393 }, { "epoch": 0.6861722427172116, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000133514404296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.490952491760254, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8722734451293945, "num_tokens": 204881676.0, "step": 5394 }, { "epoch": 0.6862994529958021, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9793577194213867, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8575522303581238, "num_tokens": 204920316.0, "step": 5395 }, { "epoch": 0.6864266632743926, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6252152919769287, "learning_rate": 1e-05, "loss": 0.5115, "mean_token_accuracy": 0.8480836153030396, "num_tokens": 204960269.0, "step": 5396 }, { "epoch": 0.686553873552983, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.9848203659057617, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8594276905059814, "num_tokens": 204993238.0, "step": 5397 }, { "epoch": 0.6866810838315736, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7762868404388428, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.8497741222381592, "num_tokens": 205029550.0, "step": 5398 }, { "epoch": 0.6868082941101641, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0288257598876953, "learning_rate": 1e-05, "loss": 0.4973, "mean_token_accuracy": 0.8522659540176392, "num_tokens": 205073198.0, "step": 5399 }, { "epoch": 0.6869355043887546, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9197137355804443, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8637956380844116, "num_tokens": 205107581.0, "step": 5400 }, { "epoch": 0.6870627146673451, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2036056518554688, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8576286435127258, "num_tokens": 205144030.0, "step": 5401 }, { "epoch": 0.6871899249459357, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.953317165374756, "learning_rate": 1e-05, "loss": 0.5068, "mean_token_accuracy": 0.8555353879928589, "num_tokens": 205180583.0, "step": 5402 }, { "epoch": 0.6873171352245261, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.385286331176758, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8605002164840698, "num_tokens": 205215889.0, "step": 5403 }, { "epoch": 0.6874443455031166, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.777801513671875, "learning_rate": 1e-05, "loss": 0.5374, "mean_token_accuracy": 0.8408302068710327, "num_tokens": 205254725.0, "step": 5404 }, { "epoch": 0.6875715557817071, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.4148221015930176, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8679362535476685, "num_tokens": 205289533.0, "step": 5405 }, { "epoch": 0.6876987660602977, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6059606075286865, "learning_rate": 1e-05, "loss": 0.5659, "mean_token_accuracy": 0.837777853012085, "num_tokens": 205330462.0, "step": 5406 }, { "epoch": 0.6878259763388882, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3775267601013184, "learning_rate": 1e-05, "loss": 0.5352, "mean_token_accuracy": 0.8452764749526978, "num_tokens": 205371462.0, "step": 5407 }, { "epoch": 0.6879531866174787, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4488799571990967, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8790926933288574, "num_tokens": 205404629.0, "step": 5408 }, { "epoch": 0.6880803968960693, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3277668952941895, "learning_rate": 1e-05, "loss": 0.5102, "mean_token_accuracy": 0.8468499183654785, "num_tokens": 205440600.0, "step": 5409 }, { "epoch": 0.6882076071746597, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.340435743331909, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8617744445800781, "num_tokens": 205473696.0, "step": 5410 }, { "epoch": 0.6883348174532502, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2668378353118896, "learning_rate": 1e-05, "loss": 0.5201, "mean_token_accuracy": 0.848298192024231, "num_tokens": 205519889.0, "step": 5411 }, { "epoch": 0.6884620277318407, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2832844257354736, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8738986253738403, "num_tokens": 205552303.0, "step": 5412 }, { "epoch": 0.6885892380104313, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.222423553466797, "learning_rate": 1e-05, "loss": 0.5139, "mean_token_accuracy": 0.844359278678894, "num_tokens": 205588918.0, "step": 5413 }, { "epoch": 0.6887164482890218, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.267416000366211, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8589490056037903, "num_tokens": 205624634.0, "step": 5414 }, { "epoch": 0.6888436585676123, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1579360961914062, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8640552759170532, "num_tokens": 205668700.0, "step": 5415 }, { "epoch": 0.6889708688462027, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.263164520263672, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.853748619556427, "num_tokens": 205711051.0, "step": 5416 }, { "epoch": 0.6890980791247933, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1136159896850586, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8632743954658508, "num_tokens": 205750349.0, "step": 5417 }, { "epoch": 0.6892252894033838, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.298560619354248, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8715257048606873, "num_tokens": 205789525.0, "step": 5418 }, { "epoch": 0.6893524996819743, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0509724617004395, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8584186434745789, "num_tokens": 205827266.0, "step": 5419 }, { "epoch": 0.6894797099605648, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3842878341674805, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8658607006072998, "num_tokens": 205866170.0, "step": 5420 }, { "epoch": 0.6896069202391554, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.9863367080688477, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8670569658279419, "num_tokens": 205905613.0, "step": 5421 }, { "epoch": 0.6897341305177458, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5172665119171143, "learning_rate": 1e-05, "loss": 0.5432, "mean_token_accuracy": 0.8379237651824951, "num_tokens": 205943142.0, "step": 5422 }, { "epoch": 0.6898613407963363, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.9159634113311768, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8583282232284546, "num_tokens": 205983321.0, "step": 5423 }, { "epoch": 0.6899885510749268, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7313950061798096, "learning_rate": 1e-05, "loss": 0.5275, "mean_token_accuracy": 0.8472774028778076, "num_tokens": 206019013.0, "step": 5424 }, { "epoch": 0.6901157613535174, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.9146013259887695, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8585435152053833, "num_tokens": 206059958.0, "step": 5425 }, { "epoch": 0.6902429716321079, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.99340558052063, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8620519638061523, "num_tokens": 206096536.0, "step": 5426 }, { "epoch": 0.6903701819106984, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1193525791168213, "learning_rate": 1e-05, "loss": 0.5071, "mean_token_accuracy": 0.8536313772201538, "num_tokens": 206138399.0, "step": 5427 }, { "epoch": 0.6904973921892888, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.1146240234375, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8733959197998047, "num_tokens": 206171666.0, "step": 5428 }, { "epoch": 0.6906246024678794, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.431764602661133, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8655523657798767, "num_tokens": 206212439.0, "step": 5429 }, { "epoch": 0.6907518127464699, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.98223614692688, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.858974814414978, "num_tokens": 206245213.0, "step": 5430 }, { "epoch": 0.6908790230250604, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5473146438598633, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8604715466499329, "num_tokens": 206283875.0, "step": 5431 }, { "epoch": 0.691006233303651, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7602827548980713, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8536470532417297, "num_tokens": 206319442.0, "step": 5432 }, { "epoch": 0.6911334435822415, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5167813301086426, "learning_rate": 1e-05, "loss": 0.5326, "mean_token_accuracy": 0.8431313037872314, "num_tokens": 206358210.0, "step": 5433 }, { "epoch": 0.6912606538608319, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5650877952575684, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8701112866401672, "num_tokens": 206390900.0, "step": 5434 }, { "epoch": 0.6913878641394224, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4057323932647705, "learning_rate": 1e-05, "loss": 0.5422, "mean_token_accuracy": 0.8409827947616577, "num_tokens": 206426296.0, "step": 5435 }, { "epoch": 0.691515074418013, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.445864200592041, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8588995933532715, "num_tokens": 206457010.0, "step": 5436 }, { "epoch": 0.6916422846966035, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.294518232345581, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8658332228660583, "num_tokens": 206496324.0, "step": 5437 }, { "epoch": 0.691769494975194, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.390085458755493, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.8535280823707581, "num_tokens": 206530507.0, "step": 5438 }, { "epoch": 0.6918967052537845, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.206388473510742, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8644205331802368, "num_tokens": 206565123.0, "step": 5439 }, { "epoch": 0.692023915532375, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3729653358459473, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8622180223464966, "num_tokens": 206602026.0, "step": 5440 }, { "epoch": 0.6921511258109655, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.117224931716919, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8548358678817749, "num_tokens": 206643371.0, "step": 5441 }, { "epoch": 0.692278336089556, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.426886796951294, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8712204694747925, "num_tokens": 206675822.0, "step": 5442 }, { "epoch": 0.6924055463681466, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0422158241271973, "learning_rate": 1e-05, "loss": 0.3608, "mean_token_accuracy": 0.8945213556289673, "num_tokens": 206703973.0, "step": 5443 }, { "epoch": 0.6925327566467371, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.52644681930542, "learning_rate": 1e-05, "loss": 0.5093, "mean_token_accuracy": 0.853916347026825, "num_tokens": 206740645.0, "step": 5444 }, { "epoch": 0.6926599669253276, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.966952323913574, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8677061796188354, "num_tokens": 206785620.0, "step": 5445 }, { "epoch": 0.692787177203918, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.708817481994629, "learning_rate": 1e-05, "loss": 0.5117, "mean_token_accuracy": 0.8528735637664795, "num_tokens": 206818761.0, "step": 5446 }, { "epoch": 0.6929143874825086, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.984924793243408, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8686943650245667, "num_tokens": 206857523.0, "step": 5447 }, { "epoch": 0.6930415977610991, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9020016193389893, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.852366030216217, "num_tokens": 206893389.0, "step": 5448 }, { "epoch": 0.6931688080396896, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.1545445919036865, "learning_rate": 1e-05, "loss": 0.5051, "mean_token_accuracy": 0.8539217710494995, "num_tokens": 206924997.0, "step": 5449 }, { "epoch": 0.6932960183182801, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.965481758117676, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8552788496017456, "num_tokens": 206957032.0, "step": 5450 }, { "epoch": 0.6934232285968707, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.355360269546509, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8625770807266235, "num_tokens": 206995407.0, "step": 5451 }, { "epoch": 0.6935504388754611, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8358311653137207, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.864611804485321, "num_tokens": 207031563.0, "step": 5452 }, { "epoch": 0.6936776491540516, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.4587419033050537, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8634169101715088, "num_tokens": 207064641.0, "step": 5453 }, { "epoch": 0.6938048594326421, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.619840145111084, "learning_rate": 1e-05, "loss": 0.5531, "mean_token_accuracy": 0.8420258164405823, "num_tokens": 207099474.0, "step": 5454 }, { "epoch": 0.6939320697112327, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.421811103820801, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8693971633911133, "num_tokens": 207139250.0, "step": 5455 }, { "epoch": 0.6940592799898232, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4384708404541016, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8600876331329346, "num_tokens": 207179523.0, "step": 5456 }, { "epoch": 0.6941864902684137, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3775904178619385, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8658189177513123, "num_tokens": 207218143.0, "step": 5457 }, { "epoch": 0.6943137005470043, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2886123657226562, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8608185052871704, "num_tokens": 207256946.0, "step": 5458 }, { "epoch": 0.6944409108255947, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00013446807861328125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3358712196350098, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.873078465461731, "num_tokens": 207296793.0, "step": 5459 }, { "epoch": 0.6945681211041852, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2116963863372803, "learning_rate": 1e-05, "loss": 0.5051, "mean_token_accuracy": 0.8478083610534668, "num_tokens": 207333602.0, "step": 5460 }, { "epoch": 0.6946953313827757, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.313352584838867, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8489409685134888, "num_tokens": 207375806.0, "step": 5461 }, { "epoch": 0.6948225416613663, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1489973068237305, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8801034688949585, "num_tokens": 207407856.0, "step": 5462 }, { "epoch": 0.6949497519399568, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.322701930999756, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.8451595306396484, "num_tokens": 207447000.0, "step": 5463 }, { "epoch": 0.6950769622185473, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0905039310455322, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8694595098495483, "num_tokens": 207479860.0, "step": 5464 }, { "epoch": 0.6952041724971377, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3769290447235107, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8732836842536926, "num_tokens": 207515454.0, "step": 5465 }, { "epoch": 0.6953313827757283, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.011179208755493, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.8549149632453918, "num_tokens": 207559259.0, "step": 5466 }, { "epoch": 0.6954585930543188, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.485318422317505, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8502306938171387, "num_tokens": 207595909.0, "step": 5467 }, { "epoch": 0.6955858033329093, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.9466207027435303, "learning_rate": 1e-05, "loss": 0.568, "mean_token_accuracy": 0.8331905007362366, "num_tokens": 207636063.0, "step": 5468 }, { "epoch": 0.6957130136114998, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6447689533233643, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8625235557556152, "num_tokens": 207675990.0, "step": 5469 }, { "epoch": 0.6958402238900904, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.932988166809082, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8626738786697388, "num_tokens": 207708996.0, "step": 5470 }, { "epoch": 0.6959674341686808, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8564019203186035, "learning_rate": 1e-05, "loss": 0.5513, "mean_token_accuracy": 0.8425251841545105, "num_tokens": 207750433.0, "step": 5471 }, { "epoch": 0.6960946444472713, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0160889625549316, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8521872758865356, "num_tokens": 207792479.0, "step": 5472 }, { "epoch": 0.6962218547258618, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.038476943969727, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8604698181152344, "num_tokens": 207833420.0, "step": 5473 }, { "epoch": 0.6963490650044524, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2605361938476562, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8715182542800903, "num_tokens": 207875933.0, "step": 5474 }, { "epoch": 0.6964762752830429, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.035909175872803, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8575543165206909, "num_tokens": 207915481.0, "step": 5475 }, { "epoch": 0.6966034855616334, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.4847207069396973, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.857068657875061, "num_tokens": 207952024.0, "step": 5476 }, { "epoch": 0.6967306958402238, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8442113399505615, "learning_rate": 1e-05, "loss": 0.5296, "mean_token_accuracy": 0.8498271703720093, "num_tokens": 207990938.0, "step": 5477 }, { "epoch": 0.6968579061188144, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4974489212036133, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8583496809005737, "num_tokens": 208032492.0, "step": 5478 }, { "epoch": 0.6969851163974049, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6448941230773926, "learning_rate": 1e-05, "loss": 0.5173, "mean_token_accuracy": 0.849265992641449, "num_tokens": 208068817.0, "step": 5479 }, { "epoch": 0.6971123266759954, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.45583438873291, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8639912009239197, "num_tokens": 208105090.0, "step": 5480 }, { "epoch": 0.697239536954586, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4841394424438477, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8579357862472534, "num_tokens": 208136653.0, "step": 5481 }, { "epoch": 0.6973667472331765, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.356856346130371, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8603417873382568, "num_tokens": 208175556.0, "step": 5482 }, { "epoch": 0.6974939575117669, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3612232208251953, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8674613833427429, "num_tokens": 208214549.0, "step": 5483 }, { "epoch": 0.6976211677903574, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2998294830322266, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8755009174346924, "num_tokens": 208242582.0, "step": 5484 }, { "epoch": 0.697748378068948, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.308716058731079, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8539508581161499, "num_tokens": 208281639.0, "step": 5485 }, { "epoch": 0.6978755883475385, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.219553232192993, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8751891851425171, "num_tokens": 208323543.0, "step": 5486 }, { "epoch": 0.698002798626129, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2849977016448975, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8588358163833618, "num_tokens": 208359881.0, "step": 5487 }, { "epoch": 0.6981300089047195, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1549906730651855, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8497503995895386, "num_tokens": 208404541.0, "step": 5488 }, { "epoch": 0.69825721918331, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3079724311828613, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8645223379135132, "num_tokens": 208437731.0, "step": 5489 }, { "epoch": 0.6983844294619005, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.100426197052002, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8586519360542297, "num_tokens": 208475171.0, "step": 5490 }, { "epoch": 0.698511639740491, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3594422340393066, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8594280481338501, "num_tokens": 208511410.0, "step": 5491 }, { "epoch": 0.6986388500190815, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.018632173538208, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8635513782501221, "num_tokens": 208551133.0, "step": 5492 }, { "epoch": 0.6987660602976721, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.467719316482544, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8636338710784912, "num_tokens": 208593394.0, "step": 5493 }, { "epoch": 0.6988932705762626, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.954502582550049, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8498027324676514, "num_tokens": 208627160.0, "step": 5494 }, { "epoch": 0.699020480854853, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6431145668029785, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8576894998550415, "num_tokens": 208664896.0, "step": 5495 }, { "epoch": 0.6991476911334435, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.916539430618286, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8534998893737793, "num_tokens": 208702118.0, "step": 5496 }, { "epoch": 0.6992749014120341, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8983476161956787, "learning_rate": 1e-05, "loss": 0.5213, "mean_token_accuracy": 0.8504647016525269, "num_tokens": 208741243.0, "step": 5497 }, { "epoch": 0.6994021116906246, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.0298500061035156, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8680598735809326, "num_tokens": 208778914.0, "step": 5498 }, { "epoch": 0.6995293219692151, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.089770793914795, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8574622869491577, "num_tokens": 208815683.0, "step": 5499 }, { "epoch": 0.6996565322478057, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.330277681350708, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8603166341781616, "num_tokens": 208853812.0, "step": 5500 }, { "epoch": 0.6997837425263961, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.022918701171875, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8697540163993835, "num_tokens": 208894326.0, "step": 5501 }, { "epoch": 0.6999109528049866, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5156409740448, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8669331073760986, "num_tokens": 208929644.0, "step": 5502 }, { "epoch": 0.7000381630835771, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.807615041732788, "learning_rate": 1e-05, "loss": 0.5171, "mean_token_accuracy": 0.8565638661384583, "num_tokens": 208972836.0, "step": 5503 }, { "epoch": 0.7001653733621677, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5083277225494385, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8595818281173706, "num_tokens": 209016255.0, "step": 5504 }, { "epoch": 0.7002925836407582, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6202239990234375, "learning_rate": 1e-05, "loss": 0.531, "mean_token_accuracy": 0.8471812605857849, "num_tokens": 209054576.0, "step": 5505 }, { "epoch": 0.7004197939193487, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4138457775115967, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8688762187957764, "num_tokens": 209097881.0, "step": 5506 }, { "epoch": 0.7005470041979391, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4707462787628174, "learning_rate": 1e-05, "loss": 0.5188, "mean_token_accuracy": 0.850595235824585, "num_tokens": 209137354.0, "step": 5507 }, { "epoch": 0.7006742144765297, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.349069118499756, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8554772734642029, "num_tokens": 209170505.0, "step": 5508 }, { "epoch": 0.7008014247551202, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.376816511154175, "learning_rate": 1e-05, "loss": 0.4949, "mean_token_accuracy": 0.8524187207221985, "num_tokens": 209207745.0, "step": 5509 }, { "epoch": 0.7009286350337107, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2550344467163086, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8538320660591125, "num_tokens": 209246674.0, "step": 5510 }, { "epoch": 0.7010558453123013, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3244142532348633, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.862183153629303, "num_tokens": 209289019.0, "step": 5511 }, { "epoch": 0.7011830555908918, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1971096992492676, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8594937324523926, "num_tokens": 209323053.0, "step": 5512 }, { "epoch": 0.7013102658694823, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.330996036529541, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8595379590988159, "num_tokens": 209358810.0, "step": 5513 }, { "epoch": 0.7014374761480727, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1188058853149414, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.8619834184646606, "num_tokens": 209401868.0, "step": 5514 }, { "epoch": 0.7015646864266633, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3797409534454346, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8478200435638428, "num_tokens": 209435212.0, "step": 5515 }, { "epoch": 0.7016918967052538, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0455472469329834, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8673338294029236, "num_tokens": 209472357.0, "step": 5516 }, { "epoch": 0.7018191069838443, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.483783006668091, "learning_rate": 1e-05, "loss": 0.5251, "mean_token_accuracy": 0.8447563648223877, "num_tokens": 209512687.0, "step": 5517 }, { "epoch": 0.7019463172624348, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.980787515640259, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8645100593566895, "num_tokens": 209547223.0, "step": 5518 }, { "epoch": 0.7020735275410254, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6443464756011963, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8618916273117065, "num_tokens": 209583299.0, "step": 5519 }, { "epoch": 0.7022007378196158, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.9518210887908936, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8600177764892578, "num_tokens": 209618792.0, "step": 5520 }, { "epoch": 0.7023279480982063, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8548808097839355, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8550658822059631, "num_tokens": 209654324.0, "step": 5521 }, { "epoch": 0.7024551583767968, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0524237155914307, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8727713227272034, "num_tokens": 209684963.0, "step": 5522 }, { "epoch": 0.7025823686553874, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.000100612640381, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8741326928138733, "num_tokens": 209724197.0, "step": 5523 }, { "epoch": 0.7027095789339779, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2776825428009033, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8671660423278809, "num_tokens": 209759397.0, "step": 5524 }, { "epoch": 0.7028367892125684, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9613382816314697, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8619904518127441, "num_tokens": 209802504.0, "step": 5525 }, { "epoch": 0.7029639994911588, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.4586873054504395, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8628989458084106, "num_tokens": 209838798.0, "step": 5526 }, { "epoch": 0.7030912097697494, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7581589221954346, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8499317169189453, "num_tokens": 209880935.0, "step": 5527 }, { "epoch": 0.7032184200483399, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.4874303340911865, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.854202389717102, "num_tokens": 209921514.0, "step": 5528 }, { "epoch": 0.7033456303269304, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5403757095336914, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8676501512527466, "num_tokens": 209960082.0, "step": 5529 }, { "epoch": 0.703472840605521, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.412017822265625, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8609122037887573, "num_tokens": 209996338.0, "step": 5530 }, { "epoch": 0.7036000508841115, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.381629705429077, "learning_rate": 1e-05, "loss": 0.5148, "mean_token_accuracy": 0.8494532108306885, "num_tokens": 210039140.0, "step": 5531 }, { "epoch": 0.7037272611627019, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3545520305633545, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8634642362594604, "num_tokens": 210074166.0, "step": 5532 }, { "epoch": 0.7038544714412924, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2789461612701416, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8729917407035828, "num_tokens": 210110295.0, "step": 5533 }, { "epoch": 0.703981681719883, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3066766262054443, "learning_rate": 1e-05, "loss": 0.5017, "mean_token_accuracy": 0.8513615131378174, "num_tokens": 210151994.0, "step": 5534 }, { "epoch": 0.7041088919984735, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.210689067840576, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8634405136108398, "num_tokens": 210184769.0, "step": 5535 }, { "epoch": 0.704236102277064, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2848424911499023, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8549174070358276, "num_tokens": 210225561.0, "step": 5536 }, { "epoch": 0.7043633125556545, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.153559446334839, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8731967210769653, "num_tokens": 210265544.0, "step": 5537 }, { "epoch": 0.704490522834245, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.282393455505371, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8719958662986755, "num_tokens": 210298178.0, "step": 5538 }, { "epoch": 0.7046177331128355, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.1148264408111572, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8521820306777954, "num_tokens": 210342391.0, "step": 5539 }, { "epoch": 0.704744943391426, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2984468936920166, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8655015230178833, "num_tokens": 210386673.0, "step": 5540 }, { "epoch": 0.7048721536700165, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0776479244232178, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8515917062759399, "num_tokens": 210424050.0, "step": 5541 }, { "epoch": 0.7049993639486071, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3502774238586426, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8686062097549438, "num_tokens": 210462593.0, "step": 5542 }, { "epoch": 0.7051265742271976, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.021350860595703, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8609601259231567, "num_tokens": 210496534.0, "step": 5543 }, { "epoch": 0.705253784505788, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4584250450134277, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8524808883666992, "num_tokens": 210534945.0, "step": 5544 }, { "epoch": 0.7053809947843785, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 3.9577484130859375e-05, "grad_norm": 2.9584081172943115, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8703965544700623, "num_tokens": 210576742.0, "step": 5545 }, { "epoch": 0.7055082050629691, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.618643045425415, "learning_rate": 1e-05, "loss": 0.532, "mean_token_accuracy": 0.8421580791473389, "num_tokens": 210616248.0, "step": 5546 }, { "epoch": 0.7056354153415596, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.8957674503326416, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.853926420211792, "num_tokens": 210658229.0, "step": 5547 }, { "epoch": 0.7057626256201501, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8596322536468506, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8716906309127808, "num_tokens": 210696312.0, "step": 5548 }, { "epoch": 0.7058898358987407, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.973095178604126, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8541814088821411, "num_tokens": 210736916.0, "step": 5549 }, { "epoch": 0.7060170461773311, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.11171293258667, "learning_rate": 1e-05, "loss": 0.5652, "mean_token_accuracy": 0.8411217331886292, "num_tokens": 210777779.0, "step": 5550 }, { "epoch": 0.7061442564559216, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.2686991691589355, "learning_rate": 1e-05, "loss": 0.5494, "mean_token_accuracy": 0.8403069376945496, "num_tokens": 210823134.0, "step": 5551 }, { "epoch": 0.7062714667345121, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.129653453826904, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8562449812889099, "num_tokens": 210862792.0, "step": 5552 }, { "epoch": 0.7063986770131027, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.534619092941284, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8651646375656128, "num_tokens": 210905411.0, "step": 5553 }, { "epoch": 0.7065258872916932, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.930270195007324, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8503810167312622, "num_tokens": 210944090.0, "step": 5554 }, { "epoch": 0.7066530975702837, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5729455947875977, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8569637537002563, "num_tokens": 210977372.0, "step": 5555 }, { "epoch": 0.7067803078488741, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7096312046051025, "learning_rate": 1e-05, "loss": 0.5323, "mean_token_accuracy": 0.8507308959960938, "num_tokens": 211021800.0, "step": 5556 }, { "epoch": 0.7069075181274647, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.471004009246826, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8551592826843262, "num_tokens": 211063568.0, "step": 5557 }, { "epoch": 0.7070347284060552, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.552462100982666, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.861024796962738, "num_tokens": 211100422.0, "step": 5558 }, { "epoch": 0.7071619386846457, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.368759870529175, "learning_rate": 1e-05, "loss": 0.5085, "mean_token_accuracy": 0.8519035577774048, "num_tokens": 211138644.0, "step": 5559 }, { "epoch": 0.7072891489632362, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4645237922668457, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8568753600120544, "num_tokens": 211173033.0, "step": 5560 }, { "epoch": 0.7074163592418268, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2642173767089844, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8665388822555542, "num_tokens": 211212902.0, "step": 5561 }, { "epoch": 0.7075435695204173, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4169347286224365, "learning_rate": 1e-05, "loss": 0.5435, "mean_token_accuracy": 0.838029146194458, "num_tokens": 211258812.0, "step": 5562 }, { "epoch": 0.7076707797990077, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.176891803741455, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8540163636207581, "num_tokens": 211296496.0, "step": 5563 }, { "epoch": 0.7077979900775982, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4350273609161377, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8556679487228394, "num_tokens": 211337081.0, "step": 5564 }, { "epoch": 0.7079252003561888, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.1015872955322266, "learning_rate": 1e-05, "loss": 0.5115, "mean_token_accuracy": 0.849431037902832, "num_tokens": 211373845.0, "step": 5565 }, { "epoch": 0.7080524106347793, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.502082109451294, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8572090864181519, "num_tokens": 211412994.0, "step": 5566 }, { "epoch": 0.7081796209133698, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0209245681762695, "learning_rate": 1e-05, "loss": 0.5082, "mean_token_accuracy": 0.8496907353401184, "num_tokens": 211456678.0, "step": 5567 }, { "epoch": 0.7083068311919604, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6584248542785645, "learning_rate": 1e-05, "loss": 0.5093, "mean_token_accuracy": 0.851963996887207, "num_tokens": 211497040.0, "step": 5568 }, { "epoch": 0.7084340414705508, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.004762887954712, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8493210673332214, "num_tokens": 211534113.0, "step": 5569 }, { "epoch": 0.7085612517491413, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.819744348526001, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8646582365036011, "num_tokens": 211569518.0, "step": 5570 }, { "epoch": 0.7086884620277318, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.090334415435791, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8704861402511597, "num_tokens": 211607918.0, "step": 5571 }, { "epoch": 0.7088156723063224, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9430031776428223, "learning_rate": 1e-05, "loss": 0.5378, "mean_token_accuracy": 0.8474497199058533, "num_tokens": 211647222.0, "step": 5572 }, { "epoch": 0.7089428825849129, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2907826900482178, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8675441741943359, "num_tokens": 211684584.0, "step": 5573 }, { "epoch": 0.7090700928635034, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8761448860168457, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8621552586555481, "num_tokens": 211718968.0, "step": 5574 }, { "epoch": 0.7091973031420938, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4320991039276123, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8610473871231079, "num_tokens": 211756230.0, "step": 5575 }, { "epoch": 0.7093245134206844, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6608569622039795, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8771973848342896, "num_tokens": 211794155.0, "step": 5576 }, { "epoch": 0.7094517236992749, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.433915376663208, "learning_rate": 1e-05, "loss": 0.4924, "mean_token_accuracy": 0.858482301235199, "num_tokens": 211834483.0, "step": 5577 }, { "epoch": 0.7095789339778654, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4649147987365723, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8488134145736694, "num_tokens": 211873369.0, "step": 5578 }, { "epoch": 0.709706144256456, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.415785551071167, "learning_rate": 1e-05, "loss": 0.5286, "mean_token_accuracy": 0.8491159677505493, "num_tokens": 211911985.0, "step": 5579 }, { "epoch": 0.7098333545350465, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.281047821044922, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8620465397834778, "num_tokens": 211953739.0, "step": 5580 }, { "epoch": 0.7099605648136369, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.393183469772339, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8683521151542664, "num_tokens": 211988817.0, "step": 5581 }, { "epoch": 0.7100877750922274, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1636645793914795, "learning_rate": 1e-05, "loss": 0.5206, "mean_token_accuracy": 0.8484116792678833, "num_tokens": 212030891.0, "step": 5582 }, { "epoch": 0.710214985370818, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4118871688842773, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8709678649902344, "num_tokens": 212065727.0, "step": 5583 }, { "epoch": 0.7103421956494085, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.084311008453369, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8569833040237427, "num_tokens": 212098718.0, "step": 5584 }, { "epoch": 0.710469405927999, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.456160545349121, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8568845987319946, "num_tokens": 212142597.0, "step": 5585 }, { "epoch": 0.7105966162065895, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0149106979370117, "learning_rate": 1e-05, "loss": 0.537, "mean_token_accuracy": 0.8357409238815308, "num_tokens": 212182878.0, "step": 5586 }, { "epoch": 0.71072382648518, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.571411609649658, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8664488792419434, "num_tokens": 212216759.0, "step": 5587 }, { "epoch": 0.7108510367637705, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.978775978088379, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8517385721206665, "num_tokens": 212253559.0, "step": 5588 }, { "epoch": 0.710978247042361, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.712289810180664, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8619257211685181, "num_tokens": 212296821.0, "step": 5589 }, { "epoch": 0.7111054573209515, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.9930081367492676, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.865105390548706, "num_tokens": 212334872.0, "step": 5590 }, { "epoch": 0.7112326675995421, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8819515705108643, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8622502684593201, "num_tokens": 212375692.0, "step": 5591 }, { "epoch": 0.7113598778781326, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.150402307510376, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8507527112960815, "num_tokens": 212407301.0, "step": 5592 }, { "epoch": 0.711487088156723, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9268252849578857, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8606477975845337, "num_tokens": 212443831.0, "step": 5593 }, { "epoch": 0.7116142984353135, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.307110071182251, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8673546314239502, "num_tokens": 212477595.0, "step": 5594 }, { "epoch": 0.7117415087139041, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.84720516204834, "learning_rate": 1e-05, "loss": 0.5531, "mean_token_accuracy": 0.8407877683639526, "num_tokens": 212513786.0, "step": 5595 }, { "epoch": 0.7118687189924946, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3969504833221436, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8563440442085266, "num_tokens": 212548111.0, "step": 5596 }, { "epoch": 0.7119959292710851, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6800637245178223, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.8534542322158813, "num_tokens": 212585395.0, "step": 5597 }, { "epoch": 0.7121231395496757, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.382744312286377, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.8700339794158936, "num_tokens": 212625351.0, "step": 5598 }, { "epoch": 0.7122503498282661, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.554056167602539, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8516234159469604, "num_tokens": 212657689.0, "step": 5599 }, { "epoch": 0.7123775601068566, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3296899795532227, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8736023306846619, "num_tokens": 212694741.0, "step": 5600 }, { "epoch": 0.7125047703854471, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4237730503082275, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8546097278594971, "num_tokens": 212727109.0, "step": 5601 }, { "epoch": 0.7126319806640377, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2811150550842285, "learning_rate": 1e-05, "loss": 0.5188, "mean_token_accuracy": 0.8468512892723083, "num_tokens": 212763031.0, "step": 5602 }, { "epoch": 0.7127591909426282, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.338263750076294, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.861984133720398, "num_tokens": 212803601.0, "step": 5603 }, { "epoch": 0.7128864012212187, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.218397617340088, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8740304112434387, "num_tokens": 212848331.0, "step": 5604 }, { "epoch": 0.7130136114998091, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3088128566741943, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8562599420547485, "num_tokens": 212890364.0, "step": 5605 }, { "epoch": 0.7131408217783997, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.165574312210083, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8635079264640808, "num_tokens": 212931234.0, "step": 5606 }, { "epoch": 0.7132680320569902, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.327371597290039, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8605023622512817, "num_tokens": 212970103.0, "step": 5607 }, { "epoch": 0.7133952423355807, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1217381954193115, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8532904982566833, "num_tokens": 213004895.0, "step": 5608 }, { "epoch": 0.7135224526141712, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.367380142211914, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8620439767837524, "num_tokens": 213044263.0, "step": 5609 }, { "epoch": 0.7136496628927618, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0611093044281006, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.855313777923584, "num_tokens": 213081685.0, "step": 5610 }, { "epoch": 0.7137768731713523, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4445531368255615, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.862255334854126, "num_tokens": 213116986.0, "step": 5611 }, { "epoch": 0.7139040834499427, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.9943182468414307, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8683594465255737, "num_tokens": 213156313.0, "step": 5612 }, { "epoch": 0.7140312937285332, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.6036064624786377, "learning_rate": 1e-05, "loss": 0.5263, "mean_token_accuracy": 0.8472614288330078, "num_tokens": 213187216.0, "step": 5613 }, { "epoch": 0.7141585040071238, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.9638617038726807, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8542299866676331, "num_tokens": 213223380.0, "step": 5614 }, { "epoch": 0.7142857142857143, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.785001754760742, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.874061107635498, "num_tokens": 213259100.0, "step": 5615 }, { "epoch": 0.7144129245643048, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0210869312286377, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.8503454923629761, "num_tokens": 213300103.0, "step": 5616 }, { "epoch": 0.7145401348428954, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9478280544281006, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8691773414611816, "num_tokens": 213338283.0, "step": 5617 }, { "epoch": 0.7146673451214858, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.1985924243927, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8657351732254028, "num_tokens": 213382211.0, "step": 5618 }, { "epoch": 0.7147945554000763, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9558510780334473, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8610552549362183, "num_tokens": 213418790.0, "step": 5619 }, { "epoch": 0.7149217656786668, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3931257724761963, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8601217269897461, "num_tokens": 213457843.0, "step": 5620 }, { "epoch": 0.7150489759572574, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.802366256713867, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8588405847549438, "num_tokens": 213496182.0, "step": 5621 }, { "epoch": 0.7151761862358479, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.4432013034820557, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8757482767105103, "num_tokens": 213535254.0, "step": 5622 }, { "epoch": 0.7153033965144384, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5965425968170166, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8629446625709534, "num_tokens": 213574331.0, "step": 5623 }, { "epoch": 0.7154306067930288, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.41188907623291, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8582970499992371, "num_tokens": 213608185.0, "step": 5624 }, { "epoch": 0.7155578170716194, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4251270294189453, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8673141598701477, "num_tokens": 213642032.0, "step": 5625 }, { "epoch": 0.7156850273502099, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3677985668182373, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8646310567855835, "num_tokens": 213674652.0, "step": 5626 }, { "epoch": 0.7158122376288004, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2988624572753906, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8658314347267151, "num_tokens": 213711112.0, "step": 5627 }, { "epoch": 0.715939447907391, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3157269954681396, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8505896329879761, "num_tokens": 213750319.0, "step": 5628 }, { "epoch": 0.7160666581859815, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.220125913619995, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8659576773643494, "num_tokens": 213788408.0, "step": 5629 }, { "epoch": 0.7161938684645719, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.294386386871338, "learning_rate": 1e-05, "loss": 0.5422, "mean_token_accuracy": 0.8398698568344116, "num_tokens": 213826822.0, "step": 5630 }, { "epoch": 0.7163210787431624, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1811633110046387, "learning_rate": 1e-05, "loss": 0.5452, "mean_token_accuracy": 0.84332275390625, "num_tokens": 213864046.0, "step": 5631 }, { "epoch": 0.716448289021753, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2593202590942383, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8640825748443604, "num_tokens": 213907235.0, "step": 5632 }, { "epoch": 0.7165754993003435, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1439247131347656, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8625677824020386, "num_tokens": 213942126.0, "step": 5633 }, { "epoch": 0.716702709578934, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2654895782470703, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8594416379928589, "num_tokens": 213980486.0, "step": 5634 }, { "epoch": 0.7168299198575245, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.1055805683135986, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8707662224769592, "num_tokens": 214016991.0, "step": 5635 }, { "epoch": 0.716957130136115, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3044772148132324, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8643310070037842, "num_tokens": 214049424.0, "step": 5636 }, { "epoch": 0.7170843404147055, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0439341068267822, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8602924346923828, "num_tokens": 214090422.0, "step": 5637 }, { "epoch": 0.717211550693296, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3822836875915527, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.869616687297821, "num_tokens": 214123075.0, "step": 5638 }, { "epoch": 0.7173387609718865, "ewc_loss": 0.05322265625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 3.065830707550049, "learning_rate": 1e-05, "loss": 0.5573, "mean_token_accuracy": 0.8394522666931152, "num_tokens": 214158962.0, "step": 5639 }, { "epoch": 0.7174659712504771, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.434394359588623, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8665516376495361, "num_tokens": 214201867.0, "step": 5640 }, { "epoch": 0.7175931815290676, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.9483718872070312, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8677805662155151, "num_tokens": 214241300.0, "step": 5641 }, { "epoch": 0.717720391807658, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5894880294799805, "learning_rate": 1e-05, "loss": 0.4946, "mean_token_accuracy": 0.8558658361434937, "num_tokens": 214276455.0, "step": 5642 }, { "epoch": 0.7178476020862485, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.900362730026245, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8634074926376343, "num_tokens": 214313737.0, "step": 5643 }, { "epoch": 0.7179748123648391, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8294789791107178, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8623180985450745, "num_tokens": 214351179.0, "step": 5644 }, { "epoch": 0.7181020226434296, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 2.9453189373016357, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8733981847763062, "num_tokens": 214387606.0, "step": 5645 }, { "epoch": 0.7182292329220201, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.100382328033447, "learning_rate": 1e-05, "loss": 0.5284, "mean_token_accuracy": 0.8514561653137207, "num_tokens": 214428149.0, "step": 5646 }, { "epoch": 0.7183564432006107, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2213387489318848, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8667261004447937, "num_tokens": 214468878.0, "step": 5647 }, { "epoch": 0.7184836534792011, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.166654109954834, "learning_rate": 1e-05, "loss": 0.5248, "mean_token_accuracy": 0.8512624502182007, "num_tokens": 214508813.0, "step": 5648 }, { "epoch": 0.7186108637577916, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5194764137268066, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8611634969711304, "num_tokens": 214553896.0, "step": 5649 }, { "epoch": 0.7187380740363821, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.980623245239258, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8643052577972412, "num_tokens": 214593727.0, "step": 5650 }, { "epoch": 0.7188652843149727, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6020424365997314, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8627155423164368, "num_tokens": 214627029.0, "step": 5651 }, { "epoch": 0.7189924945935632, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.757528305053711, "learning_rate": 1e-05, "loss": 0.5382, "mean_token_accuracy": 0.8448935747146606, "num_tokens": 214665484.0, "step": 5652 }, { "epoch": 0.7191197048721537, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.488374710083008, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8681817650794983, "num_tokens": 214708654.0, "step": 5653 }, { "epoch": 0.7192469151507441, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6013174057006836, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8538565635681152, "num_tokens": 214746962.0, "step": 5654 }, { "epoch": 0.7193741254293347, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.381727695465088, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8593766689300537, "num_tokens": 214784691.0, "step": 5655 }, { "epoch": 0.7195013357079252, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5106019973754883, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.86921626329422, "num_tokens": 214817182.0, "step": 5656 }, { "epoch": 0.7196285459865157, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2801616191864014, "learning_rate": 1e-05, "loss": 0.5157, "mean_token_accuracy": 0.8454264402389526, "num_tokens": 214857053.0, "step": 5657 }, { "epoch": 0.7197557562651062, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4781744480133057, "learning_rate": 1e-05, "loss": 0.5828, "mean_token_accuracy": 0.8357656598091125, "num_tokens": 214894908.0, "step": 5658 }, { "epoch": 0.7198829665436968, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1770007610321045, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8573206663131714, "num_tokens": 214930235.0, "step": 5659 }, { "epoch": 0.7200101768222873, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4888222217559814, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.869163990020752, "num_tokens": 214967140.0, "step": 5660 }, { "epoch": 0.7201373871008777, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.09381365776062, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.858820915222168, "num_tokens": 215004039.0, "step": 5661 }, { "epoch": 0.7202645973794682, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5848824977874756, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8692100048065186, "num_tokens": 215039305.0, "step": 5662 }, { "epoch": 0.7203918076580588, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0356366634368896, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8697155714035034, "num_tokens": 215076455.0, "step": 5663 }, { "epoch": 0.7205190179366493, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7294013500213623, "learning_rate": 1e-05, "loss": 0.5189, "mean_token_accuracy": 0.8507056832313538, "num_tokens": 215116826.0, "step": 5664 }, { "epoch": 0.7206462282152398, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0590827465057373, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.849725067615509, "num_tokens": 215162302.0, "step": 5665 }, { "epoch": 0.7207734384938304, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.882133722305298, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8544256687164307, "num_tokens": 215192402.0, "step": 5666 }, { "epoch": 0.7209006487724208, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2155983448028564, "learning_rate": 1e-05, "loss": 0.5167, "mean_token_accuracy": 0.8549526333808899, "num_tokens": 215228521.0, "step": 5667 }, { "epoch": 0.7210278590510113, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.874959945678711, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8608156442642212, "num_tokens": 215264997.0, "step": 5668 }, { "epoch": 0.7211550693296018, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.353425979614258, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8797910809516907, "num_tokens": 215301997.0, "step": 5669 }, { "epoch": 0.7212822796081924, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.731881856918335, "learning_rate": 1e-05, "loss": 0.5124, "mean_token_accuracy": 0.8526067733764648, "num_tokens": 215343637.0, "step": 5670 }, { "epoch": 0.7214094898867829, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4274559020996094, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8613649010658264, "num_tokens": 215381966.0, "step": 5671 }, { "epoch": 0.7215367001653734, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5141348838806152, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8568549156188965, "num_tokens": 215422930.0, "step": 5672 }, { "epoch": 0.7216639104439638, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4102437496185303, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8711371421813965, "num_tokens": 215461204.0, "step": 5673 }, { "epoch": 0.7217911207225544, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3356666564941406, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8593829274177551, "num_tokens": 215493771.0, "step": 5674 }, { "epoch": 0.7219183310011449, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.406003952026367, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8591769933700562, "num_tokens": 215529398.0, "step": 5675 }, { "epoch": 0.7220455412797354, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1776373386383057, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8661764860153198, "num_tokens": 215565794.0, "step": 5676 }, { "epoch": 0.7221727515583259, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001354217529296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4247989654541016, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8513296842575073, "num_tokens": 215604005.0, "step": 5677 }, { "epoch": 0.7222999618369165, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.081123113632202, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8714479804039001, "num_tokens": 215635325.0, "step": 5678 }, { "epoch": 0.7224271721155069, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.485980749130249, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.862683892250061, "num_tokens": 215667477.0, "step": 5679 }, { "epoch": 0.7225543823940974, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0077035427093506, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8510261178016663, "num_tokens": 215707087.0, "step": 5680 }, { "epoch": 0.722681592672688, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.602545976638794, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8642152547836304, "num_tokens": 215746182.0, "step": 5681 }, { "epoch": 0.7228088029512785, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.968888998031616, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8549518585205078, "num_tokens": 215785941.0, "step": 5682 }, { "epoch": 0.722936013229869, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.777623176574707, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.862832248210907, "num_tokens": 215823562.0, "step": 5683 }, { "epoch": 0.7230632235084595, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0323421955108643, "learning_rate": 1e-05, "loss": 0.5337, "mean_token_accuracy": 0.8394784331321716, "num_tokens": 215860454.0, "step": 5684 }, { "epoch": 0.72319043378705, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.935811996459961, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8681355714797974, "num_tokens": 215895088.0, "step": 5685 }, { "epoch": 0.7233176440656405, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2000834941864014, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8534978628158569, "num_tokens": 215935170.0, "step": 5686 }, { "epoch": 0.723444854344231, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.945955753326416, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8723954558372498, "num_tokens": 215977323.0, "step": 5687 }, { "epoch": 0.7235720646228215, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3698806762695312, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.871172308921814, "num_tokens": 216016729.0, "step": 5688 }, { "epoch": 0.7236992749014121, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8225646018981934, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8593475818634033, "num_tokens": 216058514.0, "step": 5689 }, { "epoch": 0.7238264851800026, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.43711256980896, "learning_rate": 1e-05, "loss": 0.5496, "mean_token_accuracy": 0.8401113748550415, "num_tokens": 216098633.0, "step": 5690 }, { "epoch": 0.723953695458593, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.627981662750244, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8804567456245422, "num_tokens": 216136124.0, "step": 5691 }, { "epoch": 0.7240809057371835, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3974194526672363, "learning_rate": 1e-05, "loss": 0.4973, "mean_token_accuracy": 0.8568013906478882, "num_tokens": 216178799.0, "step": 5692 }, { "epoch": 0.7242081160157741, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4738404750823975, "learning_rate": 1e-05, "loss": 0.5274, "mean_token_accuracy": 0.8453204035758972, "num_tokens": 216215233.0, "step": 5693 }, { "epoch": 0.7243353262943646, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3480453491210938, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8652247786521912, "num_tokens": 216252258.0, "step": 5694 }, { "epoch": 0.7244625365729551, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.360201120376587, "learning_rate": 1e-05, "loss": 0.5124, "mean_token_accuracy": 0.8479917049407959, "num_tokens": 216287845.0, "step": 5695 }, { "epoch": 0.7245897468515456, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.278188943862915, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8679032325744629, "num_tokens": 216325490.0, "step": 5696 }, { "epoch": 0.7247169571301361, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.3108432292938232, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8526003360748291, "num_tokens": 216359353.0, "step": 5697 }, { "epoch": 0.7248441674087266, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2394421100616455, "learning_rate": 1e-05, "loss": 0.5222, "mean_token_accuracy": 0.8510311841964722, "num_tokens": 216393620.0, "step": 5698 }, { "epoch": 0.7249713776873171, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.280440330505371, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8651270270347595, "num_tokens": 216427400.0, "step": 5699 }, { "epoch": 0.7250985879659076, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1750590801239014, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8573262691497803, "num_tokens": 216463817.0, "step": 5700 }, { "epoch": 0.7252257982444982, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2857470512390137, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8578651547431946, "num_tokens": 216495596.0, "step": 5701 }, { "epoch": 0.7253530085230887, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.128007650375366, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8647267818450928, "num_tokens": 216533003.0, "step": 5702 }, { "epoch": 0.7254802188016791, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3155086040496826, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.86318039894104, "num_tokens": 216569280.0, "step": 5703 }, { "epoch": 0.7256074290802697, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.076094627380371, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8712228536605835, "num_tokens": 216603073.0, "step": 5704 }, { "epoch": 0.7257346393588602, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001373291015625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.381833553314209, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8547549247741699, "num_tokens": 216641971.0, "step": 5705 }, { "epoch": 0.7258618496374507, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0081799030303955, "learning_rate": 1e-05, "loss": 0.5294, "mean_token_accuracy": 0.8433998823165894, "num_tokens": 216683119.0, "step": 5706 }, { "epoch": 0.7259890599160412, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.511948823928833, "learning_rate": 1e-05, "loss": 0.5522, "mean_token_accuracy": 0.836341142654419, "num_tokens": 216722139.0, "step": 5707 }, { "epoch": 0.7261162701946318, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.955580234527588, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8708360195159912, "num_tokens": 216757712.0, "step": 5708 }, { "epoch": 0.7262434804732223, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.687070369720459, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8511227369308472, "num_tokens": 216794132.0, "step": 5709 }, { "epoch": 0.7263706907518127, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.9241535663604736, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8558446764945984, "num_tokens": 216833937.0, "step": 5710 }, { "epoch": 0.7264979010304032, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.928250312805176, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8573824167251587, "num_tokens": 216872979.0, "step": 5711 }, { "epoch": 0.7266251113089938, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.0761635303497314, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8693221807479858, "num_tokens": 216915369.0, "step": 5712 }, { "epoch": 0.7267523215875843, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0831403732299805, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8640835285186768, "num_tokens": 216948876.0, "step": 5713 }, { "epoch": 0.7268795318661748, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3588781356811523, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8611412048339844, "num_tokens": 216988254.0, "step": 5714 }, { "epoch": 0.7270067421447653, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.992021083831787, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8583670258522034, "num_tokens": 217024792.0, "step": 5715 }, { "epoch": 0.7271339524233558, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5133981704711914, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8593460321426392, "num_tokens": 217063228.0, "step": 5716 }, { "epoch": 0.7272611627019463, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7901153564453125, "learning_rate": 1e-05, "loss": 0.562, "mean_token_accuracy": 0.8336156606674194, "num_tokens": 217102601.0, "step": 5717 }, { "epoch": 0.7273883729805368, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.492316961288452, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8647218942642212, "num_tokens": 217137839.0, "step": 5718 }, { "epoch": 0.7275155832591274, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.591700315475464, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.864174485206604, "num_tokens": 217179862.0, "step": 5719 }, { "epoch": 0.7276427935377179, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4190497398376465, "learning_rate": 1e-05, "loss": 0.5581, "mean_token_accuracy": 0.8365731239318848, "num_tokens": 217217728.0, "step": 5720 }, { "epoch": 0.7277700038163084, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.449492931365967, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8589811325073242, "num_tokens": 217255410.0, "step": 5721 }, { "epoch": 0.7278972140948988, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.330113172531128, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8510265350341797, "num_tokens": 217294365.0, "step": 5722 }, { "epoch": 0.7280244243734894, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3578765392303467, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8634785413742065, "num_tokens": 217333451.0, "step": 5723 }, { "epoch": 0.7281516346520799, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.269667625427246, "learning_rate": 1e-05, "loss": 0.4824, "mean_token_accuracy": 0.8552026748657227, "num_tokens": 217368737.0, "step": 5724 }, { "epoch": 0.7282788449306704, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3169033527374268, "learning_rate": 1e-05, "loss": 0.5251, "mean_token_accuracy": 0.8428509831428528, "num_tokens": 217409476.0, "step": 5725 }, { "epoch": 0.7284060552092609, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1940805912017822, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8677504658699036, "num_tokens": 217446481.0, "step": 5726 }, { "epoch": 0.7285332654878515, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3191630840301514, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8684351444244385, "num_tokens": 217482647.0, "step": 5727 }, { "epoch": 0.7286604757664419, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.134305238723755, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8561908602714539, "num_tokens": 217520783.0, "step": 5728 }, { "epoch": 0.7287876860450324, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.362792730331421, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8584939241409302, "num_tokens": 217557061.0, "step": 5729 }, { "epoch": 0.7289148963236229, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0639641284942627, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8700628280639648, "num_tokens": 217586833.0, "step": 5730 }, { "epoch": 0.7290421066022135, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.437196969985962, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8504658937454224, "num_tokens": 217622644.0, "step": 5731 }, { "epoch": 0.729169316880804, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.9941599369049072, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8635618686676025, "num_tokens": 217659345.0, "step": 5732 }, { "epoch": 0.7292965271593945, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.572826385498047, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8678520321846008, "num_tokens": 217698134.0, "step": 5733 }, { "epoch": 0.7294237374379849, "ewc_loss": 0.054931640625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.960216522216797, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8493514060974121, "num_tokens": 217730578.0, "step": 5734 }, { "epoch": 0.7295509477165755, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7546463012695312, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8756426572799683, "num_tokens": 217768090.0, "step": 5735 }, { "epoch": 0.729678157995166, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.9787423610687256, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8596440553665161, "num_tokens": 217809350.0, "step": 5736 }, { "epoch": 0.7298053682737565, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9545507431030273, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8552957773208618, "num_tokens": 217852478.0, "step": 5737 }, { "epoch": 0.7299325785523471, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.164297103881836, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8674644827842712, "num_tokens": 217885463.0, "step": 5738 }, { "epoch": 0.7300597888309376, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.021290302276611, "learning_rate": 1e-05, "loss": 0.5254, "mean_token_accuracy": 0.8514811396598816, "num_tokens": 217925712.0, "step": 5739 }, { "epoch": 0.730186999109528, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3889119625091553, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8574622869491577, "num_tokens": 217969628.0, "step": 5740 }, { "epoch": 0.7303142093881185, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.886720657348633, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8611807227134705, "num_tokens": 218012027.0, "step": 5741 }, { "epoch": 0.7304414196667091, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4753353595733643, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8543409109115601, "num_tokens": 218054018.0, "step": 5742 }, { "epoch": 0.7305686299452996, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6802029609680176, "learning_rate": 1e-05, "loss": 0.532, "mean_token_accuracy": 0.846839189529419, "num_tokens": 218099181.0, "step": 5743 }, { "epoch": 0.7306958402238901, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4596641063690186, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8664007186889648, "num_tokens": 218135330.0, "step": 5744 }, { "epoch": 0.7308230505024806, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.505080223083496, "learning_rate": 1e-05, "loss": 0.5387, "mean_token_accuracy": 0.8441264033317566, "num_tokens": 218171785.0, "step": 5745 }, { "epoch": 0.7309502607810711, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.407245397567749, "learning_rate": 1e-05, "loss": 0.5535, "mean_token_accuracy": 0.8368518352508545, "num_tokens": 218204230.0, "step": 5746 }, { "epoch": 0.7310774710596616, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.357964515686035, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8721116781234741, "num_tokens": 218237755.0, "step": 5747 }, { "epoch": 0.7312046813382521, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.337568521499634, "learning_rate": 1e-05, "loss": 0.5224, "mean_token_accuracy": 0.8445364832878113, "num_tokens": 218271850.0, "step": 5748 }, { "epoch": 0.7313318916168426, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2602016925811768, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8641201853752136, "num_tokens": 218308466.0, "step": 5749 }, { "epoch": 0.7314591018954332, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2924082279205322, "learning_rate": 1e-05, "loss": 0.5163, "mean_token_accuracy": 0.8454667925834656, "num_tokens": 218346133.0, "step": 5750 }, { "epoch": 0.7315863121740237, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2111897468566895, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8594403862953186, "num_tokens": 218384341.0, "step": 5751 }, { "epoch": 0.7317135224526141, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2613730430603027, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8620167970657349, "num_tokens": 218418314.0, "step": 5752 }, { "epoch": 0.7318407327312046, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.186513662338257, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8504805564880371, "num_tokens": 218455258.0, "step": 5753 }, { "epoch": 0.7319679430097952, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2467477321624756, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8593432307243347, "num_tokens": 218498733.0, "step": 5754 }, { "epoch": 0.7320951532883857, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1729824542999268, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8528605103492737, "num_tokens": 218541365.0, "step": 5755 }, { "epoch": 0.7322223635669762, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2478976249694824, "learning_rate": 1e-05, "loss": 0.5115, "mean_token_accuracy": 0.8494241237640381, "num_tokens": 218578292.0, "step": 5756 }, { "epoch": 0.7323495738455668, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1201930046081543, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8584228754043579, "num_tokens": 218617988.0, "step": 5757 }, { "epoch": 0.7324767841241572, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2680201530456543, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8558425903320312, "num_tokens": 218657759.0, "step": 5758 }, { "epoch": 0.7326039944027477, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.095844030380249, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8597326278686523, "num_tokens": 218693619.0, "step": 5759 }, { "epoch": 0.7327312046813382, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3173913955688477, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8640507459640503, "num_tokens": 218725978.0, "step": 5760 }, { "epoch": 0.7328584149599288, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.02188777923584, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8643884658813477, "num_tokens": 218769616.0, "step": 5761 }, { "epoch": 0.7329856252385193, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.416924238204956, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.863749086856842, "num_tokens": 218806053.0, "step": 5762 }, { "epoch": 0.7331128355171098, "ewc_loss": 0.0537109375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.9428229331970215, "learning_rate": 1e-05, "loss": 0.4924, "mean_token_accuracy": 0.8506306409835815, "num_tokens": 218844650.0, "step": 5763 }, { "epoch": 0.7332400457957003, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5969858169555664, "learning_rate": 1e-05, "loss": 0.5051, "mean_token_accuracy": 0.8506478071212769, "num_tokens": 218885943.0, "step": 5764 }, { "epoch": 0.7333672560742908, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.864804744720459, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8592239022254944, "num_tokens": 218923939.0, "step": 5765 }, { "epoch": 0.7334944663528813, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.890927791595459, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.860023021697998, "num_tokens": 218964252.0, "step": 5766 }, { "epoch": 0.7336216766314718, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 2.929905891418457, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8541090488433838, "num_tokens": 219009815.0, "step": 5767 }, { "epoch": 0.7337488869100623, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.213627338409424, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8644026517868042, "num_tokens": 219047998.0, "step": 5768 }, { "epoch": 0.7338760971886529, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.319331645965576, "learning_rate": 1e-05, "loss": 0.5423, "mean_token_accuracy": 0.843956470489502, "num_tokens": 219082447.0, "step": 5769 }, { "epoch": 0.7340033074672434, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00013828277587890625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.244865417480469, "learning_rate": 1e-05, "loss": 0.536, "mean_token_accuracy": 0.8483331203460693, "num_tokens": 219119393.0, "step": 5770 }, { "epoch": 0.7341305177458338, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.653017282485962, "learning_rate": 1e-05, "loss": 0.5527, "mean_token_accuracy": 0.8483963012695312, "num_tokens": 219155954.0, "step": 5771 }, { "epoch": 0.7342577280244243, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9875006675720215, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8577775955200195, "num_tokens": 219195960.0, "step": 5772 }, { "epoch": 0.7343849383030149, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.637279510498047, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8581190705299377, "num_tokens": 219238323.0, "step": 5773 }, { "epoch": 0.7345121485816054, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7805070877075195, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8634469509124756, "num_tokens": 219279044.0, "step": 5774 }, { "epoch": 0.7346393588601959, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5215046405792236, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8660240173339844, "num_tokens": 219312635.0, "step": 5775 }, { "epoch": 0.7347665691387865, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.636300802230835, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8744681477546692, "num_tokens": 219349424.0, "step": 5776 }, { "epoch": 0.7348937794173769, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3863134384155273, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8554391860961914, "num_tokens": 219389230.0, "step": 5777 }, { "epoch": 0.7350209896959674, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5720200538635254, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8585734367370605, "num_tokens": 219424082.0, "step": 5778 }, { "epoch": 0.7351481999745579, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2670648097991943, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8615020513534546, "num_tokens": 219463324.0, "step": 5779 }, { "epoch": 0.7352754102531485, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.558690309524536, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8785765171051025, "num_tokens": 219500653.0, "step": 5780 }, { "epoch": 0.735402620531739, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.155565023422241, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8689965009689331, "num_tokens": 219538906.0, "step": 5781 }, { "epoch": 0.7355298308103295, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6370396614074707, "learning_rate": 1e-05, "loss": 0.5454, "mean_token_accuracy": 0.8371915817260742, "num_tokens": 219578950.0, "step": 5782 }, { "epoch": 0.7356570410889199, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.086643934249878, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8733603358268738, "num_tokens": 219614298.0, "step": 5783 }, { "epoch": 0.7357842513675105, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7766494750976562, "learning_rate": 1e-05, "loss": 0.5097, "mean_token_accuracy": 0.8530527353286743, "num_tokens": 219655521.0, "step": 5784 }, { "epoch": 0.735911461646101, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1213462352752686, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8634160757064819, "num_tokens": 219689112.0, "step": 5785 }, { "epoch": 0.7360386719246915, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8985702991485596, "learning_rate": 1e-05, "loss": 0.57, "mean_token_accuracy": 0.834126353263855, "num_tokens": 219727432.0, "step": 5786 }, { "epoch": 0.736165882203282, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2521681785583496, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8575341701507568, "num_tokens": 219767231.0, "step": 5787 }, { "epoch": 0.7362930924818726, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8963592052459717, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8650919198989868, "num_tokens": 219798999.0, "step": 5788 }, { "epoch": 0.736420302760463, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3985443115234375, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8691431283950806, "num_tokens": 219844867.0, "step": 5789 }, { "epoch": 0.7365475130390535, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.730323314666748, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8603569269180298, "num_tokens": 219882817.0, "step": 5790 }, { "epoch": 0.736674723317644, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4679946899414062, "learning_rate": 1e-05, "loss": 0.5161, "mean_token_accuracy": 0.84445720911026, "num_tokens": 219916776.0, "step": 5791 }, { "epoch": 0.7368019335962346, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4992260932922363, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8778321743011475, "num_tokens": 219953471.0, "step": 5792 }, { "epoch": 0.7369291438748251, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.457043170928955, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8499103784561157, "num_tokens": 219983313.0, "step": 5793 }, { "epoch": 0.7370563541534156, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2904515266418457, "learning_rate": 1e-05, "loss": 0.5443, "mean_token_accuracy": 0.8392238020896912, "num_tokens": 220028222.0, "step": 5794 }, { "epoch": 0.737183564432006, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4355008602142334, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8593310117721558, "num_tokens": 220069005.0, "step": 5795 }, { "epoch": 0.7373107747105966, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.158825635910034, "learning_rate": 1e-05, "loss": 0.5021, "mean_token_accuracy": 0.8523333668708801, "num_tokens": 220103557.0, "step": 5796 }, { "epoch": 0.7374379849891871, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.458048105239868, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8592536449432373, "num_tokens": 220141405.0, "step": 5797 }, { "epoch": 0.7375651952677776, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.047545909881592, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8575659990310669, "num_tokens": 220184913.0, "step": 5798 }, { "epoch": 0.7376924055463682, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.544146776199341, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8558486700057983, "num_tokens": 220221886.0, "step": 5799 }, { "epoch": 0.7378196158249587, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.987987756729126, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8658187985420227, "num_tokens": 220260945.0, "step": 5800 }, { "epoch": 0.7379468261035491, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.698512315750122, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.862881064414978, "num_tokens": 220299580.0, "step": 5801 }, { "epoch": 0.7380740363821396, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.984154462814331, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8698498010635376, "num_tokens": 220338063.0, "step": 5802 }, { "epoch": 0.7382012466607302, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8797457218170166, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8665733337402344, "num_tokens": 220381212.0, "step": 5803 }, { "epoch": 0.7383284569393207, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1100168228149414, "learning_rate": 1e-05, "loss": 0.4993, "mean_token_accuracy": 0.8582031726837158, "num_tokens": 220422665.0, "step": 5804 }, { "epoch": 0.7384556672179112, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9890663623809814, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8681313991546631, "num_tokens": 220462044.0, "step": 5805 }, { "epoch": 0.7385828774965018, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3399815559387207, "learning_rate": 1e-05, "loss": 0.5475, "mean_token_accuracy": 0.8375917077064514, "num_tokens": 220497373.0, "step": 5806 }, { "epoch": 0.7387100877750922, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8890268802642822, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8720847368240356, "num_tokens": 220537612.0, "step": 5807 }, { "epoch": 0.7388372980536827, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4352400302886963, "learning_rate": 1e-05, "loss": 0.5084, "mean_token_accuracy": 0.8528261184692383, "num_tokens": 220574874.0, "step": 5808 }, { "epoch": 0.7389645083322732, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7130930423736572, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8716302514076233, "num_tokens": 220615699.0, "step": 5809 }, { "epoch": 0.7390917186108638, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4275667667388916, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8502694368362427, "num_tokens": 220654517.0, "step": 5810 }, { "epoch": 0.7392189288894543, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5502493381500244, "learning_rate": 1e-05, "loss": 0.5759, "mean_token_accuracy": 0.8325963020324707, "num_tokens": 220695915.0, "step": 5811 }, { "epoch": 0.7393461391680448, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.363785743713379, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8577516674995422, "num_tokens": 220734880.0, "step": 5812 }, { "epoch": 0.7394733494466353, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4425501823425293, "learning_rate": 1e-05, "loss": 0.4969, "mean_token_accuracy": 0.8547005653381348, "num_tokens": 220767800.0, "step": 5813 }, { "epoch": 0.7396005597252258, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.299158811569214, "learning_rate": 1e-05, "loss": 0.5225, "mean_token_accuracy": 0.847983717918396, "num_tokens": 220812253.0, "step": 5814 }, { "epoch": 0.7397277700038163, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3673791885375977, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8729585409164429, "num_tokens": 220847777.0, "step": 5815 }, { "epoch": 0.7398549802824068, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.228482961654663, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8597195744514465, "num_tokens": 220884974.0, "step": 5816 }, { "epoch": 0.7399821905609973, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3419573307037354, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8566925525665283, "num_tokens": 220926236.0, "step": 5817 }, { "epoch": 0.7401094008395879, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1708810329437256, "learning_rate": 1e-05, "loss": 0.5232, "mean_token_accuracy": 0.8445645570755005, "num_tokens": 220967985.0, "step": 5818 }, { "epoch": 0.7402366111181784, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3599936962127686, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8565717935562134, "num_tokens": 221004308.0, "step": 5819 }, { "epoch": 0.7403638213967688, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.0992910861968994, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8769594430923462, "num_tokens": 221043455.0, "step": 5820 }, { "epoch": 0.7404910316753593, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4225893020629883, "learning_rate": 1e-05, "loss": 0.5113, "mean_token_accuracy": 0.8483197689056396, "num_tokens": 221084634.0, "step": 5821 }, { "epoch": 0.7406182419539499, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0474588871002197, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8602091670036316, "num_tokens": 221117350.0, "step": 5822 }, { "epoch": 0.7407454522325404, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5264742374420166, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8592864274978638, "num_tokens": 221150864.0, "step": 5823 }, { "epoch": 0.7408726625111309, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.9819071292877197, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8729449510574341, "num_tokens": 221186501.0, "step": 5824 }, { "epoch": 0.7409998727897215, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6916277408599854, "learning_rate": 1e-05, "loss": 0.5574, "mean_token_accuracy": 0.8399890661239624, "num_tokens": 221227734.0, "step": 5825 }, { "epoch": 0.7411270830683119, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0019140243530273, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.857967734336853, "num_tokens": 221260615.0, "step": 5826 }, { "epoch": 0.7412542933469024, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00014019012451171875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8649158477783203, "learning_rate": 1e-05, "loss": 0.5222, "mean_token_accuracy": 0.8493683934211731, "num_tokens": 221298553.0, "step": 5827 }, { "epoch": 0.7413815036254929, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1134424209594727, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8608331680297852, "num_tokens": 221338803.0, "step": 5828 }, { "epoch": 0.7415087139040835, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9429564476013184, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8615819215774536, "num_tokens": 221379557.0, "step": 5829 }, { "epoch": 0.741635924182674, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3275818824768066, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8590545654296875, "num_tokens": 221412242.0, "step": 5830 }, { "epoch": 0.7417631344612645, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8326008319854736, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8754000663757324, "num_tokens": 221449382.0, "step": 5831 }, { "epoch": 0.7418903447398549, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.404270887374878, "learning_rate": 1e-05, "loss": 0.5162, "mean_token_accuracy": 0.8494753241539001, "num_tokens": 221489480.0, "step": 5832 }, { "epoch": 0.7420175550184455, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.654101610183716, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.854511022567749, "num_tokens": 221525264.0, "step": 5833 }, { "epoch": 0.742144765297036, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3996829986572266, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8648952841758728, "num_tokens": 221568447.0, "step": 5834 }, { "epoch": 0.7422719755756265, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4683163166046143, "learning_rate": 1e-05, "loss": 0.5584, "mean_token_accuracy": 0.8346837759017944, "num_tokens": 221612943.0, "step": 5835 }, { "epoch": 0.742399185854217, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3984806537628174, "learning_rate": 1e-05, "loss": 0.577, "mean_token_accuracy": 0.829809844493866, "num_tokens": 221645449.0, "step": 5836 }, { "epoch": 0.7425263961328076, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.308913469314575, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8730466961860657, "num_tokens": 221684041.0, "step": 5837 }, { "epoch": 0.742653606411398, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3573503494262695, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8684622049331665, "num_tokens": 221720200.0, "step": 5838 }, { "epoch": 0.7427808166899885, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.209672212600708, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8579071164131165, "num_tokens": 221758997.0, "step": 5839 }, { "epoch": 0.742908026968579, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3464558124542236, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8608666062355042, "num_tokens": 221800713.0, "step": 5840 }, { "epoch": 0.7430352372471696, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1471617221832275, "learning_rate": 1e-05, "loss": 0.5223, "mean_token_accuracy": 0.8469270467758179, "num_tokens": 221840336.0, "step": 5841 }, { "epoch": 0.7431624475257601, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3530919551849365, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8788102865219116, "num_tokens": 221876286.0, "step": 5842 }, { "epoch": 0.7432896578043506, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0869574546813965, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8525286316871643, "num_tokens": 221917596.0, "step": 5843 }, { "epoch": 0.743416868082941, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.401576519012451, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8664373755455017, "num_tokens": 221954045.0, "step": 5844 }, { "epoch": 0.7435440783615316, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.023385763168335, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8615463972091675, "num_tokens": 221996493.0, "step": 5845 }, { "epoch": 0.7436712886401221, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5043904781341553, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.856765866279602, "num_tokens": 222036608.0, "step": 5846 }, { "epoch": 0.7437984989187126, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.9613468647003174, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8630931377410889, "num_tokens": 222080647.0, "step": 5847 }, { "epoch": 0.7439257091973032, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6878161430358887, "learning_rate": 1e-05, "loss": 0.5174, "mean_token_accuracy": 0.8505855202674866, "num_tokens": 222114039.0, "step": 5848 }, { "epoch": 0.7440529194758937, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.9506075382232666, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.8524688482284546, "num_tokens": 222149233.0, "step": 5849 }, { "epoch": 0.7441801297544841, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.897747755050659, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8720675706863403, "num_tokens": 222185842.0, "step": 5850 }, { "epoch": 0.7443073400330746, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.0488100051879883, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8649560213088989, "num_tokens": 222233576.0, "step": 5851 }, { "epoch": 0.7444345503116652, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.076107501983643, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8616266846656799, "num_tokens": 222267822.0, "step": 5852 }, { "epoch": 0.7445617605902557, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.3207943439483643, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.87021803855896, "num_tokens": 222304694.0, "step": 5853 }, { "epoch": 0.7446889708688462, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0286431312561035, "learning_rate": 1e-05, "loss": 0.5169, "mean_token_accuracy": 0.8512539863586426, "num_tokens": 222340793.0, "step": 5854 }, { "epoch": 0.7448161811474368, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5017662048339844, "learning_rate": 1e-05, "loss": 0.5045, "mean_token_accuracy": 0.8564369082450867, "num_tokens": 222380316.0, "step": 5855 }, { "epoch": 0.7449433914260272, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8309216499328613, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8562893867492676, "num_tokens": 222418867.0, "step": 5856 }, { "epoch": 0.7450706017046177, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.496782064437866, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8649617433547974, "num_tokens": 222456585.0, "step": 5857 }, { "epoch": 0.7451978119832082, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6450154781341553, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8700683116912842, "num_tokens": 222492894.0, "step": 5858 }, { "epoch": 0.7453250222617988, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.421797752380371, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.862155020236969, "num_tokens": 222538214.0, "step": 5859 }, { "epoch": 0.7454522325403893, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4934427738189697, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.859220027923584, "num_tokens": 222577643.0, "step": 5860 }, { "epoch": 0.7455794428189798, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.347653388977051, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8571406602859497, "num_tokens": 222615080.0, "step": 5861 }, { "epoch": 0.7457066530975703, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3984293937683105, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8695791959762573, "num_tokens": 222650358.0, "step": 5862 }, { "epoch": 0.7458338633761608, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2700355052948, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8461483716964722, "num_tokens": 222691558.0, "step": 5863 }, { "epoch": 0.7459610736547513, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.34505295753479, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.858181357383728, "num_tokens": 222734632.0, "step": 5864 }, { "epoch": 0.7460882839333418, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2118449211120605, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8574944734573364, "num_tokens": 222769422.0, "step": 5865 }, { "epoch": 0.7462154942119323, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.329923629760742, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8792912364006042, "num_tokens": 222808667.0, "step": 5866 }, { "epoch": 0.7463427044905229, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1484880447387695, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8645514249801636, "num_tokens": 222844291.0, "step": 5867 }, { "epoch": 0.7464699147691134, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3689019680023193, "learning_rate": 1e-05, "loss": 0.5328, "mean_token_accuracy": 0.843660295009613, "num_tokens": 222885812.0, "step": 5868 }, { "epoch": 0.7465971250477038, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0792651176452637, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8609843850135803, "num_tokens": 222920768.0, "step": 5869 }, { "epoch": 0.7467243353262943, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.445321798324585, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8644500970840454, "num_tokens": 222961973.0, "step": 5870 }, { "epoch": 0.7468515456048849, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.005495071411133, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8640415072441101, "num_tokens": 222997356.0, "step": 5871 }, { "epoch": 0.7469787558834754, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.591200590133667, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8661126494407654, "num_tokens": 223035933.0, "step": 5872 }, { "epoch": 0.7471059661620659, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.9565582275390625, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8534029722213745, "num_tokens": 223074597.0, "step": 5873 }, { "epoch": 0.7472331764406565, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.796126127243042, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.87228924036026, "num_tokens": 223113736.0, "step": 5874 }, { "epoch": 0.7473603867192469, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.997917652130127, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8533952236175537, "num_tokens": 223153974.0, "step": 5875 }, { "epoch": 0.7474875969978374, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.016736030578613, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8666967153549194, "num_tokens": 223187043.0, "step": 5876 }, { "epoch": 0.7476148072764279, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000141143798828125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.229011058807373, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8722800612449646, "num_tokens": 223221265.0, "step": 5877 }, { "epoch": 0.7477420175550185, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.028987407684326, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8577260971069336, "num_tokens": 223262167.0, "step": 5878 }, { "epoch": 0.747869227833609, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4535913467407227, "learning_rate": 1e-05, "loss": 0.5256, "mean_token_accuracy": 0.8499021530151367, "num_tokens": 223302333.0, "step": 5879 }, { "epoch": 0.7479964381121995, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8531200885772705, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.869978666305542, "num_tokens": 223343337.0, "step": 5880 }, { "epoch": 0.7481236483907899, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.490896463394165, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8737648725509644, "num_tokens": 223382905.0, "step": 5881 }, { "epoch": 0.7482508586693805, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6473612785339355, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8628541231155396, "num_tokens": 223422177.0, "step": 5882 }, { "epoch": 0.748378068947971, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.444586992263794, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.858890175819397, "num_tokens": 223465725.0, "step": 5883 }, { "epoch": 0.7485052792265615, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4761650562286377, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8523456454277039, "num_tokens": 223507246.0, "step": 5884 }, { "epoch": 0.748632489505152, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3635053634643555, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8684882521629333, "num_tokens": 223550736.0, "step": 5885 }, { "epoch": 0.7487596997837426, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3716986179351807, "learning_rate": 1e-05, "loss": 0.522, "mean_token_accuracy": 0.8461971282958984, "num_tokens": 223588096.0, "step": 5886 }, { "epoch": 0.748886910062333, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.300926685333252, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8608953356742859, "num_tokens": 223629414.0, "step": 5887 }, { "epoch": 0.7490141203409235, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.309999465942383, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8585978746414185, "num_tokens": 223664468.0, "step": 5888 }, { "epoch": 0.749141330619514, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2494778633117676, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8703508377075195, "num_tokens": 223699832.0, "step": 5889 }, { "epoch": 0.7492685408981046, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.274538516998291, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8584245443344116, "num_tokens": 223736774.0, "step": 5890 }, { "epoch": 0.7493957511766951, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.202322483062744, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8627275228500366, "num_tokens": 223774622.0, "step": 5891 }, { "epoch": 0.7495229614552856, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2657296657562256, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8565443754196167, "num_tokens": 223816845.0, "step": 5892 }, { "epoch": 0.749650171733876, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.161210775375366, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8758751153945923, "num_tokens": 223849964.0, "step": 5893 }, { "epoch": 0.7497773820124666, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.295421600341797, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.8511567711830139, "num_tokens": 223887821.0, "step": 5894 }, { "epoch": 0.7499045922910571, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.103825092315674, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8749442100524902, "num_tokens": 223929420.0, "step": 5895 }, { "epoch": 0.7500318025696476, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.341456890106201, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8682727813720703, "num_tokens": 223964910.0, "step": 5896 }, { "epoch": 0.7501590128482382, "ewc_loss": 0.05419921875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 3.051382064819336, "learning_rate": 1e-05, "loss": 0.5158, "mean_token_accuracy": 0.8429687023162842, "num_tokens": 224001131.0, "step": 5897 }, { "epoch": 0.7502862231268287, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4353067874908447, "learning_rate": 1e-05, "loss": 0.5328, "mean_token_accuracy": 0.8436827659606934, "num_tokens": 224038969.0, "step": 5898 }, { "epoch": 0.7504134334054191, "ewc_loss": 0.053955078125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.965117931365967, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8495547771453857, "num_tokens": 224081044.0, "step": 5899 }, { "epoch": 0.7505406436840096, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.612941026687622, "learning_rate": 1e-05, "loss": 0.5698, "mean_token_accuracy": 0.8347532749176025, "num_tokens": 224117967.0, "step": 5900 }, { "epoch": 0.7506678539626002, "ewc_loss": 0.054443359375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.926758050918579, "learning_rate": 1e-05, "loss": 0.5178, "mean_token_accuracy": 0.8452690839767456, "num_tokens": 224148890.0, "step": 5901 }, { "epoch": 0.7507950642411907, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.847898483276367, "learning_rate": 1e-05, "loss": 0.5413, "mean_token_accuracy": 0.843399167060852, "num_tokens": 224183592.0, "step": 5902 }, { "epoch": 0.7509222745197812, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.9533157348632812, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8628566265106201, "num_tokens": 224225983.0, "step": 5903 }, { "epoch": 0.7510494847983717, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1141676902771, "learning_rate": 1e-05, "loss": 0.5443, "mean_token_accuracy": 0.8471726775169373, "num_tokens": 224267068.0, "step": 5904 }, { "epoch": 0.7511766950769622, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.233232021331787, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8791699409484863, "num_tokens": 224313060.0, "step": 5905 }, { "epoch": 0.7513039053555527, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.18874454498291, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8533976078033447, "num_tokens": 224355066.0, "step": 5906 }, { "epoch": 0.7514311156341432, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.582728862762451, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8703226447105408, "num_tokens": 224383930.0, "step": 5907 }, { "epoch": 0.7515583259127337, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.964409351348877, "learning_rate": 1e-05, "loss": 0.5428, "mean_token_accuracy": 0.8438457250595093, "num_tokens": 224418462.0, "step": 5908 }, { "epoch": 0.7516855361913243, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.601309299468994, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8634323477745056, "num_tokens": 224453672.0, "step": 5909 }, { "epoch": 0.7518127464699148, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7400171756744385, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8640683889389038, "num_tokens": 224494093.0, "step": 5910 }, { "epoch": 0.7519399567485053, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5004100799560547, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8644875288009644, "num_tokens": 224531829.0, "step": 5911 }, { "epoch": 0.7520671670270958, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.582641363143921, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.850530743598938, "num_tokens": 224569403.0, "step": 5912 }, { "epoch": 0.7521943773056863, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4094974994659424, "learning_rate": 1e-05, "loss": 0.5072, "mean_token_accuracy": 0.8544427156448364, "num_tokens": 224605189.0, "step": 5913 }, { "epoch": 0.7523215875842768, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.477419853210449, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8582335114479065, "num_tokens": 224638729.0, "step": 5914 }, { "epoch": 0.7524487978628673, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.301985263824463, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8630222678184509, "num_tokens": 224673442.0, "step": 5915 }, { "epoch": 0.7525760081414579, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00014209747314453125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.441211462020874, "learning_rate": 1e-05, "loss": 0.5234, "mean_token_accuracy": 0.8456266522407532, "num_tokens": 224705945.0, "step": 5916 }, { "epoch": 0.7527032184200484, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1895694732666016, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8608025312423706, "num_tokens": 224740035.0, "step": 5917 }, { "epoch": 0.7528304286986388, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4503254890441895, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8729397058486938, "num_tokens": 224782047.0, "step": 5918 }, { "epoch": 0.7529576389772293, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0950210094451904, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8518268465995789, "num_tokens": 224821755.0, "step": 5919 }, { "epoch": 0.7530848492558199, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5445220470428467, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8690474033355713, "num_tokens": 224854822.0, "step": 5920 }, { "epoch": 0.7532120595344104, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0157546997070312, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8653647303581238, "num_tokens": 224899570.0, "step": 5921 }, { "epoch": 0.7533392698130009, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7081220149993896, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.8640807867050171, "num_tokens": 224934897.0, "step": 5922 }, { "epoch": 0.7534664800915915, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.014901638031006, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.871304988861084, "num_tokens": 224972866.0, "step": 5923 }, { "epoch": 0.7535936903701819, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8996617794036865, "learning_rate": 1e-05, "loss": 0.5412, "mean_token_accuracy": 0.8456395864486694, "num_tokens": 225015075.0, "step": 5924 }, { "epoch": 0.7537209006487724, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.171381950378418, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8613649010658264, "num_tokens": 225046729.0, "step": 5925 }, { "epoch": 0.7538481109273629, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.976332187652588, "learning_rate": 1e-05, "loss": 0.5157, "mean_token_accuracy": 0.8558064699172974, "num_tokens": 225080117.0, "step": 5926 }, { "epoch": 0.7539753212059535, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3632471561431885, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8601424098014832, "num_tokens": 225116854.0, "step": 5927 }, { "epoch": 0.754102531484544, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8505210876464844, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8672090768814087, "num_tokens": 225154044.0, "step": 5928 }, { "epoch": 0.7542297417631345, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.462000608444214, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.881373941898346, "num_tokens": 225191393.0, "step": 5929 }, { "epoch": 0.7543569520417249, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.640049934387207, "learning_rate": 1e-05, "loss": 0.5138, "mean_token_accuracy": 0.850185215473175, "num_tokens": 225232577.0, "step": 5930 }, { "epoch": 0.7544841623203155, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4636294841766357, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8478682041168213, "num_tokens": 225270458.0, "step": 5931 }, { "epoch": 0.754611372598906, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.43361496925354, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8591315746307373, "num_tokens": 225306777.0, "step": 5932 }, { "epoch": 0.7547385828774965, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.416914701461792, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8690469264984131, "num_tokens": 225345602.0, "step": 5933 }, { "epoch": 0.754865793156087, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2857108116149902, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8676666617393494, "num_tokens": 225378885.0, "step": 5934 }, { "epoch": 0.7549930034346776, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4069409370422363, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8617982864379883, "num_tokens": 225408421.0, "step": 5935 }, { "epoch": 0.755120213713268, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1805200576782227, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8620920181274414, "num_tokens": 225441428.0, "step": 5936 }, { "epoch": 0.7552474239918585, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.420130968093872, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8506616353988647, "num_tokens": 225474912.0, "step": 5937 }, { "epoch": 0.755374634270449, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.093574285507202, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8568031787872314, "num_tokens": 225511876.0, "step": 5938 }, { "epoch": 0.7555018445490396, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.470616340637207, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8580560684204102, "num_tokens": 225548754.0, "step": 5939 }, { "epoch": 0.7556290548276301, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001430511474609375, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.025613307952881, "learning_rate": 1e-05, "loss": 0.4993, "mean_token_accuracy": 0.8506155014038086, "num_tokens": 225585554.0, "step": 5940 }, { "epoch": 0.7557562651062206, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014400482177734375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5705153942108154, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.865670919418335, "num_tokens": 225622727.0, "step": 5941 }, { "epoch": 0.755883475384811, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00014400482177734375, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.971280813217163, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8707588911056519, "num_tokens": 225662163.0, "step": 5942 }, { "epoch": 0.7560106856634016, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7527215480804443, "learning_rate": 1e-05, "loss": 0.5139, "mean_token_accuracy": 0.8484349250793457, "num_tokens": 225698932.0, "step": 5943 }, { "epoch": 0.7561378959419921, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.0060927867889404, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8711237907409668, "num_tokens": 225735587.0, "step": 5944 }, { "epoch": 0.7562651062205826, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9196486473083496, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8728047609329224, "num_tokens": 225777013.0, "step": 5945 }, { "epoch": 0.7563923164991732, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.165090560913086, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8638471364974976, "num_tokens": 225808496.0, "step": 5946 }, { "epoch": 0.7565195267777637, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9826929569244385, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8669700622558594, "num_tokens": 225848629.0, "step": 5947 }, { "epoch": 0.7566467370563541, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3552026748657227, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8698400855064392, "num_tokens": 225882284.0, "step": 5948 }, { "epoch": 0.7567739473349446, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.876589775085449, "learning_rate": 1e-05, "loss": 0.5173, "mean_token_accuracy": 0.8505837917327881, "num_tokens": 225919294.0, "step": 5949 }, { "epoch": 0.7569011576135352, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4459452629089355, "learning_rate": 1e-05, "loss": 0.5239, "mean_token_accuracy": 0.8472762703895569, "num_tokens": 225954742.0, "step": 5950 }, { "epoch": 0.7570283678921257, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6823654174804688, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8850786089897156, "num_tokens": 225992532.0, "step": 5951 }, { "epoch": 0.7571555781707162, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4145398139953613, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8613612651824951, "num_tokens": 226034272.0, "step": 5952 }, { "epoch": 0.7572827884493067, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.52616810798645, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.85368412733078, "num_tokens": 226066607.0, "step": 5953 }, { "epoch": 0.7574099987278972, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.367835521697998, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.8531129360198975, "num_tokens": 226103077.0, "step": 5954 }, { "epoch": 0.7575372090064877, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3884873390197754, "learning_rate": 1e-05, "loss": 0.4125, "mean_token_accuracy": 0.8761399984359741, "num_tokens": 226140488.0, "step": 5955 }, { "epoch": 0.7576644192850782, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.306104898452759, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8766968250274658, "num_tokens": 226177139.0, "step": 5956 }, { "epoch": 0.7577916295636687, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.315434455871582, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8643382787704468, "num_tokens": 226216340.0, "step": 5957 }, { "epoch": 0.7579188398422593, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2614693641662598, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8788138628005981, "num_tokens": 226253507.0, "step": 5958 }, { "epoch": 0.7580460501208498, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.275263786315918, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.8553392291069031, "num_tokens": 226296617.0, "step": 5959 }, { "epoch": 0.7581732603994403, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2179129123687744, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8562116622924805, "num_tokens": 226335483.0, "step": 5960 }, { "epoch": 0.7583004706780307, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2652487754821777, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8492305278778076, "num_tokens": 226373773.0, "step": 5961 }, { "epoch": 0.7584276809566213, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1719460487365723, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8645234704017639, "num_tokens": 226410333.0, "step": 5962 }, { "epoch": 0.7585548912352118, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2699990272521973, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8624387979507446, "num_tokens": 226451058.0, "step": 5963 }, { "epoch": 0.7586821015138023, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.128981828689575, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8654908537864685, "num_tokens": 226487099.0, "step": 5964 }, { "epoch": 0.7588093117923929, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3075735569000244, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8672561049461365, "num_tokens": 226521462.0, "step": 5965 }, { "epoch": 0.7589365220709834, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0823512077331543, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8605000972747803, "num_tokens": 226560573.0, "step": 5966 }, { "epoch": 0.7590637323495738, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.371157646179199, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8574584722518921, "num_tokens": 226603717.0, "step": 5967 }, { "epoch": 0.7591909426281643, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 3.000952959060669, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8664011359214783, "num_tokens": 226647049.0, "step": 5968 }, { "epoch": 0.7593181529067549, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.5040383338928223, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8708117604255676, "num_tokens": 226682459.0, "step": 5969 }, { "epoch": 0.7594453631853454, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9259603023529053, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8555372953414917, "num_tokens": 226727381.0, "step": 5970 }, { "epoch": 0.7595725734639359, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.723414659500122, "learning_rate": 1e-05, "loss": 0.5366, "mean_token_accuracy": 0.8422254323959351, "num_tokens": 226766732.0, "step": 5971 }, { "epoch": 0.7596997837425264, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.9119603633880615, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.856163740158081, "num_tokens": 226808165.0, "step": 5972 }, { "epoch": 0.7598269940211169, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.014431476593018, "learning_rate": 1e-05, "loss": 0.5199, "mean_token_accuracy": 0.8540228009223938, "num_tokens": 226848113.0, "step": 5973 }, { "epoch": 0.7599542042997074, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.0998177528381348, "learning_rate": 1e-05, "loss": 0.5473, "mean_token_accuracy": 0.839949369430542, "num_tokens": 226889390.0, "step": 5974 }, { "epoch": 0.7600814145782979, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.20469331741333, "learning_rate": 1e-05, "loss": 0.53, "mean_token_accuracy": 0.8507076501846313, "num_tokens": 226929111.0, "step": 5975 }, { "epoch": 0.7602086248568884, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.443551778793335, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8751525282859802, "num_tokens": 226966677.0, "step": 5976 }, { "epoch": 0.760335835135479, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.084686279296875, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8576826453208923, "num_tokens": 227004083.0, "step": 5977 }, { "epoch": 0.7604630454140695, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6077396869659424, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8578150272369385, "num_tokens": 227049954.0, "step": 5978 }, { "epoch": 0.7605902556926599, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.85518479347229, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8605117201805115, "num_tokens": 227088532.0, "step": 5979 }, { "epoch": 0.7607174659712505, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.549332857131958, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8761754035949707, "num_tokens": 227129471.0, "step": 5980 }, { "epoch": 0.760844676249841, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6762421131134033, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8574104905128479, "num_tokens": 227167443.0, "step": 5981 }, { "epoch": 0.7609718865284315, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.440033435821533, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8706135749816895, "num_tokens": 227207577.0, "step": 5982 }, { "epoch": 0.761099096807022, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.553891897201538, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8616863489151001, "num_tokens": 227247424.0, "step": 5983 }, { "epoch": 0.7612263070856126, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.32885479927063, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8613230586051941, "num_tokens": 227286361.0, "step": 5984 }, { "epoch": 0.761353517364203, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4891088008880615, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8680894374847412, "num_tokens": 227326786.0, "step": 5985 }, { "epoch": 0.7614807276427935, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2255187034606934, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.85068678855896, "num_tokens": 227370237.0, "step": 5986 }, { "epoch": 0.761607937921384, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5075247287750244, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8812756538391113, "num_tokens": 227399711.0, "step": 5987 }, { "epoch": 0.7617351481999746, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1395456790924072, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.868303656578064, "num_tokens": 227436468.0, "step": 5988 }, { "epoch": 0.7618623584785651, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5677340030670166, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.8599782586097717, "num_tokens": 227476128.0, "step": 5989 }, { "epoch": 0.7619895687571556, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0767693519592285, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8630539178848267, "num_tokens": 227510499.0, "step": 5990 }, { "epoch": 0.762116779035746, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.705493688583374, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.857930064201355, "num_tokens": 227542886.0, "step": 5991 }, { "epoch": 0.7622439893143366, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0523698329925537, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8705225586891174, "num_tokens": 227583244.0, "step": 5992 }, { "epoch": 0.7623711995929271, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.854461431503296, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8674551248550415, "num_tokens": 227622690.0, "step": 5993 }, { "epoch": 0.7624984098715176, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.158477306365967, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8591417074203491, "num_tokens": 227663537.0, "step": 5994 }, { "epoch": 0.7626256201501082, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9328901767730713, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.862226128578186, "num_tokens": 227700679.0, "step": 5995 }, { "epoch": 0.7627528304286987, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.351590156555176, "learning_rate": 1e-05, "loss": 0.5154, "mean_token_accuracy": 0.8516545295715332, "num_tokens": 227738410.0, "step": 5996 }, { "epoch": 0.7628800407072891, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8041558265686035, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8661178350448608, "num_tokens": 227780849.0, "step": 5997 }, { "epoch": 0.7630072509858796, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4435477256774902, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8691722750663757, "num_tokens": 227815488.0, "step": 5998 }, { "epoch": 0.7631344612644702, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6004600524902344, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8682107925415039, "num_tokens": 227848627.0, "step": 5999 }, { "epoch": 0.7632616715430607, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4406425952911377, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8613746166229248, "num_tokens": 227888394.0, "step": 6000 }, { "epoch": 0.7633888818216512, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00014495849609375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3860626220703125, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.869073748588562, "num_tokens": 227923785.0, "step": 6001 }, { "epoch": 0.7635160921002417, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4297633171081543, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8531517386436462, "num_tokens": 227958714.0, "step": 6002 }, { "epoch": 0.7636433023788322, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2184834480285645, "learning_rate": 1e-05, "loss": 0.5474, "mean_token_accuracy": 0.8353402018547058, "num_tokens": 228004414.0, "step": 6003 }, { "epoch": 0.7637705126574227, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4452829360961914, "learning_rate": 1e-05, "loss": 0.5251, "mean_token_accuracy": 0.8472366333007812, "num_tokens": 228044291.0, "step": 6004 }, { "epoch": 0.7638977229360132, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.1066479682922363, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8679195642471313, "num_tokens": 228080656.0, "step": 6005 }, { "epoch": 0.7640249332146037, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5031049251556396, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8631780743598938, "num_tokens": 228117787.0, "step": 6006 }, { "epoch": 0.7641521434931943, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0401690006256104, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8549166917800903, "num_tokens": 228150620.0, "step": 6007 }, { "epoch": 0.7642793537717848, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.7335357666015625, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8716303110122681, "num_tokens": 228188330.0, "step": 6008 }, { "epoch": 0.7644065640503753, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.9862937927246094, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8680371046066284, "num_tokens": 228231369.0, "step": 6009 }, { "epoch": 0.7645337743289657, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8523428440093994, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8694822788238525, "num_tokens": 228269321.0, "step": 6010 }, { "epoch": 0.7646609846075563, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.0848824977874756, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8698387145996094, "num_tokens": 228305240.0, "step": 6011 }, { "epoch": 0.7647881948861468, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.988401412963867, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8633223176002502, "num_tokens": 228345100.0, "step": 6012 }, { "epoch": 0.7649154051647373, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.292875289916992, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8661617040634155, "num_tokens": 228379215.0, "step": 6013 }, { "epoch": 0.7650426154433279, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.935886859893799, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8582453727722168, "num_tokens": 228415954.0, "step": 6014 }, { "epoch": 0.7651698257219184, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4379773139953613, "learning_rate": 1e-05, "loss": 0.521, "mean_token_accuracy": 0.8469715118408203, "num_tokens": 228453569.0, "step": 6015 }, { "epoch": 0.7652970360005088, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7742974758148193, "learning_rate": 1e-05, "loss": 0.5489, "mean_token_accuracy": 0.8454395532608032, "num_tokens": 228491384.0, "step": 6016 }, { "epoch": 0.7654242462790993, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.445852279663086, "learning_rate": 1e-05, "loss": 0.5282, "mean_token_accuracy": 0.844847559928894, "num_tokens": 228528118.0, "step": 6017 }, { "epoch": 0.7655514565576899, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5891456604003906, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8554221391677856, "num_tokens": 228568207.0, "step": 6018 }, { "epoch": 0.7656786668362804, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3877480030059814, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8600652813911438, "num_tokens": 228604313.0, "step": 6019 }, { "epoch": 0.7658058771148709, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4611551761627197, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8763440847396851, "num_tokens": 228645615.0, "step": 6020 }, { "epoch": 0.7659330873934614, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3208870887756348, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8572994470596313, "num_tokens": 228679137.0, "step": 6021 }, { "epoch": 0.7660602976720519, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.388216495513916, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8677840828895569, "num_tokens": 228717383.0, "step": 6022 }, { "epoch": 0.7661875079506424, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2537007331848145, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8739039897918701, "num_tokens": 228747260.0, "step": 6023 }, { "epoch": 0.7663147182292329, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.370082378387451, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8546695709228516, "num_tokens": 228780852.0, "step": 6024 }, { "epoch": 0.7664419285078234, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.183844804763794, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8656755089759827, "num_tokens": 228820220.0, "step": 6025 }, { "epoch": 0.766569138786414, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.375340461730957, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8674054145812988, "num_tokens": 228858903.0, "step": 6026 }, { "epoch": 0.7666963490650045, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1087098121643066, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8744053244590759, "num_tokens": 228894998.0, "step": 6027 }, { "epoch": 0.7668235593435949, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4440340995788574, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8667352199554443, "num_tokens": 228933333.0, "step": 6028 }, { "epoch": 0.7669507696221854, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0451788902282715, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8637056350708008, "num_tokens": 228972960.0, "step": 6029 }, { "epoch": 0.767077979900776, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5633881092071533, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.852094292640686, "num_tokens": 229010967.0, "step": 6030 }, { "epoch": 0.7672051901793665, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.986023187637329, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8670804500579834, "num_tokens": 229052424.0, "step": 6031 }, { "epoch": 0.767332400457957, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7434487342834473, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8597325086593628, "num_tokens": 229088106.0, "step": 6032 }, { "epoch": 0.7674596107365476, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.024507522583008, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8652425408363342, "num_tokens": 229122340.0, "step": 6033 }, { "epoch": 0.767586821015138, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9184229373931885, "learning_rate": 1e-05, "loss": 0.5257, "mean_token_accuracy": 0.8523792028427124, "num_tokens": 229158299.0, "step": 6034 }, { "epoch": 0.7677140312937285, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.187356948852539, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.8597391843795776, "num_tokens": 229187870.0, "step": 6035 }, { "epoch": 0.767841241572319, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.932502508163452, "learning_rate": 1e-05, "loss": 0.519, "mean_token_accuracy": 0.8492351770401001, "num_tokens": 229226459.0, "step": 6036 }, { "epoch": 0.7679684518509096, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.363990306854248, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8630685806274414, "num_tokens": 229257973.0, "step": 6037 }, { "epoch": 0.7680956621295001, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8030190467834473, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8661119937896729, "num_tokens": 229297870.0, "step": 6038 }, { "epoch": 0.7682228724080906, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.4348161220550537, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.8547732830047607, "num_tokens": 229331295.0, "step": 6039 }, { "epoch": 0.768350082686681, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5951249599456787, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8603390455245972, "num_tokens": 229369491.0, "step": 6040 }, { "epoch": 0.7684772929652716, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.423731803894043, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8545424938201904, "num_tokens": 229407836.0, "step": 6041 }, { "epoch": 0.7686045032438621, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.403409481048584, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8660521507263184, "num_tokens": 229444806.0, "step": 6042 }, { "epoch": 0.7687317135224526, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3876864910125732, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8664520382881165, "num_tokens": 229485588.0, "step": 6043 }, { "epoch": 0.7688589238010431, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2688300609588623, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8680624961853027, "num_tokens": 229521924.0, "step": 6044 }, { "epoch": 0.7689861340796337, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.378145217895508, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8661863803863525, "num_tokens": 229561511.0, "step": 6045 }, { "epoch": 0.7691133443582241, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.172569513320923, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8612375259399414, "num_tokens": 229607193.0, "step": 6046 }, { "epoch": 0.7692405546368146, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.386847496032715, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8709720373153687, "num_tokens": 229642775.0, "step": 6047 }, { "epoch": 0.7693677649154052, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.119375467300415, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8476548194885254, "num_tokens": 229680960.0, "step": 6048 }, { "epoch": 0.7694949751939957, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4289236068725586, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8753284215927124, "num_tokens": 229713473.0, "step": 6049 }, { "epoch": 0.7696221854725862, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0401344299316406, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8736351132392883, "num_tokens": 229753919.0, "step": 6050 }, { "epoch": 0.7697493957511767, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.535827875137329, "learning_rate": 1e-05, "loss": 0.5339, "mean_token_accuracy": 0.847021222114563, "num_tokens": 229788635.0, "step": 6051 }, { "epoch": 0.7698766060297672, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.970695734024048, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8676479458808899, "num_tokens": 229831221.0, "step": 6052 }, { "epoch": 0.7700038163083577, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7163047790527344, "learning_rate": 1e-05, "loss": 0.5106, "mean_token_accuracy": 0.8506680130958557, "num_tokens": 229866762.0, "step": 6053 }, { "epoch": 0.7701310265869482, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.9694297313690186, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8674705028533936, "num_tokens": 229904715.0, "step": 6054 }, { "epoch": 0.7702582368655387, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9168543815612793, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8649113178253174, "num_tokens": 229942938.0, "step": 6055 }, { "epoch": 0.7703854471441293, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.0918450355529785, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.867396354675293, "num_tokens": 229982468.0, "step": 6056 }, { "epoch": 0.7705126574227198, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.058620929718018, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8541584014892578, "num_tokens": 230022386.0, "step": 6057 }, { "epoch": 0.7706398677013102, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3646745681762695, "learning_rate": 1e-05, "loss": 0.528, "mean_token_accuracy": 0.849859356880188, "num_tokens": 230054871.0, "step": 6058 }, { "epoch": 0.7707670779799007, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.961869478225708, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8703827261924744, "num_tokens": 230090996.0, "step": 6059 }, { "epoch": 0.7708942882584913, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.479738235473633, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8689223527908325, "num_tokens": 230124575.0, "step": 6060 }, { "epoch": 0.7710214985370818, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7771196365356445, "learning_rate": 1e-05, "loss": 0.5483, "mean_token_accuracy": 0.8427543640136719, "num_tokens": 230168225.0, "step": 6061 }, { "epoch": 0.7711487088156723, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4684255123138428, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.8598822355270386, "num_tokens": 230208004.0, "step": 6062 }, { "epoch": 0.7712759190942629, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5982370376586914, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.854523777961731, "num_tokens": 230243247.0, "step": 6063 }, { "epoch": 0.7714031293728534, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3899214267730713, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8682288527488708, "num_tokens": 230286646.0, "step": 6064 }, { "epoch": 0.7715303396514438, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4791033267974854, "learning_rate": 1e-05, "loss": 0.5298, "mean_token_accuracy": 0.8418459892272949, "num_tokens": 230317705.0, "step": 6065 }, { "epoch": 0.7716575499300343, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3118155002593994, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8543187379837036, "num_tokens": 230362523.0, "step": 6066 }, { "epoch": 0.7717847602086249, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.408086061477661, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8538318872451782, "num_tokens": 230402678.0, "step": 6067 }, { "epoch": 0.7719119704872154, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2394180297851562, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.861185610294342, "num_tokens": 230440557.0, "step": 6068 }, { "epoch": 0.7720391807658059, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4008960723876953, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8613934516906738, "num_tokens": 230477602.0, "step": 6069 }, { "epoch": 0.7721663910443964, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1605887413024902, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8799103498458862, "num_tokens": 230518381.0, "step": 6070 }, { "epoch": 0.7722936013229869, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4203591346740723, "learning_rate": 1e-05, "loss": 0.5093, "mean_token_accuracy": 0.8546078205108643, "num_tokens": 230559615.0, "step": 6071 }, { "epoch": 0.7724208116015774, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0924875736236572, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8595766425132751, "num_tokens": 230595270.0, "step": 6072 }, { "epoch": 0.7725480218801679, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5056865215301514, "learning_rate": 1e-05, "loss": 0.5103, "mean_token_accuracy": 0.8507776260375977, "num_tokens": 230630405.0, "step": 6073 }, { "epoch": 0.7726752321587584, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0375211238861084, "learning_rate": 1e-05, "loss": 0.5327, "mean_token_accuracy": 0.8420094847679138, "num_tokens": 230670836.0, "step": 6074 }, { "epoch": 0.772802442437349, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6289515495300293, "learning_rate": 1e-05, "loss": 0.5249, "mean_token_accuracy": 0.8542155623435974, "num_tokens": 230709150.0, "step": 6075 }, { "epoch": 0.7729296527159395, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0194180011749268, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8579729795455933, "num_tokens": 230740306.0, "step": 6076 }, { "epoch": 0.7730568629945299, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7891030311584473, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8613700866699219, "num_tokens": 230776499.0, "step": 6077 }, { "epoch": 0.7731840732731204, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.058384418487549, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8629584312438965, "num_tokens": 230816614.0, "step": 6078 }, { "epoch": 0.773311283551711, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9116578102111816, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8546241521835327, "num_tokens": 230865271.0, "step": 6079 }, { "epoch": 0.7734384938303015, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.227419137954712, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8567758798599243, "num_tokens": 230903092.0, "step": 6080 }, { "epoch": 0.773565704108892, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8971524238586426, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8616619110107422, "num_tokens": 230942726.0, "step": 6081 }, { "epoch": 0.7736929143874826, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4534592628479004, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8566765189170837, "num_tokens": 230978694.0, "step": 6082 }, { "epoch": 0.773820124666073, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6797924041748047, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8570140600204468, "num_tokens": 231015466.0, "step": 6083 }, { "epoch": 0.7739473349446635, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4110770225524902, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.862828254699707, "num_tokens": 231054823.0, "step": 6084 }, { "epoch": 0.774074545223254, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.477818012237549, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8697482347488403, "num_tokens": 231087893.0, "step": 6085 }, { "epoch": 0.7742017555018446, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4014289379119873, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8605166077613831, "num_tokens": 231123446.0, "step": 6086 }, { "epoch": 0.7743289657804351, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3213155269622803, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8644781708717346, "num_tokens": 231151558.0, "step": 6087 }, { "epoch": 0.7744561760590256, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.375547409057617, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8638813495635986, "num_tokens": 231190715.0, "step": 6088 }, { "epoch": 0.774583386337616, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2051897048950195, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8702467083930969, "num_tokens": 231224634.0, "step": 6089 }, { "epoch": 0.7747105966162066, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3823394775390625, "learning_rate": 1e-05, "loss": 0.5166, "mean_token_accuracy": 0.851704478263855, "num_tokens": 231265497.0, "step": 6090 }, { "epoch": 0.7748378068947971, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1230509281158447, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8638873100280762, "num_tokens": 231301886.0, "step": 6091 }, { "epoch": 0.7749650171733876, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4277515411376953, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8490025997161865, "num_tokens": 231339580.0, "step": 6092 }, { "epoch": 0.7750922274519781, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.063863754272461, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8614722490310669, "num_tokens": 231376176.0, "step": 6093 }, { "epoch": 0.7752194377305687, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5141899585723877, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8611635565757751, "num_tokens": 231409975.0, "step": 6094 }, { "epoch": 0.7753466480091591, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9956259727478027, "learning_rate": 1e-05, "loss": 0.4924, "mean_token_accuracy": 0.8560876250267029, "num_tokens": 231447998.0, "step": 6095 }, { "epoch": 0.7754738582877496, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.665351629257202, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8697837591171265, "num_tokens": 231481810.0, "step": 6096 }, { "epoch": 0.7756010685663401, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.959735631942749, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8693250417709351, "num_tokens": 231521866.0, "step": 6097 }, { "epoch": 0.7757282788449307, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.872298240661621, "learning_rate": 1e-05, "loss": 0.5314, "mean_token_accuracy": 0.8506795167922974, "num_tokens": 231558247.0, "step": 6098 }, { "epoch": 0.7758554891235212, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.063864231109619, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8636595606803894, "num_tokens": 231594183.0, "step": 6099 }, { "epoch": 0.7759826994021117, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.021711826324463, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8609646558761597, "num_tokens": 231628991.0, "step": 6100 }, { "epoch": 0.7761099096807021, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.275587797164917, "learning_rate": 1e-05, "loss": 0.5265, "mean_token_accuracy": 0.847597599029541, "num_tokens": 231669413.0, "step": 6101 }, { "epoch": 0.7762371199592927, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.998196840286255, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8625752329826355, "num_tokens": 231708765.0, "step": 6102 }, { "epoch": 0.7763643302378832, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.444951295852661, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8583581447601318, "num_tokens": 231747577.0, "step": 6103 }, { "epoch": 0.7764915405164737, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.829159736633301, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8541289567947388, "num_tokens": 231788256.0, "step": 6104 }, { "epoch": 0.7766187507950643, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.4769837856292725, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8595187664031982, "num_tokens": 231825709.0, "step": 6105 }, { "epoch": 0.7767459610736548, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6449804306030273, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8643573522567749, "num_tokens": 231862240.0, "step": 6106 }, { "epoch": 0.7768731713522452, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4286253452301025, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8720880150794983, "num_tokens": 231897711.0, "step": 6107 }, { "epoch": 0.7770003816308357, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.490095853805542, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.854210376739502, "num_tokens": 231932383.0, "step": 6108 }, { "epoch": 0.7771275919094263, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.375028133392334, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8547017574310303, "num_tokens": 231960919.0, "step": 6109 }, { "epoch": 0.7772548021880168, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3695123195648193, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8793842792510986, "num_tokens": 231996435.0, "step": 6110 }, { "epoch": 0.7773820124666073, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.301448106765747, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8519288301467896, "num_tokens": 232036146.0, "step": 6111 }, { "epoch": 0.7775092227451978, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3086483478546143, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.8535889983177185, "num_tokens": 232077656.0, "step": 6112 }, { "epoch": 0.7776364330237884, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.252347946166992, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8673948645591736, "num_tokens": 232113370.0, "step": 6113 }, { "epoch": 0.7777636433023788, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2723124027252197, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8667855262756348, "num_tokens": 232155450.0, "step": 6114 }, { "epoch": 0.7778908535809693, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2018134593963623, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8664306998252869, "num_tokens": 232198076.0, "step": 6115 }, { "epoch": 0.7780180638595598, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.268454074859619, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8543763160705566, "num_tokens": 232246472.0, "step": 6116 }, { "epoch": 0.7781452741381504, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1773126125335693, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8654414415359497, "num_tokens": 232280576.0, "step": 6117 }, { "epoch": 0.7782724844167409, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2819252014160156, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8669111132621765, "num_tokens": 232322172.0, "step": 6118 }, { "epoch": 0.7783996946953314, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.119502067565918, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8625175952911377, "num_tokens": 232359317.0, "step": 6119 }, { "epoch": 0.7785269049739219, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.34366774559021, "learning_rate": 1e-05, "loss": 0.5215, "mean_token_accuracy": 0.8456543684005737, "num_tokens": 232395209.0, "step": 6120 }, { "epoch": 0.7786541152525124, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0561585426330566, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8530464172363281, "num_tokens": 232433303.0, "step": 6121 }, { "epoch": 0.7787813255311029, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4311959743499756, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8738785982131958, "num_tokens": 232464465.0, "step": 6122 }, { "epoch": 0.7789085358096934, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9744012355804443, "learning_rate": 1e-05, "loss": 0.4157, "mean_token_accuracy": 0.8773537874221802, "num_tokens": 232499846.0, "step": 6123 }, { "epoch": 0.779035746088284, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6036806106567383, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8728362321853638, "num_tokens": 232528714.0, "step": 6124 }, { "epoch": 0.7791629563668745, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.9292712211608887, "learning_rate": 1e-05, "loss": 0.4969, "mean_token_accuracy": 0.8533840775489807, "num_tokens": 232561155.0, "step": 6125 }, { "epoch": 0.7792901666454649, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.837857961654663, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8756236433982849, "num_tokens": 232595685.0, "step": 6126 }, { "epoch": 0.7794173769240554, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 2.9551405906677246, "learning_rate": 1e-05, "loss": 0.5366, "mean_token_accuracy": 0.8475707769393921, "num_tokens": 232638683.0, "step": 6127 }, { "epoch": 0.779544587202646, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.102193832397461, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8624844551086426, "num_tokens": 232677492.0, "step": 6128 }, { "epoch": 0.7796717974812365, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.221214771270752, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8487292528152466, "num_tokens": 232717997.0, "step": 6129 }, { "epoch": 0.779799007759827, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.184241771697998, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8549899458885193, "num_tokens": 232756784.0, "step": 6130 }, { "epoch": 0.7799262180384176, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5401952266693115, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8593763113021851, "num_tokens": 232795194.0, "step": 6131 }, { "epoch": 0.780053428317008, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.006994724273682, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8651736378669739, "num_tokens": 232826505.0, "step": 6132 }, { "epoch": 0.7801806385955985, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.607780933380127, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8713955879211426, "num_tokens": 232860111.0, "step": 6133 }, { "epoch": 0.780307848874189, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8014206886291504, "learning_rate": 1e-05, "loss": 0.5154, "mean_token_accuracy": 0.8532203435897827, "num_tokens": 232895558.0, "step": 6134 }, { "epoch": 0.7804350591527796, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5121395587921143, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8673068284988403, "num_tokens": 232935737.0, "step": 6135 }, { "epoch": 0.7805622694313701, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001468658447265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6366660594940186, "learning_rate": 1e-05, "loss": 0.537, "mean_token_accuracy": 0.8426914811134338, "num_tokens": 232977917.0, "step": 6136 }, { "epoch": 0.7806894797099606, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.403038501739502, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8748869895935059, "num_tokens": 233016748.0, "step": 6137 }, { "epoch": 0.780816689988551, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5375595092773438, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8641601800918579, "num_tokens": 233052814.0, "step": 6138 }, { "epoch": 0.7809439002671416, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.295032262802124, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.859783947467804, "num_tokens": 233088727.0, "step": 6139 }, { "epoch": 0.7810711105457321, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.498936653137207, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8581598997116089, "num_tokens": 233126402.0, "step": 6140 }, { "epoch": 0.7811983208243226, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2100350856781006, "learning_rate": 1e-05, "loss": 0.523, "mean_token_accuracy": 0.8468585014343262, "num_tokens": 233161204.0, "step": 6141 }, { "epoch": 0.7813255311029131, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.527812957763672, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8617188930511475, "num_tokens": 233196527.0, "step": 6142 }, { "epoch": 0.7814527413815037, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1108360290527344, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8543687462806702, "num_tokens": 233236164.0, "step": 6143 }, { "epoch": 0.7815799516600941, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6190497875213623, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8545529842376709, "num_tokens": 233268762.0, "step": 6144 }, { "epoch": 0.7817071619386846, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0663421154022217, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.8542208671569824, "num_tokens": 233307388.0, "step": 6145 }, { "epoch": 0.7818343722172751, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.747015953063965, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8652060031890869, "num_tokens": 233347902.0, "step": 6146 }, { "epoch": 0.7819615824958657, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.117504119873047, "learning_rate": 1e-05, "loss": 0.5202, "mean_token_accuracy": 0.8450876474380493, "num_tokens": 233378959.0, "step": 6147 }, { "epoch": 0.7820887927744562, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.848104476928711, "learning_rate": 1e-05, "loss": 0.5046, "mean_token_accuracy": 0.8559784293174744, "num_tokens": 233420580.0, "step": 6148 }, { "epoch": 0.7822160030530467, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2192697525024414, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8685564994812012, "num_tokens": 233451628.0, "step": 6149 }, { "epoch": 0.7823432133316371, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8451426029205322, "learning_rate": 1e-05, "loss": 0.5113, "mean_token_accuracy": 0.8548490405082703, "num_tokens": 233493994.0, "step": 6150 }, { "epoch": 0.7824704236102277, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3376107215881348, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8657756447792053, "num_tokens": 233534156.0, "step": 6151 }, { "epoch": 0.7825976338888182, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7169055938720703, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.860145628452301, "num_tokens": 233573179.0, "step": 6152 }, { "epoch": 0.7827248441674087, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4095945358276367, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8683658242225647, "num_tokens": 233606424.0, "step": 6153 }, { "epoch": 0.7828520544459993, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.513960599899292, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8556956052780151, "num_tokens": 233649155.0, "step": 6154 }, { "epoch": 0.7829792647245898, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4267523288726807, "learning_rate": 1e-05, "loss": 0.5128, "mean_token_accuracy": 0.8468191623687744, "num_tokens": 233677163.0, "step": 6155 }, { "epoch": 0.7831064750031802, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.327641487121582, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8494109511375427, "num_tokens": 233713120.0, "step": 6156 }, { "epoch": 0.7832336852817707, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.404130458831787, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8580304384231567, "num_tokens": 233751160.0, "step": 6157 }, { "epoch": 0.7833608955603613, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.19124436378479, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8593453168869019, "num_tokens": 233786474.0, "step": 6158 }, { "epoch": 0.7834881058389518, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.430968999862671, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8666794300079346, "num_tokens": 233818256.0, "step": 6159 }, { "epoch": 0.7836153161175423, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1012182235717773, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8514604568481445, "num_tokens": 233856519.0, "step": 6160 }, { "epoch": 0.7837425263961328, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.476508378982544, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8611469864845276, "num_tokens": 233896132.0, "step": 6161 }, { "epoch": 0.7838697366747234, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.024808883666992, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8611266613006592, "num_tokens": 233931668.0, "step": 6162 }, { "epoch": 0.7839969469533138, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.603391170501709, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8702532052993774, "num_tokens": 233965821.0, "step": 6163 }, { "epoch": 0.7841241572319043, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.970944404602051, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.862450897693634, "num_tokens": 234003488.0, "step": 6164 }, { "epoch": 0.7842513675104948, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00014781951904296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7765626907348633, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8547921180725098, "num_tokens": 234044270.0, "step": 6165 }, { "epoch": 0.7843785777890854, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 2.998500347137451, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8798573613166809, "num_tokens": 234081754.0, "step": 6166 }, { "epoch": 0.7845057880676759, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9769740104675293, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8768597841262817, "num_tokens": 234118961.0, "step": 6167 }, { "epoch": 0.7846329983462664, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.18782114982605, "learning_rate": 1e-05, "loss": 0.5279, "mean_token_accuracy": 0.8443785905838013, "num_tokens": 234159084.0, "step": 6168 }, { "epoch": 0.7847602086248568, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0278801918029785, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8651882410049438, "num_tokens": 234201296.0, "step": 6169 }, { "epoch": 0.7848874189034474, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4105217456817627, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8672478199005127, "num_tokens": 234240661.0, "step": 6170 }, { "epoch": 0.7850146291820379, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.89630126953125, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8506529927253723, "num_tokens": 234277192.0, "step": 6171 }, { "epoch": 0.7851418394606284, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.48836612701416, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8584727644920349, "num_tokens": 234321527.0, "step": 6172 }, { "epoch": 0.785269049739219, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.698456048965454, "learning_rate": 1e-05, "loss": 0.5161, "mean_token_accuracy": 0.8492351174354553, "num_tokens": 234355578.0, "step": 6173 }, { "epoch": 0.7853962600178095, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.4579036235809326, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8686637878417969, "num_tokens": 234395240.0, "step": 6174 }, { "epoch": 0.7855234702963999, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5145487785339355, "learning_rate": 1e-05, "loss": 0.5353, "mean_token_accuracy": 0.8466148376464844, "num_tokens": 234436914.0, "step": 6175 }, { "epoch": 0.7856506805749904, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3931665420532227, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8603955507278442, "num_tokens": 234477811.0, "step": 6176 }, { "epoch": 0.785777890853581, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.381950616836548, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.866061806678772, "num_tokens": 234515489.0, "step": 6177 }, { "epoch": 0.7859051011321715, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.356637716293335, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8575695753097534, "num_tokens": 234544846.0, "step": 6178 }, { "epoch": 0.786032311410762, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2984189987182617, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8710345029830933, "num_tokens": 234581497.0, "step": 6179 }, { "epoch": 0.7861595216893525, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2995998859405518, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.853706419467926, "num_tokens": 234621837.0, "step": 6180 }, { "epoch": 0.786286731967943, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.250373125076294, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.865842342376709, "num_tokens": 234655049.0, "step": 6181 }, { "epoch": 0.7864139422465335, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2715675830841064, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.854087233543396, "num_tokens": 234689177.0, "step": 6182 }, { "epoch": 0.786541152525124, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.216573715209961, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8735533952713013, "num_tokens": 234722744.0, "step": 6183 }, { "epoch": 0.7866683628037145, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.240170478820801, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8653855323791504, "num_tokens": 234762998.0, "step": 6184 }, { "epoch": 0.7867955730823051, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2119009494781494, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8455246686935425, "num_tokens": 234799283.0, "step": 6185 }, { "epoch": 0.7869227833608956, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2312381267547607, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8513872623443604, "num_tokens": 234838019.0, "step": 6186 }, { "epoch": 0.787049993639486, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.211703300476074, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.85381019115448, "num_tokens": 234874449.0, "step": 6187 }, { "epoch": 0.7871772039180766, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2079553604125977, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8617838621139526, "num_tokens": 234913868.0, "step": 6188 }, { "epoch": 0.7873044141966671, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1983749866485596, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8793219327926636, "num_tokens": 234949060.0, "step": 6189 }, { "epoch": 0.7874316244752576, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1947648525238037, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8593725562095642, "num_tokens": 234987732.0, "step": 6190 }, { "epoch": 0.7875588347538481, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.197376251220703, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.870121955871582, "num_tokens": 235023030.0, "step": 6191 }, { "epoch": 0.7876860450324387, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2043888568878174, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8645177483558655, "num_tokens": 235061560.0, "step": 6192 }, { "epoch": 0.7878132553110291, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.188471794128418, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8662484288215637, "num_tokens": 235098826.0, "step": 6193 }, { "epoch": 0.7879404655896196, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1989214420318604, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8576852083206177, "num_tokens": 235142116.0, "step": 6194 }, { "epoch": 0.7880676758682101, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.181396484375, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.856694221496582, "num_tokens": 235182920.0, "step": 6195 }, { "epoch": 0.7881948861468007, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.218672037124634, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.857947051525116, "num_tokens": 235214669.0, "step": 6196 }, { "epoch": 0.7883220964253912, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1615347862243652, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8634576797485352, "num_tokens": 235251723.0, "step": 6197 }, { "epoch": 0.7884493067039817, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.228219985961914, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8672084808349609, "num_tokens": 235285931.0, "step": 6198 }, { "epoch": 0.7885765169825721, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.135610342025757, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8692457675933838, "num_tokens": 235324571.0, "step": 6199 }, { "epoch": 0.7887037272611627, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2627556324005127, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8720762729644775, "num_tokens": 235359348.0, "step": 6200 }, { "epoch": 0.7888309375397532, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.094675064086914, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8733536601066589, "num_tokens": 235397998.0, "step": 6201 }, { "epoch": 0.7889581478183437, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3314342498779297, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8594001531600952, "num_tokens": 235436991.0, "step": 6202 }, { "epoch": 0.7890853580969343, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0103795528411865, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8785978555679321, "num_tokens": 235478129.0, "step": 6203 }, { "epoch": 0.7892125683755248, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4688875675201416, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8674790859222412, "num_tokens": 235514630.0, "step": 6204 }, { "epoch": 0.7893397786541152, "ewc_loss": 0.0546875, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 3.981590270996094e-05, "grad_norm": 2.919037342071533, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8711982369422913, "num_tokens": 235550212.0, "step": 6205 }, { "epoch": 0.7894669889327057, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.695709466934204, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8712029457092285, "num_tokens": 235592602.0, "step": 6206 }, { "epoch": 0.7895941992112963, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.8328940868377686, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8732616901397705, "num_tokens": 235634199.0, "step": 6207 }, { "epoch": 0.7897214094898868, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0971903800964355, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8691830635070801, "num_tokens": 235670086.0, "step": 6208 }, { "epoch": 0.7898486197684773, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000148773193359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.0180094242095947, "learning_rate": 1e-05, "loss": 0.5383, "mean_token_accuracy": 0.8456713557243347, "num_tokens": 235709749.0, "step": 6209 }, { "epoch": 0.7899758300470678, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.460175514221191, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8641566038131714, "num_tokens": 235750574.0, "step": 6210 }, { "epoch": 0.7901030403256584, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.597325086593628, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8661047220230103, "num_tokens": 235785234.0, "step": 6211 }, { "epoch": 0.7902302506042488, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.312279224395752, "learning_rate": 1e-05, "loss": 0.5306, "mean_token_accuracy": 0.8488726019859314, "num_tokens": 235822931.0, "step": 6212 }, { "epoch": 0.7903574608828393, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.856665849685669, "learning_rate": 1e-05, "loss": 0.516, "mean_token_accuracy": 0.8504285216331482, "num_tokens": 235863022.0, "step": 6213 }, { "epoch": 0.7904846711614298, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9957752227783203, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8624076843261719, "num_tokens": 235893275.0, "step": 6214 }, { "epoch": 0.7906118814400204, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00014972686767578125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.72525954246521, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.881531834602356, "num_tokens": 235933908.0, "step": 6215 }, { "epoch": 0.7907390917186109, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7994296550750732, "learning_rate": 1e-05, "loss": 0.5111, "mean_token_accuracy": 0.8530807495117188, "num_tokens": 235975296.0, "step": 6216 }, { "epoch": 0.7908663019972014, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.547067403793335, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8710445165634155, "num_tokens": 236012872.0, "step": 6217 }, { "epoch": 0.7909935122757918, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.68186354637146, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8642115592956543, "num_tokens": 236055762.0, "step": 6218 }, { "epoch": 0.7911207225543824, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3819594383239746, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.866573691368103, "num_tokens": 236099968.0, "step": 6219 }, { "epoch": 0.7912479328329729, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.660881996154785, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8526265621185303, "num_tokens": 236134000.0, "step": 6220 }, { "epoch": 0.7913751431115634, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2470157146453857, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8672112226486206, "num_tokens": 236174427.0, "step": 6221 }, { "epoch": 0.791502353390154, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.720393180847168, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.8535000681877136, "num_tokens": 236205643.0, "step": 6222 }, { "epoch": 0.7916295636687445, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1634464263916016, "learning_rate": 1e-05, "loss": 0.5246, "mean_token_accuracy": 0.8452000617980957, "num_tokens": 236252232.0, "step": 6223 }, { "epoch": 0.7917567739473349, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8421669006347656, "learning_rate": 1e-05, "loss": 0.4924, "mean_token_accuracy": 0.8577637076377869, "num_tokens": 236293694.0, "step": 6224 }, { "epoch": 0.7918839842259254, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2259509563446045, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8670580387115479, "num_tokens": 236335129.0, "step": 6225 }, { "epoch": 0.792011194504516, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.9269442558288574, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8735879063606262, "num_tokens": 236370294.0, "step": 6226 }, { "epoch": 0.7921384047831065, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.286691188812256, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8707125186920166, "num_tokens": 236410728.0, "step": 6227 }, { "epoch": 0.792265615061697, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.937026023864746, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8616973757743835, "num_tokens": 236453053.0, "step": 6228 }, { "epoch": 0.7923928253402875, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5110387802124023, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8704867362976074, "num_tokens": 236491746.0, "step": 6229 }, { "epoch": 0.792520035618878, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.7303013801574707, "learning_rate": 1e-05, "loss": 0.5123, "mean_token_accuracy": 0.8481782674789429, "num_tokens": 236530938.0, "step": 6230 }, { "epoch": 0.7926472458974685, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4577932357788086, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.8871678709983826, "num_tokens": 236567765.0, "step": 6231 }, { "epoch": 0.792774456176059, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.5560407638549805, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8609517812728882, "num_tokens": 236596522.0, "step": 6232 }, { "epoch": 0.7929016664546495, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4177823066711426, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8549189567565918, "num_tokens": 236637493.0, "step": 6233 }, { "epoch": 0.7930288767332401, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.401601552963257, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8593251705169678, "num_tokens": 236668405.0, "step": 6234 }, { "epoch": 0.7931560870118306, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3751909732818604, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.866268515586853, "num_tokens": 236703105.0, "step": 6235 }, { "epoch": 0.793283297290421, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2728965282440186, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8637382388114929, "num_tokens": 236747017.0, "step": 6236 }, { "epoch": 0.7934105075690115, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.342231035232544, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8687534332275391, "num_tokens": 236785659.0, "step": 6237 }, { "epoch": 0.7935377178476021, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.2114076614379883, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8536239266395569, "num_tokens": 236819603.0, "step": 6238 }, { "epoch": 0.7936649281261926, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.322277069091797, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8673335909843445, "num_tokens": 236859440.0, "step": 6239 }, { "epoch": 0.7937921384047831, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1694247722625732, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8625907301902771, "num_tokens": 236892138.0, "step": 6240 }, { "epoch": 0.7939193486833737, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.3238325119018555, "learning_rate": 1e-05, "loss": 0.5194, "mean_token_accuracy": 0.8449811935424805, "num_tokens": 236930559.0, "step": 6241 }, { "epoch": 0.7940465589619641, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.117955207824707, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8596255779266357, "num_tokens": 236969275.0, "step": 6242 }, { "epoch": 0.7941737692405546, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.352313995361328, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8697961568832397, "num_tokens": 237007662.0, "step": 6243 }, { "epoch": 0.7943009795191451, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0578980445861816, "learning_rate": 1e-05, "loss": 0.5097, "mean_token_accuracy": 0.8455988168716431, "num_tokens": 237046553.0, "step": 6244 }, { "epoch": 0.7944281897977357, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.441298246383667, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8601844310760498, "num_tokens": 237082309.0, "step": 6245 }, { "epoch": 0.7945554000763262, "ewc_loss": 0.05517578125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.00543212890625e-05, "grad_norm": 2.9871726036071777, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8698359131813049, "num_tokens": 237117997.0, "step": 6246 }, { "epoch": 0.7946826103549167, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.591665506362915, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.8476909399032593, "num_tokens": 237159169.0, "step": 6247 }, { "epoch": 0.7948098206335071, "ewc_loss": 0.055419921875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.029273986816406e-05, "grad_norm": 2.9227702617645264, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8591660857200623, "num_tokens": 237200863.0, "step": 6248 }, { "epoch": 0.7949370309120977, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.845621347427368, "learning_rate": 1e-05, "loss": 0.5191, "mean_token_accuracy": 0.8489972949028015, "num_tokens": 237237318.0, "step": 6249 }, { "epoch": 0.7950642411906882, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.9658453464508057, "learning_rate": 1e-05, "loss": 0.5179, "mean_token_accuracy": 0.8439462780952454, "num_tokens": 237278724.0, "step": 6250 }, { "epoch": 0.7951914514692787, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.126250743865967, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.866204023361206, "num_tokens": 237314190.0, "step": 6251 }, { "epoch": 0.7953186617478692, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.250474214553833, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8732497692108154, "num_tokens": 237348560.0, "step": 6252 }, { "epoch": 0.7954458720264598, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.179749011993408, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.8630165457725525, "num_tokens": 237388421.0, "step": 6253 }, { "epoch": 0.7955730823050502, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.572951316833496, "learning_rate": 1e-05, "loss": 0.5304, "mean_token_accuracy": 0.8465451598167419, "num_tokens": 237424517.0, "step": 6254 }, { "epoch": 0.7957002925836407, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.954446792602539, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8652499914169312, "num_tokens": 237466195.0, "step": 6255 }, { "epoch": 0.7958275028622313, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.607973575592041, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8655766248703003, "num_tokens": 237500189.0, "step": 6256 }, { "epoch": 0.7959547131408218, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.750623941421509, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8563117384910583, "num_tokens": 237534186.0, "step": 6257 }, { "epoch": 0.7960819234194123, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5151679515838623, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.858191967010498, "num_tokens": 237572382.0, "step": 6258 }, { "epoch": 0.7962091336980028, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.574632167816162, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8587309122085571, "num_tokens": 237614300.0, "step": 6259 }, { "epoch": 0.7963363439765934, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.413620948791504, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8637498617172241, "num_tokens": 237650952.0, "step": 6260 }, { "epoch": 0.7964635542551838, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4534151554107666, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.8529046773910522, "num_tokens": 237689901.0, "step": 6261 }, { "epoch": 0.7965907645337743, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.327376365661621, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8468714952468872, "num_tokens": 237724889.0, "step": 6262 }, { "epoch": 0.7967179748123648, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3957130908966064, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8660218119621277, "num_tokens": 237758853.0, "step": 6263 }, { "epoch": 0.7968451850909554, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2401392459869385, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8593356609344482, "num_tokens": 237795646.0, "step": 6264 }, { "epoch": 0.7969723953695459, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.384963035583496, "learning_rate": 1e-05, "loss": 0.5142, "mean_token_accuracy": 0.8495805859565735, "num_tokens": 237832229.0, "step": 6265 }, { "epoch": 0.7970996056481364, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.155707359313965, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8486987352371216, "num_tokens": 237870459.0, "step": 6266 }, { "epoch": 0.7972268159267268, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4174482822418213, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8569682836532593, "num_tokens": 237912367.0, "step": 6267 }, { "epoch": 0.7973540262053174, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.074075222015381, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.860468327999115, "num_tokens": 237950847.0, "step": 6268 }, { "epoch": 0.7974812364839079, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5313150882720947, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8683525323867798, "num_tokens": 237985325.0, "step": 6269 }, { "epoch": 0.7976084467624984, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0106253623962402, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8516355752944946, "num_tokens": 238021710.0, "step": 6270 }, { "epoch": 0.797735657041089, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.699240207672119, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8661602139472961, "num_tokens": 238053024.0, "step": 6271 }, { "epoch": 0.7978628673196795, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.003361463546753, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8679593801498413, "num_tokens": 238089732.0, "step": 6272 }, { "epoch": 0.7979900775982699, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.880812406539917, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.861507773399353, "num_tokens": 238127583.0, "step": 6273 }, { "epoch": 0.7981172878768604, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.093122959136963, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8603061437606812, "num_tokens": 238166328.0, "step": 6274 }, { "epoch": 0.798244498155451, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.021026611328125, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8506361842155457, "num_tokens": 238206111.0, "step": 6275 }, { "epoch": 0.7983717084340415, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3326668739318848, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8626863360404968, "num_tokens": 238243793.0, "step": 6276 }, { "epoch": 0.798498918712632, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.940471649169922, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8766714930534363, "num_tokens": 238284005.0, "step": 6277 }, { "epoch": 0.7986261289912225, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.4887285232543945, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.855793833732605, "num_tokens": 238321180.0, "step": 6278 }, { "epoch": 0.798753339269813, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.713566541671753, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8669875264167786, "num_tokens": 238364383.0, "step": 6279 }, { "epoch": 0.7988805495484035, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4841654300689697, "learning_rate": 1e-05, "loss": 0.5367, "mean_token_accuracy": 0.8479584455490112, "num_tokens": 238405547.0, "step": 6280 }, { "epoch": 0.799007759826994, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.520885944366455, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8627972602844238, "num_tokens": 238436810.0, "step": 6281 }, { "epoch": 0.7991349701055845, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.441655158996582, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8531757593154907, "num_tokens": 238470062.0, "step": 6282 }, { "epoch": 0.7992621803841751, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3548190593719482, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8783082365989685, "num_tokens": 238502680.0, "step": 6283 }, { "epoch": 0.7993893906627656, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3968985080718994, "learning_rate": 1e-05, "loss": 0.4982, "mean_token_accuracy": 0.8527384996414185, "num_tokens": 238534355.0, "step": 6284 }, { "epoch": 0.799516600941356, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2442708015441895, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8586244583129883, "num_tokens": 238572225.0, "step": 6285 }, { "epoch": 0.7996438112199465, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.371671199798584, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8659703731536865, "num_tokens": 238608313.0, "step": 6286 }, { "epoch": 0.7997710214985371, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.165651321411133, "learning_rate": 1e-05, "loss": 0.5171, "mean_token_accuracy": 0.8455979824066162, "num_tokens": 238652973.0, "step": 6287 }, { "epoch": 0.7998982317771276, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3897745609283447, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8576592803001404, "num_tokens": 238688490.0, "step": 6288 }, { "epoch": 0.8000254420557181, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0944371223449707, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8694132566452026, "num_tokens": 238728641.0, "step": 6289 }, { "epoch": 0.8001526523343087, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.459747076034546, "learning_rate": 1e-05, "loss": 0.5114, "mean_token_accuracy": 0.8474479913711548, "num_tokens": 238764134.0, "step": 6290 }, { "epoch": 0.8002798626128991, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0316598415374756, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8662608861923218, "num_tokens": 238803705.0, "step": 6291 }, { "epoch": 0.8004070728914896, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5803256034851074, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.853881299495697, "num_tokens": 238836816.0, "step": 6292 }, { "epoch": 0.8005342831700801, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.9725756645202637, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8688780665397644, "num_tokens": 238870006.0, "step": 6293 }, { "epoch": 0.8006614934486707, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7589712142944336, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8598783016204834, "num_tokens": 238907018.0, "step": 6294 }, { "epoch": 0.8007887037272612, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.9972171783447266, "learning_rate": 1e-05, "loss": 0.5198, "mean_token_accuracy": 0.8467342853546143, "num_tokens": 238943319.0, "step": 6295 }, { "epoch": 0.8009159140058517, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.961564779281616, "learning_rate": 1e-05, "loss": 0.5292, "mean_token_accuracy": 0.8427895903587341, "num_tokens": 238978546.0, "step": 6296 }, { "epoch": 0.8010431242844421, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.161985397338867, "learning_rate": 1e-05, "loss": 0.5071, "mean_token_accuracy": 0.856610119342804, "num_tokens": 239014785.0, "step": 6297 }, { "epoch": 0.8011703345630327, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0583720207214355, "learning_rate": 1e-05, "loss": 0.5425, "mean_token_accuracy": 0.8409576416015625, "num_tokens": 239048212.0, "step": 6298 }, { "epoch": 0.8012975448416232, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3917155265808105, "learning_rate": 1e-05, "loss": 0.5225, "mean_token_accuracy": 0.8492685556411743, "num_tokens": 239086946.0, "step": 6299 }, { "epoch": 0.8014247551202137, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.936077356338501, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8768787384033203, "num_tokens": 239125100.0, "step": 6300 }, { "epoch": 0.8015519653988042, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001506805419921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.488460063934326, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8527660369873047, "num_tokens": 239163118.0, "step": 6301 }, { "epoch": 0.8016791756773948, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.756535530090332, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8665276765823364, "num_tokens": 239196914.0, "step": 6302 }, { "epoch": 0.8018063859559852, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.476970672607422, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8616646528244019, "num_tokens": 239229609.0, "step": 6303 }, { "epoch": 0.8019335962345757, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5586345195770264, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8797029256820679, "num_tokens": 239269829.0, "step": 6304 }, { "epoch": 0.8020608065131662, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4029641151428223, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8760325312614441, "num_tokens": 239308563.0, "step": 6305 }, { "epoch": 0.8021880167917568, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4286751747131348, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8633941411972046, "num_tokens": 239348620.0, "step": 6306 }, { "epoch": 0.8023152270703473, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.351274013519287, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8602398633956909, "num_tokens": 239380706.0, "step": 6307 }, { "epoch": 0.8024424373489378, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.350358724594116, "learning_rate": 1e-05, "loss": 0.5083, "mean_token_accuracy": 0.8500398397445679, "num_tokens": 239417365.0, "step": 6308 }, { "epoch": 0.8025696476275284, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.292226552963257, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8579651713371277, "num_tokens": 239449926.0, "step": 6309 }, { "epoch": 0.8026968579061188, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.291102886199951, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8584779500961304, "num_tokens": 239485134.0, "step": 6310 }, { "epoch": 0.8028240681847093, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.239994525909424, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8690844178199768, "num_tokens": 239521353.0, "step": 6311 }, { "epoch": 0.8029512784632998, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2812578678131104, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8580944538116455, "num_tokens": 239554902.0, "step": 6312 }, { "epoch": 0.8030784887418904, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1886496543884277, "learning_rate": 1e-05, "loss": 0.5071, "mean_token_accuracy": 0.8498961925506592, "num_tokens": 239597436.0, "step": 6313 }, { "epoch": 0.8032056990204809, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2885003089904785, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8654370307922363, "num_tokens": 239635543.0, "step": 6314 }, { "epoch": 0.8033329092990714, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.139204263687134, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8727916479110718, "num_tokens": 239673956.0, "step": 6315 }, { "epoch": 0.8034601195776618, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.323854684829712, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.857263445854187, "num_tokens": 239712556.0, "step": 6316 }, { "epoch": 0.8035873298562524, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.087219715118408, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8748204708099365, "num_tokens": 239747034.0, "step": 6317 }, { "epoch": 0.8037145401348429, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4004533290863037, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8642849326133728, "num_tokens": 239787054.0, "step": 6318 }, { "epoch": 0.8038417504134334, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.1048338413238525, "learning_rate": 1e-05, "loss": 0.5184, "mean_token_accuracy": 0.8467310667037964, "num_tokens": 239824796.0, "step": 6319 }, { "epoch": 0.803968960692024, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4503509998321533, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.8533206582069397, "num_tokens": 239862526.0, "step": 6320 }, { "epoch": 0.8040961709706145, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.9975366592407227, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8574989438056946, "num_tokens": 239897409.0, "step": 6321 }, { "epoch": 0.8042233812492049, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.575624704360962, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8678414821624756, "num_tokens": 239935202.0, "step": 6322 }, { "epoch": 0.8043505915277954, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.9481866359710693, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8634663820266724, "num_tokens": 239967603.0, "step": 6323 }, { "epoch": 0.804477801806386, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7914350032806396, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.8533629179000854, "num_tokens": 240000757.0, "step": 6324 }, { "epoch": 0.8046050120849765, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 2.96370530128479, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8540511131286621, "num_tokens": 240039747.0, "step": 6325 }, { "epoch": 0.804732222363567, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.030680179595947, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8565560579299927, "num_tokens": 240078951.0, "step": 6326 }, { "epoch": 0.8048594326421575, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.163740873336792, "learning_rate": 1e-05, "loss": 0.5373, "mean_token_accuracy": 0.8438655138015747, "num_tokens": 240120097.0, "step": 6327 }, { "epoch": 0.804986642920748, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.138491630554199, "learning_rate": 1e-05, "loss": 0.5637, "mean_token_accuracy": 0.8416743278503418, "num_tokens": 240166510.0, "step": 6328 }, { "epoch": 0.8051138531993385, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.4748451709747314, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8604036569595337, "num_tokens": 240207961.0, "step": 6329 }, { "epoch": 0.805241063477929, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.991729259490967, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8639832139015198, "num_tokens": 240243257.0, "step": 6330 }, { "epoch": 0.8053682737565195, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.551690101623535, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8659956455230713, "num_tokens": 240279844.0, "step": 6331 }, { "epoch": 0.8054954840351101, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.7977895736694336, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8644493818283081, "num_tokens": 240316836.0, "step": 6332 }, { "epoch": 0.8056226943137006, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5007903575897217, "learning_rate": 1e-05, "loss": 0.5346, "mean_token_accuracy": 0.8440937995910645, "num_tokens": 240357794.0, "step": 6333 }, { "epoch": 0.805749904592291, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6251041889190674, "learning_rate": 1e-05, "loss": 0.5265, "mean_token_accuracy": 0.8423633575439453, "num_tokens": 240397929.0, "step": 6334 }, { "epoch": 0.8058771148708815, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.411768674850464, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8725229501724243, "num_tokens": 240434319.0, "step": 6335 }, { "epoch": 0.8060043251494721, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.499542713165283, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8600086569786072, "num_tokens": 240476755.0, "step": 6336 }, { "epoch": 0.8061315354280626, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.318091869354248, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8672052025794983, "num_tokens": 240513042.0, "step": 6337 }, { "epoch": 0.8062587457066531, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.443153142929077, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8640804290771484, "num_tokens": 240554805.0, "step": 6338 }, { "epoch": 0.8063859559852437, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2210402488708496, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8752343654632568, "num_tokens": 240593196.0, "step": 6339 }, { "epoch": 0.8065131662638341, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4471564292907715, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8640110492706299, "num_tokens": 240632640.0, "step": 6340 }, { "epoch": 0.8066403765424246, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1495258808135986, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8655153512954712, "num_tokens": 240672868.0, "step": 6341 }, { "epoch": 0.8067675868210151, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.502324342727661, "learning_rate": 1e-05, "loss": 0.5191, "mean_token_accuracy": 0.8468743562698364, "num_tokens": 240713499.0, "step": 6342 }, { "epoch": 0.8068947970996057, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0540080070495605, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8712883591651917, "num_tokens": 240759684.0, "step": 6343 }, { "epoch": 0.8070220073781962, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.640047550201416, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8520891666412354, "num_tokens": 240799439.0, "step": 6344 }, { "epoch": 0.8071492176567867, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.033355474472046, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8545281887054443, "num_tokens": 240845569.0, "step": 6345 }, { "epoch": 0.8072764279353771, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.796848773956299, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.875685453414917, "num_tokens": 240879721.0, "step": 6346 }, { "epoch": 0.8074036382139677, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.09207820892334, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8632623553276062, "num_tokens": 240918433.0, "step": 6347 }, { "epoch": 0.8075308484925582, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.93335223197937, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.855588436126709, "num_tokens": 240949650.0, "step": 6348 }, { "epoch": 0.8076580587711487, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2612533569335938, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8683207035064697, "num_tokens": 240984803.0, "step": 6349 }, { "epoch": 0.8077852690497392, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8895187377929688, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8576184511184692, "num_tokens": 241023173.0, "step": 6350 }, { "epoch": 0.8079124793283298, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.396772623062134, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8655175566673279, "num_tokens": 241065235.0, "step": 6351 }, { "epoch": 0.8080396896069202, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.725055456161499, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8618274927139282, "num_tokens": 241106283.0, "step": 6352 }, { "epoch": 0.8081668998855107, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.438424825668335, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8505995273590088, "num_tokens": 241145544.0, "step": 6353 }, { "epoch": 0.8082941101641012, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5242068767547607, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8539509773254395, "num_tokens": 241183888.0, "step": 6354 }, { "epoch": 0.8084213204426918, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.413801908493042, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8770766258239746, "num_tokens": 241224650.0, "step": 6355 }, { "epoch": 0.8085485307212823, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3433618545532227, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8597389459609985, "num_tokens": 241263907.0, "step": 6356 }, { "epoch": 0.8086757409998728, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4079337120056152, "learning_rate": 1e-05, "loss": 0.5299, "mean_token_accuracy": 0.8396869897842407, "num_tokens": 241305009.0, "step": 6357 }, { "epoch": 0.8088029512784632, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2134060859680176, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8680830597877502, "num_tokens": 241344283.0, "step": 6358 }, { "epoch": 0.8089301615570538, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.407855749130249, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8627651333808899, "num_tokens": 241380243.0, "step": 6359 }, { "epoch": 0.8090573718356443, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.127220630645752, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8592547178268433, "num_tokens": 241414373.0, "step": 6360 }, { "epoch": 0.8091845821142348, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4348342418670654, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8664904832839966, "num_tokens": 241456393.0, "step": 6361 }, { "epoch": 0.8093117923928254, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.0547451972961426, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8505958318710327, "num_tokens": 241496086.0, "step": 6362 }, { "epoch": 0.8094390026714159, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.521080493927002, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8650752305984497, "num_tokens": 241537814.0, "step": 6363 }, { "epoch": 0.8095662129500064, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0029690265655518, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8547887206077576, "num_tokens": 241576762.0, "step": 6364 }, { "epoch": 0.8096934232285968, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.661940336227417, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8732329607009888, "num_tokens": 241614667.0, "step": 6365 }, { "epoch": 0.8098206335071874, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.983609199523926, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8762324452400208, "num_tokens": 241647216.0, "step": 6366 }, { "epoch": 0.8099478437857779, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8435561656951904, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8687706589698792, "num_tokens": 241684278.0, "step": 6367 }, { "epoch": 0.8100750540643684, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.0632686614990234, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8764311671257019, "num_tokens": 241723326.0, "step": 6368 }, { "epoch": 0.8102022643429589, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.976494550704956, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8657701015472412, "num_tokens": 241765251.0, "step": 6369 }, { "epoch": 0.8103294746215495, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.2653274536132812, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8708130121231079, "num_tokens": 241801862.0, "step": 6370 }, { "epoch": 0.8104566849001399, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.954802989959717, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8610297441482544, "num_tokens": 241841448.0, "step": 6371 }, { "epoch": 0.8105838951787304, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.413771867752075, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8585615754127502, "num_tokens": 241883151.0, "step": 6372 }, { "epoch": 0.810711105457321, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8029630184173584, "learning_rate": 1e-05, "loss": 0.5148, "mean_token_accuracy": 0.8528842926025391, "num_tokens": 241924643.0, "step": 6373 }, { "epoch": 0.8108383157359115, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.454923629760742, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.8568048477172852, "num_tokens": 241960361.0, "step": 6374 }, { "epoch": 0.810965526014502, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.689077377319336, "learning_rate": 1e-05, "loss": 0.5282, "mean_token_accuracy": 0.8459228873252869, "num_tokens": 241999616.0, "step": 6375 }, { "epoch": 0.8110927362930925, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4202489852905273, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8578579425811768, "num_tokens": 242035824.0, "step": 6376 }, { "epoch": 0.811219946571683, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.479893207550049, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8647816181182861, "num_tokens": 242072680.0, "step": 6377 }, { "epoch": 0.8113471568502735, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3547749519348145, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8691972494125366, "num_tokens": 242111135.0, "step": 6378 }, { "epoch": 0.811474367128864, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.374105453491211, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8681300282478333, "num_tokens": 242148993.0, "step": 6379 }, { "epoch": 0.8116015774074545, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3098502159118652, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8577091693878174, "num_tokens": 242186230.0, "step": 6380 }, { "epoch": 0.8117287876860451, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.315157651901245, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8648355007171631, "num_tokens": 242222341.0, "step": 6381 }, { "epoch": 0.8118559979646356, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.266711711883545, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.8527457118034363, "num_tokens": 242255408.0, "step": 6382 }, { "epoch": 0.811983208243226, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2953624725341797, "learning_rate": 1e-05, "loss": 0.5429, "mean_token_accuracy": 0.8404057621955872, "num_tokens": 242294664.0, "step": 6383 }, { "epoch": 0.8121104185218165, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2037978172302246, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8794859051704407, "num_tokens": 242328982.0, "step": 6384 }, { "epoch": 0.8122376288004071, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.299984931945801, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8511255979537964, "num_tokens": 242369542.0, "step": 6385 }, { "epoch": 0.8123648390789976, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.163797378540039, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8552899956703186, "num_tokens": 242406699.0, "step": 6386 }, { "epoch": 0.8124920493575881, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3252127170562744, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8534823656082153, "num_tokens": 242448057.0, "step": 6387 }, { "epoch": 0.8126192596361786, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.108769655227661, "learning_rate": 1e-05, "loss": 0.5102, "mean_token_accuracy": 0.8483928442001343, "num_tokens": 242489792.0, "step": 6388 }, { "epoch": 0.8127464699147691, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.37605357170105, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8560665845870972, "num_tokens": 242535458.0, "step": 6389 }, { "epoch": 0.8128736801933596, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0486600399017334, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8622034192085266, "num_tokens": 242577380.0, "step": 6390 }, { "epoch": 0.8130008904719501, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4805727005004883, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8567442893981934, "num_tokens": 242614353.0, "step": 6391 }, { "epoch": 0.8131281007505406, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.982037305831909, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8694844841957092, "num_tokens": 242648444.0, "step": 6392 }, { "epoch": 0.8132553110291312, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6499998569488525, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8587380647659302, "num_tokens": 242687658.0, "step": 6393 }, { "epoch": 0.8133825213077217, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.926504135131836, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8618672490119934, "num_tokens": 242726097.0, "step": 6394 }, { "epoch": 0.8135097315863121, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 4.663563251495361, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.85446697473526, "num_tokens": 242769229.0, "step": 6395 }, { "epoch": 0.8136369418649027, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.4390907287597656, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8578426837921143, "num_tokens": 242805815.0, "step": 6396 }, { "epoch": 0.8137641521434932, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.098446846008301, "learning_rate": 1e-05, "loss": 0.53, "mean_token_accuracy": 0.8457024693489075, "num_tokens": 242835929.0, "step": 6397 }, { "epoch": 0.8138913624220837, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.503826141357422, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8705112338066101, "num_tokens": 242875295.0, "step": 6398 }, { "epoch": 0.8140185727006742, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.061981678009033, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.858410656452179, "num_tokens": 242911825.0, "step": 6399 }, { "epoch": 0.8141457829792648, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.557478904724121, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8628618121147156, "num_tokens": 242950568.0, "step": 6400 }, { "epoch": 0.8142729932578552, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9363720417022705, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8676089644432068, "num_tokens": 242981223.0, "step": 6401 }, { "epoch": 0.8144002035364457, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.516007423400879, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8800399899482727, "num_tokens": 243016261.0, "step": 6402 }, { "epoch": 0.8145274138150362, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8185362815856934, "learning_rate": 1e-05, "loss": 0.5304, "mean_token_accuracy": 0.8471471667289734, "num_tokens": 243056213.0, "step": 6403 }, { "epoch": 0.8146546240936268, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4235646724700928, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8613293170928955, "num_tokens": 243097184.0, "step": 6404 }, { "epoch": 0.8147818343722173, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7451229095458984, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8598466515541077, "num_tokens": 243136995.0, "step": 6405 }, { "epoch": 0.8149090446508078, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.3421618938446045, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8752707839012146, "num_tokens": 243173495.0, "step": 6406 }, { "epoch": 0.8150362549293982, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.736445188522339, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8588334321975708, "num_tokens": 243209198.0, "step": 6407 }, { "epoch": 0.8151634652079888, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2698004245758057, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8675795197486877, "num_tokens": 243249777.0, "step": 6408 }, { "epoch": 0.8152906754865793, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.762162685394287, "learning_rate": 1e-05, "loss": 0.516, "mean_token_accuracy": 0.8503019213676453, "num_tokens": 243293094.0, "step": 6409 }, { "epoch": 0.8154178857651698, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.251800537109375, "learning_rate": 1e-05, "loss": 0.537, "mean_token_accuracy": 0.8455763459205627, "num_tokens": 243330881.0, "step": 6410 }, { "epoch": 0.8155450960437604, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.796718120574951, "learning_rate": 1e-05, "loss": 0.5103, "mean_token_accuracy": 0.8513187766075134, "num_tokens": 243371292.0, "step": 6411 }, { "epoch": 0.8156723063223509, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2886674404144287, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8470029830932617, "num_tokens": 243404925.0, "step": 6412 }, { "epoch": 0.8157995166009414, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.777127981185913, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8697969913482666, "num_tokens": 243442377.0, "step": 6413 }, { "epoch": 0.8159267268795318, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00015163421630859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.321019172668457, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8684602975845337, "num_tokens": 243476857.0, "step": 6414 }, { "epoch": 0.8160539371581224, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.730863571166992, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.879360556602478, "num_tokens": 243506641.0, "step": 6415 }, { "epoch": 0.8161811474367129, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3564441204071045, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8570283651351929, "num_tokens": 243547555.0, "step": 6416 }, { "epoch": 0.8163083577153034, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.618689775466919, "learning_rate": 1e-05, "loss": 0.5257, "mean_token_accuracy": 0.8517974615097046, "num_tokens": 243588210.0, "step": 6417 }, { "epoch": 0.8164355679938939, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.348921298980713, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8584381341934204, "num_tokens": 243626913.0, "step": 6418 }, { "epoch": 0.8165627782724845, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4989519119262695, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8720701336860657, "num_tokens": 243662645.0, "step": 6419 }, { "epoch": 0.8166899885510749, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3187341690063477, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8633846044540405, "num_tokens": 243704603.0, "step": 6420 }, { "epoch": 0.8168171988296654, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.419903516769409, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8550946712493896, "num_tokens": 243739571.0, "step": 6421 }, { "epoch": 0.8169444091082559, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2752766609191895, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8767895698547363, "num_tokens": 243782047.0, "step": 6422 }, { "epoch": 0.8170716193868465, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3650386333465576, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.845302939414978, "num_tokens": 243825159.0, "step": 6423 }, { "epoch": 0.817198829665437, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.231336832046509, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.8590612411499023, "num_tokens": 243865627.0, "step": 6424 }, { "epoch": 0.8173260399440275, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.343471050262451, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8559466004371643, "num_tokens": 243905750.0, "step": 6425 }, { "epoch": 0.8174532502226179, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.2025985717773438, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8529790639877319, "num_tokens": 243936176.0, "step": 6426 }, { "epoch": 0.8175804605012085, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3265113830566406, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8717436194419861, "num_tokens": 243979191.0, "step": 6427 }, { "epoch": 0.817707670779799, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.150398015975952, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.8509956002235413, "num_tokens": 244025074.0, "step": 6428 }, { "epoch": 0.8178348810583895, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3691468238830566, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8645175695419312, "num_tokens": 244062033.0, "step": 6429 }, { "epoch": 0.8179620913369801, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.088749885559082, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8766598701477051, "num_tokens": 244100025.0, "step": 6430 }, { "epoch": 0.8180893016155706, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4431121349334717, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.8800094723701477, "num_tokens": 244136253.0, "step": 6431 }, { "epoch": 0.818216511894161, "ewc_loss": 0.0556640625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.04032301902771, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8632352948188782, "num_tokens": 244174179.0, "step": 6432 }, { "epoch": 0.8183437221727515, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.6072628498077393, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8739438056945801, "num_tokens": 244209096.0, "step": 6433 }, { "epoch": 0.8184709324513421, "ewc_loss": 0.05615234375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.979921579360962, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8680675029754639, "num_tokens": 244248920.0, "step": 6434 }, { "epoch": 0.8185981427299326, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.77531099319458, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.860663115978241, "num_tokens": 244289338.0, "step": 6435 }, { "epoch": 0.8187253530085231, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0317649841308594, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8713721632957458, "num_tokens": 244324844.0, "step": 6436 }, { "epoch": 0.8188525632871136, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9357986450195312, "learning_rate": 1e-05, "loss": 0.5407, "mean_token_accuracy": 0.84589022397995, "num_tokens": 244364815.0, "step": 6437 }, { "epoch": 0.8189797735657041, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.1967945098876953, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8629306554794312, "num_tokens": 244407509.0, "step": 6438 }, { "epoch": 0.8191069838442946, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.946028709411621, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8748211860656738, "num_tokens": 244446663.0, "step": 6439 }, { "epoch": 0.8192341941228851, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.386284351348877, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8565090894699097, "num_tokens": 244492292.0, "step": 6440 }, { "epoch": 0.8193614044014756, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000152587890625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.801041841506958, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8610650300979614, "num_tokens": 244531390.0, "step": 6441 }, { "epoch": 0.8194886146800662, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.430788993835449, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8656617403030396, "num_tokens": 244567377.0, "step": 6442 }, { "epoch": 0.8196158249586567, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6066157817840576, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8777218461036682, "num_tokens": 244605690.0, "step": 6443 }, { "epoch": 0.8197430352372471, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4125800132751465, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8635736703872681, "num_tokens": 244643512.0, "step": 6444 }, { "epoch": 0.8198702455158376, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.429021120071411, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8499475717544556, "num_tokens": 244691048.0, "step": 6445 }, { "epoch": 0.8199974557944282, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3778915405273438, "learning_rate": 1e-05, "loss": 0.539, "mean_token_accuracy": 0.8392396569252014, "num_tokens": 244734205.0, "step": 6446 }, { "epoch": 0.8201246660730187, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.324169874191284, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8596889972686768, "num_tokens": 244768302.0, "step": 6447 }, { "epoch": 0.8202518763516092, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3411576747894287, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8626724481582642, "num_tokens": 244808880.0, "step": 6448 }, { "epoch": 0.8203790866301998, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2409090995788574, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8713181614875793, "num_tokens": 244845507.0, "step": 6449 }, { "epoch": 0.8205062969087902, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.3268744945526123, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8492805361747742, "num_tokens": 244888142.0, "step": 6450 }, { "epoch": 0.8206335071873807, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1750152111053467, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.877526044845581, "num_tokens": 244928612.0, "step": 6451 }, { "epoch": 0.8207607174659712, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.329549789428711, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8642587661743164, "num_tokens": 244969879.0, "step": 6452 }, { "epoch": 0.8208879277445618, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.1344757080078125, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8614435195922852, "num_tokens": 245004597.0, "step": 6453 }, { "epoch": 0.8210151380231523, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3624653816223145, "learning_rate": 1e-05, "loss": 0.5073, "mean_token_accuracy": 0.8558565974235535, "num_tokens": 245050087.0, "step": 6454 }, { "epoch": 0.8211423483017428, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0853025913238525, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.8543188571929932, "num_tokens": 245091626.0, "step": 6455 }, { "epoch": 0.8212695585803332, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4224259853363037, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8593685626983643, "num_tokens": 245126780.0, "step": 6456 }, { "epoch": 0.8213967688589238, "ewc_loss": 0.055908203125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 3.020413637161255, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8439805507659912, "num_tokens": 245169226.0, "step": 6457 }, { "epoch": 0.8215239791375143, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5517795085906982, "learning_rate": 1e-05, "loss": 0.519, "mean_token_accuracy": 0.8482397794723511, "num_tokens": 245204983.0, "step": 6458 }, { "epoch": 0.8216511894161048, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.944239854812622, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.868560791015625, "num_tokens": 245249162.0, "step": 6459 }, { "epoch": 0.8217783996946953, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7620861530303955, "learning_rate": 1e-05, "loss": 0.5319, "mean_token_accuracy": 0.8472452163696289, "num_tokens": 245288883.0, "step": 6460 }, { "epoch": 0.8219056099732859, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.9424829483032227, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8584842681884766, "num_tokens": 245328578.0, "step": 6461 }, { "epoch": 0.8220328202518764, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.023561000823975, "learning_rate": 1e-05, "loss": 0.5128, "mean_token_accuracy": 0.8569997549057007, "num_tokens": 245368086.0, "step": 6462 }, { "epoch": 0.8221600305304668, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.129567861557007, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8617236614227295, "num_tokens": 245405234.0, "step": 6463 }, { "epoch": 0.8222872408090574, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.180980682373047, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8692282438278198, "num_tokens": 245439759.0, "step": 6464 }, { "epoch": 0.8224144510876479, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.452585220336914, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.8571745157241821, "num_tokens": 245474689.0, "step": 6465 }, { "epoch": 0.8225416613662384, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.061622619628906, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8622504472732544, "num_tokens": 245509494.0, "step": 6466 }, { "epoch": 0.8226688716448289, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5802526473999023, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8652012944221497, "num_tokens": 245550798.0, "step": 6467 }, { "epoch": 0.8227960819234195, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8521711826324463, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8634862899780273, "num_tokens": 245588929.0, "step": 6468 }, { "epoch": 0.8229232922020099, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.532951831817627, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8635898232460022, "num_tokens": 245625952.0, "step": 6469 }, { "epoch": 0.8230505024806004, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6745622158050537, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8519441485404968, "num_tokens": 245668156.0, "step": 6470 }, { "epoch": 0.8231777127591909, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.44397234916687, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8617022633552551, "num_tokens": 245707117.0, "step": 6471 }, { "epoch": 0.8233049230377815, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.554248094558716, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8563175797462463, "num_tokens": 245747248.0, "step": 6472 }, { "epoch": 0.823432133316372, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3263633251190186, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8561287522315979, "num_tokens": 245795760.0, "step": 6473 }, { "epoch": 0.8235593435949625, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5089914798736572, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8663725256919861, "num_tokens": 245827778.0, "step": 6474 }, { "epoch": 0.8236865538735529, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.225214958190918, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8662908673286438, "num_tokens": 245866869.0, "step": 6475 }, { "epoch": 0.8238137641521435, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5169217586517334, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.875196099281311, "num_tokens": 245904686.0, "step": 6476 }, { "epoch": 0.823940974430734, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1467578411102295, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8594709634780884, "num_tokens": 245942169.0, "step": 6477 }, { "epoch": 0.8240681847093245, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.585273265838623, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8566781282424927, "num_tokens": 245979957.0, "step": 6478 }, { "epoch": 0.824195394987915, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.077573537826538, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8691142797470093, "num_tokens": 246019433.0, "step": 6479 }, { "epoch": 0.8243226052665056, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.725450277328491, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8604797720909119, "num_tokens": 246057595.0, "step": 6480 }, { "epoch": 0.824449815545096, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.060654640197754, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8721120953559875, "num_tokens": 246100141.0, "step": 6481 }, { "epoch": 0.8245770258236865, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.88196063041687, "learning_rate": 1e-05, "loss": 0.5465, "mean_token_accuracy": 0.8430327773094177, "num_tokens": 246141642.0, "step": 6482 }, { "epoch": 0.824704236102277, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2012152671813965, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8598302006721497, "num_tokens": 246177844.0, "step": 6483 }, { "epoch": 0.8248314463808676, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9223597049713135, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8579111099243164, "num_tokens": 246217470.0, "step": 6484 }, { "epoch": 0.8249586566594581, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3592543601989746, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8672833442687988, "num_tokens": 246256637.0, "step": 6485 }, { "epoch": 0.8250858669380486, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7915866374969482, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8698468208312988, "num_tokens": 246298894.0, "step": 6486 }, { "epoch": 0.8252130772166391, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.44000506401062, "learning_rate": 1e-05, "loss": 0.5026, "mean_token_accuracy": 0.8551981449127197, "num_tokens": 246340986.0, "step": 6487 }, { "epoch": 0.8253402874952296, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.594935178756714, "learning_rate": 1e-05, "loss": 0.5614, "mean_token_accuracy": 0.8349814414978027, "num_tokens": 246379201.0, "step": 6488 }, { "epoch": 0.8254674977738201, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4539289474487305, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8694916367530823, "num_tokens": 246411605.0, "step": 6489 }, { "epoch": 0.8255947080524106, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.37846040725708, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8627585768699646, "num_tokens": 246450908.0, "step": 6490 }, { "epoch": 0.8257219183310012, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.435204029083252, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8568480014801025, "num_tokens": 246493132.0, "step": 6491 }, { "epoch": 0.8258491286095917, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.226684331893921, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.850411057472229, "num_tokens": 246528532.0, "step": 6492 }, { "epoch": 0.8259763388881821, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4498350620269775, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8637141585350037, "num_tokens": 246567328.0, "step": 6493 }, { "epoch": 0.8261035491667726, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.111356258392334, "learning_rate": 1e-05, "loss": 0.5557, "mean_token_accuracy": 0.8385171890258789, "num_tokens": 246613798.0, "step": 6494 }, { "epoch": 0.8262307594453632, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.524148941040039, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8682361841201782, "num_tokens": 246649117.0, "step": 6495 }, { "epoch": 0.8263579697239537, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.023559331893921, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8649148941040039, "num_tokens": 246690304.0, "step": 6496 }, { "epoch": 0.8264851800025442, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6561923027038574, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8668533563613892, "num_tokens": 246728103.0, "step": 6497 }, { "epoch": 0.8266123902811348, "ewc_loss": 0.057373046875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.997174024581909, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8723578453063965, "num_tokens": 246763902.0, "step": 6498 }, { "epoch": 0.8267396005597252, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8484346866607666, "learning_rate": 1e-05, "loss": 0.5371, "mean_token_accuracy": 0.8457920551300049, "num_tokens": 246803899.0, "step": 6499 }, { "epoch": 0.8268668108383157, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0696895122528076, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8531997203826904, "num_tokens": 246840464.0, "step": 6500 }, { "epoch": 0.8269940211169062, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.988361120223999, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8627709150314331, "num_tokens": 246886743.0, "step": 6501 }, { "epoch": 0.8271212313954968, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2545783519744873, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8624244928359985, "num_tokens": 246925260.0, "step": 6502 }, { "epoch": 0.8272484416740873, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.001769065856934, "learning_rate": 1e-05, "loss": 0.5117, "mean_token_accuracy": 0.8540076017379761, "num_tokens": 246961523.0, "step": 6503 }, { "epoch": 0.8273756519526778, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.441243886947632, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.867103099822998, "num_tokens": 246996881.0, "step": 6504 }, { "epoch": 0.8275028622312682, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8358089923858643, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8556613922119141, "num_tokens": 247033492.0, "step": 6505 }, { "epoch": 0.8276300725098588, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4808833599090576, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8650320768356323, "num_tokens": 247065078.0, "step": 6506 }, { "epoch": 0.8277572827884493, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6446151733398438, "learning_rate": 1e-05, "loss": 0.5116, "mean_token_accuracy": 0.850591778755188, "num_tokens": 247103876.0, "step": 6507 }, { "epoch": 0.8278844930670398, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4326632022857666, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8524541854858398, "num_tokens": 247142132.0, "step": 6508 }, { "epoch": 0.8280117033456303, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.489826202392578, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8615261912345886, "num_tokens": 247182141.0, "step": 6509 }, { "epoch": 0.8281389136242209, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.373582363128662, "learning_rate": 1e-05, "loss": 0.5468, "mean_token_accuracy": 0.8434246182441711, "num_tokens": 247218240.0, "step": 6510 }, { "epoch": 0.8282661239028114, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3875694274902344, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8574893474578857, "num_tokens": 247257370.0, "step": 6511 }, { "epoch": 0.8283933341814018, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3062915802001953, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8629019856452942, "num_tokens": 247294205.0, "step": 6512 }, { "epoch": 0.8285205444599923, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.344658374786377, "learning_rate": 1e-05, "loss": 0.5151, "mean_token_accuracy": 0.8458618521690369, "num_tokens": 247329974.0, "step": 6513 }, { "epoch": 0.8286477547385829, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2442328929901123, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8570765256881714, "num_tokens": 247365279.0, "step": 6514 }, { "epoch": 0.8287749650171734, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00015354156494140625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3127095699310303, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8749020099639893, "num_tokens": 247405319.0, "step": 6515 }, { "epoch": 0.8289021752957639, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1793394088745117, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8744733333587646, "num_tokens": 247444399.0, "step": 6516 }, { "epoch": 0.8290293855743545, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.356914758682251, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8655433654785156, "num_tokens": 247475637.0, "step": 6517 }, { "epoch": 0.8291565958529449, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.113593816757202, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8661152720451355, "num_tokens": 247511009.0, "step": 6518 }, { "epoch": 0.8292838061315354, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.419187068939209, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8652594089508057, "num_tokens": 247549188.0, "step": 6519 }, { "epoch": 0.8294110164101259, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.036370277404785, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8569966554641724, "num_tokens": 247591914.0, "step": 6520 }, { "epoch": 0.8295382266887165, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5706682205200195, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8626533150672913, "num_tokens": 247626324.0, "step": 6521 }, { "epoch": 0.829665436967307, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.9755990505218506, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8599098920822144, "num_tokens": 247662846.0, "step": 6522 }, { "epoch": 0.8297926472458975, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7686331272125244, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8657533526420593, "num_tokens": 247700831.0, "step": 6523 }, { "epoch": 0.8299198575244879, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.9776651859283447, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8617370128631592, "num_tokens": 247740535.0, "step": 6524 }, { "epoch": 0.8300470678030785, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.009275913238525, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8583395481109619, "num_tokens": 247779490.0, "step": 6525 }, { "epoch": 0.830174278081669, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.172819137573242, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8524832725524902, "num_tokens": 247815867.0, "step": 6526 }, { "epoch": 0.8303014883602595, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.111637592315674, "learning_rate": 1e-05, "loss": 0.5908, "mean_token_accuracy": 0.832017183303833, "num_tokens": 247854871.0, "step": 6527 }, { "epoch": 0.83042869863885, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.4458203315734863, "learning_rate": 1e-05, "loss": 0.5109, "mean_token_accuracy": 0.8517686128616333, "num_tokens": 247897711.0, "step": 6528 }, { "epoch": 0.8305559089174406, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.955869197845459, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8588919639587402, "num_tokens": 247933792.0, "step": 6529 }, { "epoch": 0.830683119196031, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.5512030124664307, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8644809722900391, "num_tokens": 247964183.0, "step": 6530 }, { "epoch": 0.8308103294746215, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7320945262908936, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8515504598617554, "num_tokens": 248003064.0, "step": 6531 }, { "epoch": 0.830937539753212, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.493650197982788, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8666455149650574, "num_tokens": 248041025.0, "step": 6532 }, { "epoch": 0.8310647500318026, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.564438581466675, "learning_rate": 1e-05, "loss": 0.5104, "mean_token_accuracy": 0.8534303903579712, "num_tokens": 248079503.0, "step": 6533 }, { "epoch": 0.8311919603103931, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4246466159820557, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8600261211395264, "num_tokens": 248119541.0, "step": 6534 }, { "epoch": 0.8313191705889836, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4263458251953125, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8539903163909912, "num_tokens": 248163359.0, "step": 6535 }, { "epoch": 0.831446380867574, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.343796968460083, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8718345165252686, "num_tokens": 248204384.0, "step": 6536 }, { "epoch": 0.8315735911461646, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.350491762161255, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8678857088088989, "num_tokens": 248238812.0, "step": 6537 }, { "epoch": 0.8317008014247551, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2722489833831787, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8638507723808289, "num_tokens": 248277725.0, "step": 6538 }, { "epoch": 0.8318280117033456, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.335617780685425, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.8550109267234802, "num_tokens": 248317591.0, "step": 6539 }, { "epoch": 0.8319552219819362, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2023532390594482, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8698198795318604, "num_tokens": 248356330.0, "step": 6540 }, { "epoch": 0.8320824322605267, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001544952392578125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.365082025527954, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8638240098953247, "num_tokens": 248390854.0, "step": 6541 }, { "epoch": 0.8322096425391171, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.119762897491455, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8735499382019043, "num_tokens": 248431581.0, "step": 6542 }, { "epoch": 0.8323368528177076, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4116201400756836, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8696607351303101, "num_tokens": 248474009.0, "step": 6543 }, { "epoch": 0.8324640630962982, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0476906299591064, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8721654415130615, "num_tokens": 248512552.0, "step": 6544 }, { "epoch": 0.8325912733748887, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5377111434936523, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.866147518157959, "num_tokens": 248549729.0, "step": 6545 }, { "epoch": 0.8327184836534792, "ewc_loss": 0.056396484375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.981691360473633, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8614254593849182, "num_tokens": 248589481.0, "step": 6546 }, { "epoch": 0.8328456939320698, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.73618221282959, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8735790848731995, "num_tokens": 248621871.0, "step": 6547 }, { "epoch": 0.8329729042106602, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.974949359893799, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8605624437332153, "num_tokens": 248659416.0, "step": 6548 }, { "epoch": 0.8331001144892507, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.972299575805664, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8599342107772827, "num_tokens": 248700052.0, "step": 6549 }, { "epoch": 0.8332273247678412, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.1314194202423096, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8538042306900024, "num_tokens": 248731965.0, "step": 6550 }, { "epoch": 0.8333545350464318, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.107497215270996, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8567505478858948, "num_tokens": 248770030.0, "step": 6551 }, { "epoch": 0.8334817453250223, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3889718055725098, "learning_rate": 1e-05, "loss": 0.5105, "mean_token_accuracy": 0.8526564836502075, "num_tokens": 248811990.0, "step": 6552 }, { "epoch": 0.8336089556036128, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.045246124267578, "learning_rate": 1e-05, "loss": 0.5229, "mean_token_accuracy": 0.8478146195411682, "num_tokens": 248842728.0, "step": 6553 }, { "epoch": 0.8337361658822032, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5526609420776367, "learning_rate": 1e-05, "loss": 0.5225, "mean_token_accuracy": 0.8447361588478088, "num_tokens": 248885084.0, "step": 6554 }, { "epoch": 0.8338633761607938, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8062150478363037, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8698736429214478, "num_tokens": 248923701.0, "step": 6555 }, { "epoch": 0.8339905864393843, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.527621030807495, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8731526732444763, "num_tokens": 248960796.0, "step": 6556 }, { "epoch": 0.8341177967179748, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6134448051452637, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8644388914108276, "num_tokens": 249000628.0, "step": 6557 }, { "epoch": 0.8342450069965653, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4510698318481445, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8661429286003113, "num_tokens": 249034409.0, "step": 6558 }, { "epoch": 0.8343722172751559, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.474764585494995, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8706388473510742, "num_tokens": 249071227.0, "step": 6559 }, { "epoch": 0.8344994275537464, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.371187448501587, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8657435774803162, "num_tokens": 249105859.0, "step": 6560 }, { "epoch": 0.8346266378323368, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.381016969680786, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8653146624565125, "num_tokens": 249148237.0, "step": 6561 }, { "epoch": 0.8347538481109273, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.310783624649048, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8518208265304565, "num_tokens": 249181745.0, "step": 6562 }, { "epoch": 0.8348810583895179, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.334195137023926, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8567888736724854, "num_tokens": 249221839.0, "step": 6563 }, { "epoch": 0.8350082686681084, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2345902919769287, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8566091060638428, "num_tokens": 249264338.0, "step": 6564 }, { "epoch": 0.8351354789466989, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3233416080474854, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8595298528671265, "num_tokens": 249304624.0, "step": 6565 }, { "epoch": 0.8352626892252895, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1838138103485107, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8622498512268066, "num_tokens": 249342996.0, "step": 6566 }, { "epoch": 0.8353898995038799, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.345665693283081, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8580827713012695, "num_tokens": 249384795.0, "step": 6567 }, { "epoch": 0.8355171097824704, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1306025981903076, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8693941831588745, "num_tokens": 249418492.0, "step": 6568 }, { "epoch": 0.8356443200610609, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.4052138328552246, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8634875416755676, "num_tokens": 249455714.0, "step": 6569 }, { "epoch": 0.8357715303396515, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.062502145767212, "learning_rate": 1e-05, "loss": 0.5307, "mean_token_accuracy": 0.8455935716629028, "num_tokens": 249493162.0, "step": 6570 }, { "epoch": 0.835898740618242, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.507711172103882, "learning_rate": 1e-05, "loss": 0.5123, "mean_token_accuracy": 0.8514412641525269, "num_tokens": 249538154.0, "step": 6571 }, { "epoch": 0.8360259508968325, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.986215829849243, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8679282665252686, "num_tokens": 249576457.0, "step": 6572 }, { "epoch": 0.8361531611754229, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6953916549682617, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8584105968475342, "num_tokens": 249611408.0, "step": 6573 }, { "epoch": 0.8362803714540135, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.9518330097198486, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8585782051086426, "num_tokens": 249651371.0, "step": 6574 }, { "epoch": 0.836407581732604, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.943070650100708, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.860369086265564, "num_tokens": 249697669.0, "step": 6575 }, { "epoch": 0.8365347920111945, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.070936679840088, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8531636595726013, "num_tokens": 249739640.0, "step": 6576 }, { "epoch": 0.836662002289785, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.148090362548828, "learning_rate": 1e-05, "loss": 0.5162, "mean_token_accuracy": 0.8523285984992981, "num_tokens": 249778430.0, "step": 6577 }, { "epoch": 0.8367892125683756, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.379525661468506, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8708782196044922, "num_tokens": 249817875.0, "step": 6578 }, { "epoch": 0.836916422846966, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.071945667266846, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8815009593963623, "num_tokens": 249861987.0, "step": 6579 }, { "epoch": 0.8370436331255565, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5634663105010986, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8656346201896667, "num_tokens": 249903137.0, "step": 6580 }, { "epoch": 0.837170843404147, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8682124614715576, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8551320433616638, "num_tokens": 249937760.0, "step": 6581 }, { "epoch": 0.8372980536827376, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5510263442993164, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8571985960006714, "num_tokens": 249977280.0, "step": 6582 }, { "epoch": 0.8374252639613281, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6621880531311035, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8536174297332764, "num_tokens": 250022160.0, "step": 6583 }, { "epoch": 0.8375524742399186, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4643678665161133, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.86656653881073, "num_tokens": 250063328.0, "step": 6584 }, { "epoch": 0.837679684518509, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.519353151321411, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.871851921081543, "num_tokens": 250096994.0, "step": 6585 }, { "epoch": 0.8378068947970996, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3905186653137207, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8664014339447021, "num_tokens": 250126058.0, "step": 6586 }, { "epoch": 0.8379341050756901, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4142110347747803, "learning_rate": 1e-05, "loss": 0.5171, "mean_token_accuracy": 0.8475664258003235, "num_tokens": 250171965.0, "step": 6587 }, { "epoch": 0.8380613153542806, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3041064739227295, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.858117938041687, "num_tokens": 250209862.0, "step": 6588 }, { "epoch": 0.8381885256328712, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3703651428222656, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8609071969985962, "num_tokens": 250244895.0, "step": 6589 }, { "epoch": 0.8383157359114617, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2245898246765137, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8624181747436523, "num_tokens": 250283340.0, "step": 6590 }, { "epoch": 0.8384429461900521, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.377866506576538, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8605003356933594, "num_tokens": 250318671.0, "step": 6591 }, { "epoch": 0.8385701564686426, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1421613693237305, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8673268556594849, "num_tokens": 250357718.0, "step": 6592 }, { "epoch": 0.8386973667472332, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4409124851226807, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8610018491744995, "num_tokens": 250391583.0, "step": 6593 }, { "epoch": 0.8388245770258237, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0801711082458496, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8611260056495667, "num_tokens": 250420938.0, "step": 6594 }, { "epoch": 0.8389517873044142, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5356605052948, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8717359304428101, "num_tokens": 250455608.0, "step": 6595 }, { "epoch": 0.8390789975830047, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000156402587890625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0079567432403564, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8568042516708374, "num_tokens": 250489745.0, "step": 6596 }, { "epoch": 0.8392062078615952, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.710054636001587, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8602610230445862, "num_tokens": 250523711.0, "step": 6597 }, { "epoch": 0.8393334181401857, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.986985921859741, "learning_rate": 1e-05, "loss": 0.5202, "mean_token_accuracy": 0.850989043712616, "num_tokens": 250562044.0, "step": 6598 }, { "epoch": 0.8394606284187762, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9068496227264404, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8570709228515625, "num_tokens": 250605930.0, "step": 6599 }, { "epoch": 0.8395878386973668, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.106377363204956, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.860938310623169, "num_tokens": 250640621.0, "step": 6600 }, { "epoch": 0.8397150489759573, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.04897928237915, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8632752299308777, "num_tokens": 250678056.0, "step": 6601 }, { "epoch": 0.8398422592545478, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3455352783203125, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.867955207824707, "num_tokens": 250719642.0, "step": 6602 }, { "epoch": 0.8399694695331382, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.991553783416748, "learning_rate": 1e-05, "loss": 0.5561, "mean_token_accuracy": 0.8399649858474731, "num_tokens": 250757643.0, "step": 6603 }, { "epoch": 0.8400966798117288, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5147786140441895, "learning_rate": 1e-05, "loss": 0.518, "mean_token_accuracy": 0.8543813824653625, "num_tokens": 250795277.0, "step": 6604 }, { "epoch": 0.8402238900903193, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7676682472229004, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8585996627807617, "num_tokens": 250834161.0, "step": 6605 }, { "epoch": 0.8403511003689098, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5042502880096436, "learning_rate": 1e-05, "loss": 0.4997, "mean_token_accuracy": 0.8554410934448242, "num_tokens": 250872518.0, "step": 6606 }, { "epoch": 0.8404783106475003, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.564642906188965, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8671731948852539, "num_tokens": 250908516.0, "step": 6607 }, { "epoch": 0.8406055209260909, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4426329135894775, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8573606014251709, "num_tokens": 250949351.0, "step": 6608 }, { "epoch": 0.8407327312046813, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4124345779418945, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8615385293960571, "num_tokens": 250987145.0, "step": 6609 }, { "epoch": 0.8408599414832718, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3822503089904785, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8695372343063354, "num_tokens": 251024073.0, "step": 6610 }, { "epoch": 0.8409871517618623, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3110034465789795, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8530444502830505, "num_tokens": 251069537.0, "step": 6611 }, { "epoch": 0.8411143620404529, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3612987995147705, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8541430830955505, "num_tokens": 251103343.0, "step": 6612 }, { "epoch": 0.8412415723190434, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2373507022857666, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.85558021068573, "num_tokens": 251140437.0, "step": 6613 }, { "epoch": 0.8413687825976339, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.324061632156372, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.862305223941803, "num_tokens": 251178002.0, "step": 6614 }, { "epoch": 0.8414959928762245, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1889781951904297, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8723769783973694, "num_tokens": 251219694.0, "step": 6615 }, { "epoch": 0.8416232031548149, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3311824798583984, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8666186332702637, "num_tokens": 251255576.0, "step": 6616 }, { "epoch": 0.8417504134334054, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1565847396850586, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8554887771606445, "num_tokens": 251287245.0, "step": 6617 }, { "epoch": 0.8418776237119959, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.368633508682251, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8607012033462524, "num_tokens": 251320435.0, "step": 6618 }, { "epoch": 0.8420048339905865, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.083054780960083, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8697354197502136, "num_tokens": 251358893.0, "step": 6619 }, { "epoch": 0.842132044269177, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4353740215301514, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8725210428237915, "num_tokens": 251392805.0, "step": 6620 }, { "epoch": 0.8422592545477675, "ewc_loss": 0.056640625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0030272006988525, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8696069717407227, "num_tokens": 251436848.0, "step": 6621 }, { "epoch": 0.8423864648263579, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.585073232650757, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8703517913818359, "num_tokens": 251471590.0, "step": 6622 }, { "epoch": 0.8425136751049485, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.9662070274353027, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8549478650093079, "num_tokens": 251509251.0, "step": 6623 }, { "epoch": 0.842640885383539, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7923901081085205, "learning_rate": 1e-05, "loss": 0.526, "mean_token_accuracy": 0.8486752510070801, "num_tokens": 251549337.0, "step": 6624 }, { "epoch": 0.8427680956621295, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 2.9543392658233643, "learning_rate": 1e-05, "loss": 0.5419, "mean_token_accuracy": 0.8402634263038635, "num_tokens": 251593953.0, "step": 6625 }, { "epoch": 0.84289530594072, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.061359882354736, "learning_rate": 1e-05, "loss": 0.5039, "mean_token_accuracy": 0.859710156917572, "num_tokens": 251631797.0, "step": 6626 }, { "epoch": 0.8430225162193106, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.1807494163513184, "learning_rate": 1e-05, "loss": 0.5424, "mean_token_accuracy": 0.8477438688278198, "num_tokens": 251668378.0, "step": 6627 }, { "epoch": 0.843149726497901, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.177308082580566, "learning_rate": 1e-05, "loss": 0.5281, "mean_token_accuracy": 0.848675012588501, "num_tokens": 251712248.0, "step": 6628 }, { "epoch": 0.8432769367764915, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.483086109161377, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8601340055465698, "num_tokens": 251748770.0, "step": 6629 }, { "epoch": 0.843404147055082, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.036992073059082, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8656057715415955, "num_tokens": 251787262.0, "step": 6630 }, { "epoch": 0.8435313573336726, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.602912187576294, "learning_rate": 1e-05, "loss": 0.5274, "mean_token_accuracy": 0.8476312756538391, "num_tokens": 251820584.0, "step": 6631 }, { "epoch": 0.8436585676122631, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.814046621322632, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8527264595031738, "num_tokens": 251861394.0, "step": 6632 }, { "epoch": 0.8437857778908536, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.535517930984497, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.860385537147522, "num_tokens": 251902144.0, "step": 6633 }, { "epoch": 0.843912988169444, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.629371166229248, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8831902742385864, "num_tokens": 251936162.0, "step": 6634 }, { "epoch": 0.8440401984480346, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.439962387084961, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8548333644866943, "num_tokens": 251977223.0, "step": 6635 }, { "epoch": 0.8441674087266251, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.515002727508545, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.865127682685852, "num_tokens": 252013413.0, "step": 6636 }, { "epoch": 0.8442946190052156, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.34822678565979, "learning_rate": 1e-05, "loss": 0.5032, "mean_token_accuracy": 0.852289617061615, "num_tokens": 252051395.0, "step": 6637 }, { "epoch": 0.8444218292838062, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.443221092224121, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.85490483045578, "num_tokens": 252092982.0, "step": 6638 }, { "epoch": 0.8445490395623967, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2515323162078857, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8798801898956299, "num_tokens": 252128953.0, "step": 6639 }, { "epoch": 0.8446762498409871, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.437798500061035, "learning_rate": 1e-05, "loss": 0.5064, "mean_token_accuracy": 0.8492350578308105, "num_tokens": 252171438.0, "step": 6640 }, { "epoch": 0.8448034601195776, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.166273355484009, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8584764003753662, "num_tokens": 252210310.0, "step": 6641 }, { "epoch": 0.8449306703981682, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4957118034362793, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8746403455734253, "num_tokens": 252242690.0, "step": 6642 }, { "epoch": 0.8450578806767587, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.071708917617798, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8607169389724731, "num_tokens": 252281570.0, "step": 6643 }, { "epoch": 0.8451850909553492, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6245689392089844, "learning_rate": 1e-05, "loss": 0.5213, "mean_token_accuracy": 0.8482264280319214, "num_tokens": 252323478.0, "step": 6644 }, { "epoch": 0.8453123012339397, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0195858478546143, "learning_rate": 1e-05, "loss": 0.518, "mean_token_accuracy": 0.8453559279441833, "num_tokens": 252362620.0, "step": 6645 }, { "epoch": 0.8454395115125302, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8128652572631836, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8710633516311646, "num_tokens": 252397549.0, "step": 6646 }, { "epoch": 0.8455667217911207, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0671069622039795, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8545514941215515, "num_tokens": 252433021.0, "step": 6647 }, { "epoch": 0.8456939320697112, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9799256324768066, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8617089986801147, "num_tokens": 252466455.0, "step": 6648 }, { "epoch": 0.8458211423483017, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2613937854766846, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.8552301526069641, "num_tokens": 252503547.0, "step": 6649 }, { "epoch": 0.8459483526268923, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.97692608833313, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8549582362174988, "num_tokens": 252535621.0, "step": 6650 }, { "epoch": 0.8460755629054828, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.445208787918091, "learning_rate": 1e-05, "loss": 0.5094, "mean_token_accuracy": 0.8535724878311157, "num_tokens": 252573699.0, "step": 6651 }, { "epoch": 0.8462027731840732, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.784559726715088, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8688869476318359, "num_tokens": 252619816.0, "step": 6652 }, { "epoch": 0.8463299834626637, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00015735626220703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.4944400787353516, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8708208799362183, "num_tokens": 252652631.0, "step": 6653 }, { "epoch": 0.8464571937412543, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.577188491821289, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8668795824050903, "num_tokens": 252690534.0, "step": 6654 }, { "epoch": 0.8465844040198448, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4712297916412354, "learning_rate": 1e-05, "loss": 0.5142, "mean_token_accuracy": 0.8501519560813904, "num_tokens": 252723090.0, "step": 6655 }, { "epoch": 0.8467116142984353, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 5.75693416595459, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8739489316940308, "num_tokens": 252755879.0, "step": 6656 }, { "epoch": 0.8468388245770259, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.9878013134002686, "learning_rate": 1e-05, "loss": 0.5515, "mean_token_accuracy": 0.8428072929382324, "num_tokens": 252799137.0, "step": 6657 }, { "epoch": 0.8469660348556163, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001583099365234375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2562968730926514, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8595227599143982, "num_tokens": 252840871.0, "step": 6658 }, { "epoch": 0.8470932451342068, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.58769154548645, "learning_rate": 1e-05, "loss": 0.5045, "mean_token_accuracy": 0.8547852039337158, "num_tokens": 252879708.0, "step": 6659 }, { "epoch": 0.8472204554127973, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1927196979522705, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8650141954421997, "num_tokens": 252914223.0, "step": 6660 }, { "epoch": 0.8473476656913879, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.651493549346924, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8610498905181885, "num_tokens": 252956666.0, "step": 6661 }, { "epoch": 0.8474748759699784, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.0797462463378906, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8611167669296265, "num_tokens": 252997566.0, "step": 6662 }, { "epoch": 0.8476020862485689, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.794233798980713, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8682166337966919, "num_tokens": 253038545.0, "step": 6663 }, { "epoch": 0.8477292965271594, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00015926361083984375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1170270442962646, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8636000156402588, "num_tokens": 253078123.0, "step": 6664 }, { "epoch": 0.8478565068057499, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.878763437271118, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8729232549667358, "num_tokens": 253118273.0, "step": 6665 }, { "epoch": 0.8479837170843404, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.2403907775878906, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8496126532554626, "num_tokens": 253152994.0, "step": 6666 }, { "epoch": 0.8481109273629309, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.8803560733795166, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.873798668384552, "num_tokens": 253189736.0, "step": 6667 }, { "epoch": 0.8482381376415215, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3478622436523438, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.8508813381195068, "num_tokens": 253225872.0, "step": 6668 }, { "epoch": 0.848365347920112, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.754390001296997, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8745311498641968, "num_tokens": 253263902.0, "step": 6669 }, { "epoch": 0.8484925581987025, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.370269775390625, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8719223737716675, "num_tokens": 253297536.0, "step": 6670 }, { "epoch": 0.8486197684772929, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6234130859375, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8719683885574341, "num_tokens": 253334624.0, "step": 6671 }, { "epoch": 0.8487469787558835, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3344221115112305, "learning_rate": 1e-05, "loss": 0.5045, "mean_token_accuracy": 0.8560718894004822, "num_tokens": 253379203.0, "step": 6672 }, { "epoch": 0.848874189034474, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00016021728515625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.520737409591675, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8704836964607239, "num_tokens": 253417209.0, "step": 6673 }, { "epoch": 0.8490013993130645, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2893056869506836, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8492206931114197, "num_tokens": 253453909.0, "step": 6674 }, { "epoch": 0.849128609591655, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4812204837799072, "learning_rate": 1e-05, "loss": 0.5945, "mean_token_accuracy": 0.8289451599121094, "num_tokens": 253489448.0, "step": 6675 }, { "epoch": 0.8492558198702456, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.243105888366699, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.860569953918457, "num_tokens": 253517544.0, "step": 6676 }, { "epoch": 0.849383030148836, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.426518440246582, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8684521317481995, "num_tokens": 253558829.0, "step": 6677 }, { "epoch": 0.8495102404274265, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1778244972229004, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8639199733734131, "num_tokens": 253593944.0, "step": 6678 }, { "epoch": 0.849637450706017, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.439696788787842, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8574057817459106, "num_tokens": 253633579.0, "step": 6679 }, { "epoch": 0.8497646609846076, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.111067771911621, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8573154211044312, "num_tokens": 253680247.0, "step": 6680 }, { "epoch": 0.8498918712631981, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5038528442382812, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8633366823196411, "num_tokens": 253714766.0, "step": 6681 }, { "epoch": 0.8500190815417886, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.080716133117676, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.8532751202583313, "num_tokens": 253747643.0, "step": 6682 }, { "epoch": 0.850146291820379, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5826945304870605, "learning_rate": 1e-05, "loss": 0.5339, "mean_token_accuracy": 0.8422795534133911, "num_tokens": 253791252.0, "step": 6683 }, { "epoch": 0.8502735020989696, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0523557662963867, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8666616678237915, "num_tokens": 253822244.0, "step": 6684 }, { "epoch": 0.8504007123775601, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6978492736816406, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8550938367843628, "num_tokens": 253858680.0, "step": 6685 }, { "epoch": 0.8505279226561506, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.069870710372925, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8722095489501953, "num_tokens": 253896266.0, "step": 6686 }, { "epoch": 0.8506551329347412, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.797837257385254, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8624968528747559, "num_tokens": 253934059.0, "step": 6687 }, { "epoch": 0.8507823432133317, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.155809164047241, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8638070821762085, "num_tokens": 253967862.0, "step": 6688 }, { "epoch": 0.8509095534919221, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8193018436431885, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8546375632286072, "num_tokens": 254008277.0, "step": 6689 }, { "epoch": 0.8510367637705126, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2787134647369385, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8633320331573486, "num_tokens": 254039939.0, "step": 6690 }, { "epoch": 0.8511639740491032, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.722255229949951, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.866247832775116, "num_tokens": 254080497.0, "step": 6691 }, { "epoch": 0.8512911843276937, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3429651260375977, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8628808856010437, "num_tokens": 254111917.0, "step": 6692 }, { "epoch": 0.8514183946062842, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5589137077331543, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8705554604530334, "num_tokens": 254150245.0, "step": 6693 }, { "epoch": 0.8515456048848747, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.35205078125, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.86069655418396, "num_tokens": 254187106.0, "step": 6694 }, { "epoch": 0.8516728151634652, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4123783111572266, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.852331280708313, "num_tokens": 254219877.0, "step": 6695 }, { "epoch": 0.8518000254420557, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.347007989883423, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.8533187508583069, "num_tokens": 254264234.0, "step": 6696 }, { "epoch": 0.8519272357206462, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2922072410583496, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8673521280288696, "num_tokens": 254299762.0, "step": 6697 }, { "epoch": 0.8520544459992367, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3632495403289795, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8471921682357788, "num_tokens": 254336403.0, "step": 6698 }, { "epoch": 0.8521816562778273, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1835992336273193, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8663826584815979, "num_tokens": 254375971.0, "step": 6699 }, { "epoch": 0.8523088665564178, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3814926147460938, "learning_rate": 1e-05, "loss": 0.4946, "mean_token_accuracy": 0.8520795702934265, "num_tokens": 254418430.0, "step": 6700 }, { "epoch": 0.8524360768350082, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1121647357940674, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8622462153434753, "num_tokens": 254457275.0, "step": 6701 }, { "epoch": 0.8525632871135987, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.431772232055664, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8554859161376953, "num_tokens": 254499280.0, "step": 6702 }, { "epoch": 0.8526904973921893, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.053629159927368, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.855080246925354, "num_tokens": 254536571.0, "step": 6703 }, { "epoch": 0.8528177076707798, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5359609127044678, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8504727482795715, "num_tokens": 254576900.0, "step": 6704 }, { "epoch": 0.8529449179493703, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9859724044799805, "learning_rate": 1e-05, "loss": 0.5455, "mean_token_accuracy": 0.8382362127304077, "num_tokens": 254617791.0, "step": 6705 }, { "epoch": 0.8530721282279609, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.712646722793579, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8695279955863953, "num_tokens": 254658300.0, "step": 6706 }, { "epoch": 0.8531993385065513, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.9639875888824463, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8593331575393677, "num_tokens": 254698133.0, "step": 6707 }, { "epoch": 0.8533265487851418, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.929272413253784, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8652790784835815, "num_tokens": 254736099.0, "step": 6708 }, { "epoch": 0.8534537590637323, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.0786547660827637, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8730096817016602, "num_tokens": 254773460.0, "step": 6709 }, { "epoch": 0.8535809693423229, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.08566951751709, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8671422600746155, "num_tokens": 254812722.0, "step": 6710 }, { "epoch": 0.8537081796209134, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.32912278175354, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8646799921989441, "num_tokens": 254851994.0, "step": 6711 }, { "epoch": 0.8538353898995039, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.050609588623047, "learning_rate": 1e-05, "loss": 0.5196, "mean_token_accuracy": 0.8538562059402466, "num_tokens": 254889111.0, "step": 6712 }, { "epoch": 0.8539626001780944, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5085902214050293, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8573323488235474, "num_tokens": 254931779.0, "step": 6713 }, { "epoch": 0.8540898104566849, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.86456036567688, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.8647315502166748, "num_tokens": 254968709.0, "step": 6714 }, { "epoch": 0.8542170207352754, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.527318000793457, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8647728562355042, "num_tokens": 255000235.0, "step": 6715 }, { "epoch": 0.8543442310138659, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6683900356292725, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8564423322677612, "num_tokens": 255041157.0, "step": 6716 }, { "epoch": 0.8544714412924564, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4541406631469727, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8580695986747742, "num_tokens": 255077001.0, "step": 6717 }, { "epoch": 0.854598651571047, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.511810302734375, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8677668571472168, "num_tokens": 255113581.0, "step": 6718 }, { "epoch": 0.8547258618496375, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.370922803878784, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8749241828918457, "num_tokens": 255153183.0, "step": 6719 }, { "epoch": 0.8548530721282279, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.421565294265747, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8694638609886169, "num_tokens": 255192216.0, "step": 6720 }, { "epoch": 0.8549802824068184, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.294029474258423, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8665804862976074, "num_tokens": 255229969.0, "step": 6721 }, { "epoch": 0.855107492685409, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.391106128692627, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8516732454299927, "num_tokens": 255267232.0, "step": 6722 }, { "epoch": 0.8552347029639995, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2193822860717773, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8632774949073792, "num_tokens": 255304923.0, "step": 6723 }, { "epoch": 0.85536191324259, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.396202564239502, "learning_rate": 1e-05, "loss": 0.5304, "mean_token_accuracy": 0.842505931854248, "num_tokens": 255342422.0, "step": 6724 }, { "epoch": 0.8554891235211806, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1428773403167725, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8692569136619568, "num_tokens": 255380218.0, "step": 6725 }, { "epoch": 0.855616333799771, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.448421001434326, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8636568784713745, "num_tokens": 255413964.0, "step": 6726 }, { "epoch": 0.8557435440783615, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0692875385284424, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8681588172912598, "num_tokens": 255450421.0, "step": 6727 }, { "epoch": 0.855870754356952, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5639991760253906, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8628201484680176, "num_tokens": 255487239.0, "step": 6728 }, { "epoch": 0.8559979646355426, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.013911008834839, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8603562116622925, "num_tokens": 255529363.0, "step": 6729 }, { "epoch": 0.8561251749141331, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.726693868637085, "learning_rate": 1e-05, "loss": 0.3939, "mean_token_accuracy": 0.8859472870826721, "num_tokens": 255570790.0, "step": 6730 }, { "epoch": 0.8562523851927236, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0110652446746826, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8574423789978027, "num_tokens": 255608188.0, "step": 6731 }, { "epoch": 0.856379595471314, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.947531223297119, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8617141246795654, "num_tokens": 255640929.0, "step": 6732 }, { "epoch": 0.8565068057499046, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.1471035480499268, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8652915358543396, "num_tokens": 255684579.0, "step": 6733 }, { "epoch": 0.8566340160284951, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0541558265686035, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8607774376869202, "num_tokens": 255721195.0, "step": 6734 }, { "epoch": 0.8567612263070856, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.392707109451294, "learning_rate": 1e-05, "loss": 0.49, "mean_token_accuracy": 0.8617969751358032, "num_tokens": 255761697.0, "step": 6735 }, { "epoch": 0.8568884365856761, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9385223388671875, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8594812154769897, "num_tokens": 255800492.0, "step": 6736 }, { "epoch": 0.8570156468642667, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.511296272277832, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8635692596435547, "num_tokens": 255837074.0, "step": 6737 }, { "epoch": 0.8571428571428571, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.718498706817627, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8623483180999756, "num_tokens": 255876238.0, "step": 6738 }, { "epoch": 0.8572700674214476, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4988393783569336, "learning_rate": 1e-05, "loss": 0.525, "mean_token_accuracy": 0.8468464016914368, "num_tokens": 255914622.0, "step": 6739 }, { "epoch": 0.8573972777000382, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5101094245910645, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8639085292816162, "num_tokens": 255954027.0, "step": 6740 }, { "epoch": 0.8575244879786287, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.441274642944336, "learning_rate": 1e-05, "loss": 0.503, "mean_token_accuracy": 0.8536856770515442, "num_tokens": 255997416.0, "step": 6741 }, { "epoch": 0.8576516982572192, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.362792491912842, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8597321510314941, "num_tokens": 256030447.0, "step": 6742 }, { "epoch": 0.8577789085358097, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.391725540161133, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8652321100234985, "num_tokens": 256068533.0, "step": 6743 }, { "epoch": 0.8579061188144002, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2664191722869873, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8755006790161133, "num_tokens": 256104783.0, "step": 6744 }, { "epoch": 0.8580333290929907, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3664615154266357, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8654295206069946, "num_tokens": 256138272.0, "step": 6745 }, { "epoch": 0.8581605393715812, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.215775489807129, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8537284731864929, "num_tokens": 256171102.0, "step": 6746 }, { "epoch": 0.8582877496501717, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3484599590301514, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8633476495742798, "num_tokens": 256210593.0, "step": 6747 }, { "epoch": 0.8584149599287623, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1690759658813477, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8714159727096558, "num_tokens": 256248693.0, "step": 6748 }, { "epoch": 0.8585421702073528, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3512346744537354, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8675455451011658, "num_tokens": 256284875.0, "step": 6749 }, { "epoch": 0.8586693804859432, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.132105827331543, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8672466278076172, "num_tokens": 256319121.0, "step": 6750 }, { "epoch": 0.8587965907645337, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3894011974334717, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8713356256484985, "num_tokens": 256354112.0, "step": 6751 }, { "epoch": 0.8589238010431243, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.0762150287628174, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8738741874694824, "num_tokens": 256387612.0, "step": 6752 }, { "epoch": 0.8590510113217148, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4749810695648193, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8622322678565979, "num_tokens": 256419720.0, "step": 6753 }, { "epoch": 0.8591782216003053, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 3.0078284740448, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8532164692878723, "num_tokens": 256456717.0, "step": 6754 }, { "epoch": 0.8593054318788959, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6173980236053467, "learning_rate": 1e-05, "loss": 0.5353, "mean_token_accuracy": 0.8432303071022034, "num_tokens": 256496480.0, "step": 6755 }, { "epoch": 0.8594326421574863, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.953204393386841, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8654338121414185, "num_tokens": 256531974.0, "step": 6756 }, { "epoch": 0.8595598524360768, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.841503620147705, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.8553118705749512, "num_tokens": 256565983.0, "step": 6757 }, { "epoch": 0.8596870627146673, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.981454372406006, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8689790964126587, "num_tokens": 256607215.0, "step": 6758 }, { "epoch": 0.8598142729932579, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.095552444458008, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8745191097259521, "num_tokens": 256646213.0, "step": 6759 }, { "epoch": 0.8599414832718484, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2299275398254395, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8686222434043884, "num_tokens": 256687638.0, "step": 6760 }, { "epoch": 0.8600686935504389, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.187972068786621, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8584705591201782, "num_tokens": 256716852.0, "step": 6761 }, { "epoch": 0.8601959038290294, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.548542022705078, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8622218370437622, "num_tokens": 256748260.0, "step": 6762 }, { "epoch": 0.8603231141076199, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9927752017974854, "learning_rate": 1e-05, "loss": 0.5497, "mean_token_accuracy": 0.8441698551177979, "num_tokens": 256786400.0, "step": 6763 }, { "epoch": 0.8604503243862104, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.595545768737793, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8720927238464355, "num_tokens": 256820186.0, "step": 6764 }, { "epoch": 0.8605775346648009, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7851955890655518, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8626388907432556, "num_tokens": 256855735.0, "step": 6765 }, { "epoch": 0.8607047449433914, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5219993591308594, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8708291053771973, "num_tokens": 256887809.0, "step": 6766 }, { "epoch": 0.860831955221982, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6202657222747803, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8656983375549316, "num_tokens": 256921801.0, "step": 6767 }, { "epoch": 0.8609591655005725, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.410897731781006, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8770747780799866, "num_tokens": 256960399.0, "step": 6768 }, { "epoch": 0.8610863757791629, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.520127296447754, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8653146624565125, "num_tokens": 256995094.0, "step": 6769 }, { "epoch": 0.8612135860577534, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3157894611358643, "learning_rate": 1e-05, "loss": 0.5513, "mean_token_accuracy": 0.8413211107254028, "num_tokens": 257037792.0, "step": 6770 }, { "epoch": 0.861340796336344, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4814000129699707, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8653521537780762, "num_tokens": 257075376.0, "step": 6771 }, { "epoch": 0.8614680066149345, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2234444618225098, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8551284670829773, "num_tokens": 257115878.0, "step": 6772 }, { "epoch": 0.861595216893525, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.488760232925415, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8619869351387024, "num_tokens": 257150780.0, "step": 6773 }, { "epoch": 0.8617224271721156, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1267027854919434, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8645167350769043, "num_tokens": 257194704.0, "step": 6774 }, { "epoch": 0.861849637450706, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.581979751586914, "learning_rate": 1e-05, "loss": 0.5158, "mean_token_accuracy": 0.8502333164215088, "num_tokens": 257235337.0, "step": 6775 }, { "epoch": 0.8619768477292965, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.076054573059082, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.8501054048538208, "num_tokens": 257268976.0, "step": 6776 }, { "epoch": 0.862104058007887, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.717386245727539, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8628842830657959, "num_tokens": 257304638.0, "step": 6777 }, { "epoch": 0.8622312682864776, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.066774845123291, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8536403179168701, "num_tokens": 257341773.0, "step": 6778 }, { "epoch": 0.8623584785650681, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8774566650390625, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8648412227630615, "num_tokens": 257375145.0, "step": 6779 }, { "epoch": 0.8624856888436586, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.1549928188323975, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8680697679519653, "num_tokens": 257411070.0, "step": 6780 }, { "epoch": 0.862612899122249, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9669370651245117, "learning_rate": 1e-05, "loss": 0.5072, "mean_token_accuracy": 0.8510873317718506, "num_tokens": 257447283.0, "step": 6781 }, { "epoch": 0.8627401094008396, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.34751558303833, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8620298504829407, "num_tokens": 257487370.0, "step": 6782 }, { "epoch": 0.8628673196794301, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8698315620422363, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8495293855667114, "num_tokens": 257525028.0, "step": 6783 }, { "epoch": 0.8629945299580206, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.4665169715881348, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8586961030960083, "num_tokens": 257563884.0, "step": 6784 }, { "epoch": 0.8631217402366111, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6596062183380127, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8662009835243225, "num_tokens": 257599431.0, "step": 6785 }, { "epoch": 0.8632489505152017, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.471614360809326, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8604391813278198, "num_tokens": 257636060.0, "step": 6786 }, { "epoch": 0.8633761607937921, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4378621578216553, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8702265620231628, "num_tokens": 257672374.0, "step": 6787 }, { "epoch": 0.8635033710723826, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.460353374481201, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8546621203422546, "num_tokens": 257710875.0, "step": 6788 }, { "epoch": 0.8636305813509731, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.256471872329712, "learning_rate": 1e-05, "loss": 0.5145, "mean_token_accuracy": 0.8508108854293823, "num_tokens": 257756697.0, "step": 6789 }, { "epoch": 0.8637577916295637, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4747579097747803, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8556252121925354, "num_tokens": 257795281.0, "step": 6790 }, { "epoch": 0.8638850019081542, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.127018928527832, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8740613460540771, "num_tokens": 257835743.0, "step": 6791 }, { "epoch": 0.8640122121867447, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5230982303619385, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8750715255737305, "num_tokens": 257878546.0, "step": 6792 }, { "epoch": 0.8641394224653351, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0558645725250244, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8660964965820312, "num_tokens": 257915480.0, "step": 6793 }, { "epoch": 0.8642666327439257, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.638063669204712, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.8613014221191406, "num_tokens": 257951215.0, "step": 6794 }, { "epoch": 0.8643938430225162, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.020561695098877, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.8574910759925842, "num_tokens": 257987623.0, "step": 6795 }, { "epoch": 0.8645210533011067, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7922518253326416, "learning_rate": 1e-05, "loss": 0.52, "mean_token_accuracy": 0.8496759533882141, "num_tokens": 258030078.0, "step": 6796 }, { "epoch": 0.8646482635796973, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0630083084106445, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8736092448234558, "num_tokens": 258064598.0, "step": 6797 }, { "epoch": 0.8647754738582878, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9474711418151855, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8660613298416138, "num_tokens": 258100121.0, "step": 6798 }, { "epoch": 0.8649026841368782, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2078304290771484, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.857695460319519, "num_tokens": 258138398.0, "step": 6799 }, { "epoch": 0.8650298944154687, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9767236709594727, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8609290719032288, "num_tokens": 258178730.0, "step": 6800 }, { "epoch": 0.8651571046940593, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3966822624206543, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8662737011909485, "num_tokens": 258211122.0, "step": 6801 }, { "epoch": 0.8652843149726498, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8467214107513428, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.868597686290741, "num_tokens": 258249753.0, "step": 6802 }, { "epoch": 0.8654115252512403, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.440948247909546, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8597397208213806, "num_tokens": 258287951.0, "step": 6803 }, { "epoch": 0.8655387355298308, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.677048921585083, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.8528602123260498, "num_tokens": 258333216.0, "step": 6804 }, { "epoch": 0.8656659458084213, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4232935905456543, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.863414466381073, "num_tokens": 258370380.0, "step": 6805 }, { "epoch": 0.8657931560870118, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.523355007171631, "learning_rate": 1e-05, "loss": 0.527, "mean_token_accuracy": 0.8490597009658813, "num_tokens": 258413730.0, "step": 6806 }, { "epoch": 0.8659203663656023, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3686776161193848, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8646519184112549, "num_tokens": 258451644.0, "step": 6807 }, { "epoch": 0.8660475766441929, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.415860414505005, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8746784925460815, "num_tokens": 258490084.0, "step": 6808 }, { "epoch": 0.8661747869227834, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.314154624938965, "learning_rate": 1e-05, "loss": 0.5333, "mean_token_accuracy": 0.8455185890197754, "num_tokens": 258530769.0, "step": 6809 }, { "epoch": 0.8663019972013739, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3704726696014404, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.8535671234130859, "num_tokens": 258567578.0, "step": 6810 }, { "epoch": 0.8664292074799644, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2532248497009277, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8624762892723083, "num_tokens": 258605498.0, "step": 6811 }, { "epoch": 0.8665564177585549, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.342489242553711, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8790194392204285, "num_tokens": 258646931.0, "step": 6812 }, { "epoch": 0.8666836280371454, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1855897903442383, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8678728342056274, "num_tokens": 258683853.0, "step": 6813 }, { "epoch": 0.8668108383157359, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3709933757781982, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8626988530158997, "num_tokens": 258723706.0, "step": 6814 }, { "epoch": 0.8669380485943264, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1282153129577637, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8660174608230591, "num_tokens": 258762404.0, "step": 6815 }, { "epoch": 0.867065258872917, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4424173831939697, "learning_rate": 1e-05, "loss": 0.5123, "mean_token_accuracy": 0.8505481481552124, "num_tokens": 258803209.0, "step": 6816 }, { "epoch": 0.8671924691515075, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.067427158355713, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.861748456954956, "num_tokens": 258840001.0, "step": 6817 }, { "epoch": 0.8673196794300979, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00016117095947265625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5420749187469482, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8803557753562927, "num_tokens": 258871401.0, "step": 6818 }, { "epoch": 0.8674468897086884, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0287013053894043, "learning_rate": 1e-05, "loss": 0.5151, "mean_token_accuracy": 0.8469822406768799, "num_tokens": 258902898.0, "step": 6819 }, { "epoch": 0.867574099987279, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.686093807220459, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8656854629516602, "num_tokens": 258942136.0, "step": 6820 }, { "epoch": 0.8677013102658695, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.998750925064087, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8627493381500244, "num_tokens": 258976599.0, "step": 6821 }, { "epoch": 0.86782852054446, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8860466480255127, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8589150309562683, "num_tokens": 259010457.0, "step": 6822 }, { "epoch": 0.8679557308230506, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.0979902744293213, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8621953129768372, "num_tokens": 259046368.0, "step": 6823 }, { "epoch": 0.868082941101641, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.004317760467529, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8686343431472778, "num_tokens": 259087240.0, "step": 6824 }, { "epoch": 0.8682101513802315, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.2997794151306152, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8633602857589722, "num_tokens": 259128297.0, "step": 6825 }, { "epoch": 0.868337361658822, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9635696411132812, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8604854345321655, "num_tokens": 259170242.0, "step": 6826 }, { "epoch": 0.8684645719374126, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.4553444385528564, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8600503206253052, "num_tokens": 259208580.0, "step": 6827 }, { "epoch": 0.8685917822160031, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7825517654418945, "learning_rate": 1e-05, "loss": 0.5497, "mean_token_accuracy": 0.8426575660705566, "num_tokens": 259249721.0, "step": 6828 }, { "epoch": 0.8687189924945936, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.4724748134613037, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8695831298828125, "num_tokens": 259292056.0, "step": 6829 }, { "epoch": 0.868846202773184, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5898754596710205, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8700234889984131, "num_tokens": 259323390.0, "step": 6830 }, { "epoch": 0.8689734130517746, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4215357303619385, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.873228907585144, "num_tokens": 259360728.0, "step": 6831 }, { "epoch": 0.8691006233303651, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.426239252090454, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8595837354660034, "num_tokens": 259397619.0, "step": 6832 }, { "epoch": 0.8692278336089556, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3766591548919678, "learning_rate": 1e-05, "loss": 0.5155, "mean_token_accuracy": 0.8481603264808655, "num_tokens": 259440049.0, "step": 6833 }, { "epoch": 0.8693550438875461, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3208048343658447, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8683924674987793, "num_tokens": 259475657.0, "step": 6834 }, { "epoch": 0.8694822541661367, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.352391481399536, "learning_rate": 1e-05, "loss": 0.54, "mean_token_accuracy": 0.8394544720649719, "num_tokens": 259511289.0, "step": 6835 }, { "epoch": 0.8696094644447271, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2492563724517822, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.875077486038208, "num_tokens": 259546704.0, "step": 6836 }, { "epoch": 0.8697366747233176, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3115527629852295, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8742258548736572, "num_tokens": 259588490.0, "step": 6837 }, { "epoch": 0.8698638850019081, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.225139856338501, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8630853891372681, "num_tokens": 259619838.0, "step": 6838 }, { "epoch": 0.8699910952804987, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2996389865875244, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8531836867332458, "num_tokens": 259662232.0, "step": 6839 }, { "epoch": 0.8701183055590892, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1961662769317627, "learning_rate": 1e-05, "loss": 0.54, "mean_token_accuracy": 0.8395413160324097, "num_tokens": 259699824.0, "step": 6840 }, { "epoch": 0.8702455158376797, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.318333148956299, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8610926866531372, "num_tokens": 259728217.0, "step": 6841 }, { "epoch": 0.8703727261162701, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.141798973083496, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.87624192237854, "num_tokens": 259763522.0, "step": 6842 }, { "epoch": 0.8704999363948607, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3535897731781006, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.850810706615448, "num_tokens": 259797650.0, "step": 6843 }, { "epoch": 0.8706271466734512, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.08292293548584, "learning_rate": 1e-05, "loss": 0.5121, "mean_token_accuracy": 0.8534336090087891, "num_tokens": 259837265.0, "step": 6844 }, { "epoch": 0.8707543569520417, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.429812431335449, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8648901581764221, "num_tokens": 259873040.0, "step": 6845 }, { "epoch": 0.8708815672306323, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0211493968963623, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8464510440826416, "num_tokens": 259907627.0, "step": 6846 }, { "epoch": 0.8710087775092228, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5596940517425537, "learning_rate": 1e-05, "loss": 0.5399, "mean_token_accuracy": 0.8428407907485962, "num_tokens": 259948213.0, "step": 6847 }, { "epoch": 0.8711359877878132, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.944597005844116, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8556343913078308, "num_tokens": 259985239.0, "step": 6848 }, { "epoch": 0.8712631980664037, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.789057493209839, "learning_rate": 1e-05, "loss": 0.5755, "mean_token_accuracy": 0.8339913487434387, "num_tokens": 260023046.0, "step": 6849 }, { "epoch": 0.8713904083449943, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 2.9320390224456787, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8720647692680359, "num_tokens": 260060716.0, "step": 6850 }, { "epoch": 0.8715176186235848, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.093111991882324, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8803915977478027, "num_tokens": 260093939.0, "step": 6851 }, { "epoch": 0.8716448289021753, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.1518185138702393, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8624245524406433, "num_tokens": 260129282.0, "step": 6852 }, { "epoch": 0.8717720391807658, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.251835823059082, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8763918280601501, "num_tokens": 260172167.0, "step": 6853 }, { "epoch": 0.8718992494593563, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.507753372192383, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8684779405593872, "num_tokens": 260212613.0, "step": 6854 }, { "epoch": 0.8720264597379468, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.130512237548828, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8684419989585876, "num_tokens": 260248774.0, "step": 6855 }, { "epoch": 0.8721536700165373, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6702721118927, "learning_rate": 1e-05, "loss": 0.5346, "mean_token_accuracy": 0.8440955877304077, "num_tokens": 260291105.0, "step": 6856 }, { "epoch": 0.8722808802951278, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8832859992980957, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8688076734542847, "num_tokens": 260332316.0, "step": 6857 }, { "epoch": 0.8724080905737184, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.598381757736206, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8718331456184387, "num_tokens": 260365016.0, "step": 6858 }, { "epoch": 0.8725353008523089, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7142460346221924, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8802128434181213, "num_tokens": 260396131.0, "step": 6859 }, { "epoch": 0.8726625111308994, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4847400188446045, "learning_rate": 1e-05, "loss": 0.5295, "mean_token_accuracy": 0.8479724526405334, "num_tokens": 260438053.0, "step": 6860 }, { "epoch": 0.8727897214094898, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5794448852539062, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8588377237319946, "num_tokens": 260479523.0, "step": 6861 }, { "epoch": 0.8729169316880804, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3658511638641357, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8693677186965942, "num_tokens": 260516623.0, "step": 6862 }, { "epoch": 0.8730441419666709, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5248358249664307, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.867998480796814, "num_tokens": 260552499.0, "step": 6863 }, { "epoch": 0.8731713522452614, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.244479179382324, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.863982081413269, "num_tokens": 260595636.0, "step": 6864 }, { "epoch": 0.873298562523852, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5391454696655273, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8676644563674927, "num_tokens": 260639787.0, "step": 6865 }, { "epoch": 0.8734257728024425, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.156566619873047, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8766297101974487, "num_tokens": 260674117.0, "step": 6866 }, { "epoch": 0.8735529830810329, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.645613431930542, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8626438975334167, "num_tokens": 260712757.0, "step": 6867 }, { "epoch": 0.8736801933596234, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.098344087600708, "learning_rate": 1e-05, "loss": 0.5251, "mean_token_accuracy": 0.8433148860931396, "num_tokens": 260750756.0, "step": 6868 }, { "epoch": 0.873807403638214, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.784364938735962, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8578594923019409, "num_tokens": 260782929.0, "step": 6869 }, { "epoch": 0.8739346139168045, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1042113304138184, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8621066808700562, "num_tokens": 260819762.0, "step": 6870 }, { "epoch": 0.874061824195395, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.929870843887329, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8556482791900635, "num_tokens": 260862543.0, "step": 6871 }, { "epoch": 0.8741890344739855, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.237395763397217, "learning_rate": 1e-05, "loss": 0.5172, "mean_token_accuracy": 0.8495911359786987, "num_tokens": 260899835.0, "step": 6872 }, { "epoch": 0.874316244752576, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.965599298477173, "learning_rate": 1e-05, "loss": 0.5268, "mean_token_accuracy": 0.8490843772888184, "num_tokens": 260940593.0, "step": 6873 }, { "epoch": 0.8744434550311665, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4200241565704346, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8612655401229858, "num_tokens": 260977454.0, "step": 6874 }, { "epoch": 0.874570665309757, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8077094554901123, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8628724813461304, "num_tokens": 261019095.0, "step": 6875 }, { "epoch": 0.8746978755883476, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4930989742279053, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8529751300811768, "num_tokens": 261056428.0, "step": 6876 }, { "epoch": 0.8748250858669381, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5830047130584717, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.873775839805603, "num_tokens": 261090679.0, "step": 6877 }, { "epoch": 0.8749522961455286, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4788763523101807, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8588942289352417, "num_tokens": 261130640.0, "step": 6878 }, { "epoch": 0.875079506424119, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3734171390533447, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8745717406272888, "num_tokens": 261169064.0, "step": 6879 }, { "epoch": 0.8752067167027096, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001621246337890625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4706997871398926, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8644998073577881, "num_tokens": 261205625.0, "step": 6880 }, { "epoch": 0.8753339269813001, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2022275924682617, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8697735071182251, "num_tokens": 261251859.0, "step": 6881 }, { "epoch": 0.8754611372598906, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.500558614730835, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8633046746253967, "num_tokens": 261289604.0, "step": 6882 }, { "epoch": 0.8755883475384811, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0977182388305664, "learning_rate": 1e-05, "loss": 0.3948, "mean_token_accuracy": 0.8842244148254395, "num_tokens": 261330053.0, "step": 6883 }, { "epoch": 0.8757155578170717, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5849292278289795, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8538638949394226, "num_tokens": 261371042.0, "step": 6884 }, { "epoch": 0.8758427680956621, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.039790391921997, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8538732528686523, "num_tokens": 261406911.0, "step": 6885 }, { "epoch": 0.8759699783742526, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7247443199157715, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8649380803108215, "num_tokens": 261443856.0, "step": 6886 }, { "epoch": 0.8760971886528431, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.02534556388855, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8567088842391968, "num_tokens": 261485987.0, "step": 6887 }, { "epoch": 0.8762243989314337, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.904054880142212, "learning_rate": 1e-05, "loss": 0.5042, "mean_token_accuracy": 0.8555516004562378, "num_tokens": 261519888.0, "step": 6888 }, { "epoch": 0.8763516092100242, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1390764713287354, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.86307692527771, "num_tokens": 261557435.0, "step": 6889 }, { "epoch": 0.8764788194886147, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.010942459106445, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8608141541481018, "num_tokens": 261598209.0, "step": 6890 }, { "epoch": 0.8766060297672051, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3418283462524414, "learning_rate": 1e-05, "loss": 0.5145, "mean_token_accuracy": 0.8494257926940918, "num_tokens": 261637972.0, "step": 6891 }, { "epoch": 0.8767332400457957, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9428088665008545, "learning_rate": 1e-05, "loss": 0.5251, "mean_token_accuracy": 0.8482949733734131, "num_tokens": 261676717.0, "step": 6892 }, { "epoch": 0.8768604503243862, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4750418663024902, "learning_rate": 1e-05, "loss": 0.5393, "mean_token_accuracy": 0.8430132865905762, "num_tokens": 261711434.0, "step": 6893 }, { "epoch": 0.8769876606029767, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7579574584960938, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8643385171890259, "num_tokens": 261749527.0, "step": 6894 }, { "epoch": 0.8771148708815673, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.450125217437744, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8708502054214478, "num_tokens": 261792031.0, "step": 6895 }, { "epoch": 0.8772420811601578, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.593789577484131, "learning_rate": 1e-05, "loss": 0.5409, "mean_token_accuracy": 0.8445353507995605, "num_tokens": 261831862.0, "step": 6896 }, { "epoch": 0.8773692914387482, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.40234112739563, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8672575950622559, "num_tokens": 261869949.0, "step": 6897 }, { "epoch": 0.8774965017173387, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4727935791015625, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8641748428344727, "num_tokens": 261904080.0, "step": 6898 }, { "epoch": 0.8776237119959293, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.358544111251831, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8497759699821472, "num_tokens": 261932809.0, "step": 6899 }, { "epoch": 0.8777509222745198, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.38161563873291, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8613096475601196, "num_tokens": 261974313.0, "step": 6900 }, { "epoch": 0.8778781325531103, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2781975269317627, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8756800293922424, "num_tokens": 262011505.0, "step": 6901 }, { "epoch": 0.8780053428317008, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.369259834289551, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8655097484588623, "num_tokens": 262042571.0, "step": 6902 }, { "epoch": 0.8781325531102913, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2064015865325928, "learning_rate": 1e-05, "loss": 0.5002, "mean_token_accuracy": 0.853089451789856, "num_tokens": 262088038.0, "step": 6903 }, { "epoch": 0.8782597633888818, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3812925815582275, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8709659576416016, "num_tokens": 262126662.0, "step": 6904 }, { "epoch": 0.8783869736674723, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1295621395111084, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8644551038742065, "num_tokens": 262170525.0, "step": 6905 }, { "epoch": 0.8785141839460628, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4692459106445312, "learning_rate": 1e-05, "loss": 0.5321, "mean_token_accuracy": 0.8489049077033997, "num_tokens": 262210524.0, "step": 6906 }, { "epoch": 0.8786413942246534, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0684611797332764, "learning_rate": 1e-05, "loss": 0.5415, "mean_token_accuracy": 0.8435506820678711, "num_tokens": 262248636.0, "step": 6907 }, { "epoch": 0.8787686045032439, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.583937406539917, "learning_rate": 1e-05, "loss": 0.5103, "mean_token_accuracy": 0.8509310483932495, "num_tokens": 262291795.0, "step": 6908 }, { "epoch": 0.8788958147818343, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.0070152282714844, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8656690716743469, "num_tokens": 262328484.0, "step": 6909 }, { "epoch": 0.8790230250604248, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7681117057800293, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.866798996925354, "num_tokens": 262362262.0, "step": 6910 }, { "epoch": 0.8791502353390154, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.02524995803833, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8542251586914062, "num_tokens": 262399082.0, "step": 6911 }, { "epoch": 0.8792774456176059, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.968688726425171, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8634169101715088, "num_tokens": 262434968.0, "step": 6912 }, { "epoch": 0.8794046558961964, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1812055110931396, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.866152286529541, "num_tokens": 262472989.0, "step": 6913 }, { "epoch": 0.879531866174787, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.034383773803711, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8627773523330688, "num_tokens": 262512526.0, "step": 6914 }, { "epoch": 0.8796590764533775, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.408327579498291, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8694916367530823, "num_tokens": 262548082.0, "step": 6915 }, { "epoch": 0.8797862867319679, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9116063117980957, "learning_rate": 1e-05, "loss": 0.5334, "mean_token_accuracy": 0.8456350564956665, "num_tokens": 262584786.0, "step": 6916 }, { "epoch": 0.8799134970105584, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5041961669921875, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8663411736488342, "num_tokens": 262625898.0, "step": 6917 }, { "epoch": 0.880040707289149, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6831486225128174, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8635170459747314, "num_tokens": 262664041.0, "step": 6918 }, { "epoch": 0.8801679175677395, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00016307830810546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.489907741546631, "learning_rate": 1e-05, "loss": 0.5191, "mean_token_accuracy": 0.8524753451347351, "num_tokens": 262702597.0, "step": 6919 }, { "epoch": 0.88029512784633, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.493385076522827, "learning_rate": 1e-05, "loss": 0.5388, "mean_token_accuracy": 0.843705415725708, "num_tokens": 262734951.0, "step": 6920 }, { "epoch": 0.8804223381249205, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4418177604675293, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8690793514251709, "num_tokens": 262771513.0, "step": 6921 }, { "epoch": 0.880549548403511, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.334346294403076, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8548824191093445, "num_tokens": 262813045.0, "step": 6922 }, { "epoch": 0.8806767586821015, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.406355619430542, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8710719347000122, "num_tokens": 262846663.0, "step": 6923 }, { "epoch": 0.880803968960692, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2470383644104004, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8630739450454712, "num_tokens": 262880056.0, "step": 6924 }, { "epoch": 0.8809311792392825, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3883771896362305, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8696053624153137, "num_tokens": 262914819.0, "step": 6925 }, { "epoch": 0.8810583895178731, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1715035438537598, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.857514500617981, "num_tokens": 262953167.0, "step": 6926 }, { "epoch": 0.8811855997964636, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.4133124351501465, "learning_rate": 1e-05, "loss": 0.4961, "mean_token_accuracy": 0.8564214706420898, "num_tokens": 262991628.0, "step": 6927 }, { "epoch": 0.881312810075054, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1135551929473877, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8661540746688843, "num_tokens": 263029287.0, "step": 6928 }, { "epoch": 0.8814400203536445, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4688758850097656, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8615579605102539, "num_tokens": 263068982.0, "step": 6929 }, { "epoch": 0.8815672306322351, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.0557990074157715, "learning_rate": 1e-05, "loss": 0.5566, "mean_token_accuracy": 0.8393364548683167, "num_tokens": 263111260.0, "step": 6930 }, { "epoch": 0.8816944409108256, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5707571506500244, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8643057346343994, "num_tokens": 263149310.0, "step": 6931 }, { "epoch": 0.8818216511894161, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.99341082572937, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.873613715171814, "num_tokens": 263188440.0, "step": 6932 }, { "epoch": 0.8819488614680067, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7528209686279297, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.8525094389915466, "num_tokens": 263229079.0, "step": 6933 }, { "epoch": 0.8820760717465971, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.014211654663086, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8553590774536133, "num_tokens": 263260591.0, "step": 6934 }, { "epoch": 0.8822032820251876, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9461328983306885, "learning_rate": 1e-05, "loss": 0.5193, "mean_token_accuracy": 0.8507028818130493, "num_tokens": 263301256.0, "step": 6935 }, { "epoch": 0.8823304923037781, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1234662532806396, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8672543168067932, "num_tokens": 263337600.0, "step": 6936 }, { "epoch": 0.8824577025823687, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.069058895111084, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8570327758789062, "num_tokens": 263383124.0, "step": 6937 }, { "epoch": 0.8825849128609592, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3626904487609863, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.86976158618927, "num_tokens": 263416047.0, "step": 6938 }, { "epoch": 0.8827121231395497, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.004221439361572, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8753206729888916, "num_tokens": 263458515.0, "step": 6939 }, { "epoch": 0.8828393334181401, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.501645565032959, "learning_rate": 1e-05, "loss": 0.5144, "mean_token_accuracy": 0.8535618782043457, "num_tokens": 263499833.0, "step": 6940 }, { "epoch": 0.8829665436967307, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.826425313949585, "learning_rate": 1e-05, "loss": 0.49, "mean_token_accuracy": 0.8589056730270386, "num_tokens": 263531680.0, "step": 6941 }, { "epoch": 0.8830937539753212, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5047998428344727, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8491990566253662, "num_tokens": 263565889.0, "step": 6942 }, { "epoch": 0.8832209642539117, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6331570148468018, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8542959094047546, "num_tokens": 263604974.0, "step": 6943 }, { "epoch": 0.8833481745325023, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.437548875808716, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.862302303314209, "num_tokens": 263646960.0, "step": 6944 }, { "epoch": 0.8834753848110928, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.489093065261841, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8603290319442749, "num_tokens": 263686427.0, "step": 6945 }, { "epoch": 0.8836025950896832, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3798394203186035, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8714779615402222, "num_tokens": 263717520.0, "step": 6946 }, { "epoch": 0.8837298053682737, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3938663005828857, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8589835166931152, "num_tokens": 263756226.0, "step": 6947 }, { "epoch": 0.8838570156468643, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.305603504180908, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8739425539970398, "num_tokens": 263793961.0, "step": 6948 }, { "epoch": 0.8839842259254548, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3599836826324463, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8583918213844299, "num_tokens": 263827325.0, "step": 6949 }, { "epoch": 0.8841114362040453, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2465832233428955, "learning_rate": 1e-05, "loss": 0.5332, "mean_token_accuracy": 0.8401886224746704, "num_tokens": 263862854.0, "step": 6950 }, { "epoch": 0.8842386464826358, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3430051803588867, "learning_rate": 1e-05, "loss": 0.5564, "mean_token_accuracy": 0.8363834619522095, "num_tokens": 263907035.0, "step": 6951 }, { "epoch": 0.8843658567612263, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.18515682220459, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8507552742958069, "num_tokens": 263949550.0, "step": 6952 }, { "epoch": 0.8844930670398168, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.375032663345337, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8618643283843994, "num_tokens": 263988966.0, "step": 6953 }, { "epoch": 0.8846202773184073, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1355676651000977, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8751776814460754, "num_tokens": 264023442.0, "step": 6954 }, { "epoch": 0.8847474875969978, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.422633171081543, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.870193362236023, "num_tokens": 264064185.0, "step": 6955 }, { "epoch": 0.8848746978755884, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.063288450241089, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8532122373580933, "num_tokens": 264107455.0, "step": 6956 }, { "epoch": 0.8850019081541789, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.537076234817505, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8546298146247864, "num_tokens": 264143090.0, "step": 6957 }, { "epoch": 0.8851291184327693, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0021371841430664, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8578468561172485, "num_tokens": 264180002.0, "step": 6958 }, { "epoch": 0.8852563287113598, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.727123498916626, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8700786232948303, "num_tokens": 264215444.0, "step": 6959 }, { "epoch": 0.8853835389899504, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.9732046127319336, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8612269759178162, "num_tokens": 264250758.0, "step": 6960 }, { "epoch": 0.8855107492685409, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9810163974761963, "learning_rate": 1e-05, "loss": 0.5338, "mean_token_accuracy": 0.8472921848297119, "num_tokens": 264290181.0, "step": 6961 }, { "epoch": 0.8856379595471314, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.115732431411743, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8628524541854858, "num_tokens": 264327480.0, "step": 6962 }, { "epoch": 0.885765169825722, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.136952877044678, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.863261342048645, "num_tokens": 264368762.0, "step": 6963 }, { "epoch": 0.8858923801043125, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.390270233154297, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8618654608726501, "num_tokens": 264410261.0, "step": 6964 }, { "epoch": 0.8860195903829029, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.084788799285889, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8670759201049805, "num_tokens": 264449111.0, "step": 6965 }, { "epoch": 0.8861468006614934, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.580932140350342, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.8569105863571167, "num_tokens": 264487045.0, "step": 6966 }, { "epoch": 0.886274010940084, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8431670665740967, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8641844987869263, "num_tokens": 264524025.0, "step": 6967 }, { "epoch": 0.8864012212186745, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5592265129089355, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8542712926864624, "num_tokens": 264561361.0, "step": 6968 }, { "epoch": 0.886528431497265, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6423521041870117, "learning_rate": 1e-05, "loss": 0.5585, "mean_token_accuracy": 0.835952639579773, "num_tokens": 264603557.0, "step": 6969 }, { "epoch": 0.8866556417758555, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4749832153320312, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.873950719833374, "num_tokens": 264638539.0, "step": 6970 }, { "epoch": 0.886782852054446, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.494152307510376, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8629342317581177, "num_tokens": 264676592.0, "step": 6971 }, { "epoch": 0.8869100623330365, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3813037872314453, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8721574544906616, "num_tokens": 264708339.0, "step": 6972 }, { "epoch": 0.887037272611627, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4187166690826416, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8611573576927185, "num_tokens": 264741705.0, "step": 6973 }, { "epoch": 0.8871644828902175, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3073670864105225, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8602898120880127, "num_tokens": 264773095.0, "step": 6974 }, { "epoch": 0.8872916931688081, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3683786392211914, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8787733316421509, "num_tokens": 264813499.0, "step": 6975 }, { "epoch": 0.8874189034473986, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2216436862945557, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8683799505233765, "num_tokens": 264851314.0, "step": 6976 }, { "epoch": 0.887546113725989, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.41154146194458, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.863785982131958, "num_tokens": 264886908.0, "step": 6977 }, { "epoch": 0.8876733240045795, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1402323246002197, "learning_rate": 1e-05, "loss": 0.5508, "mean_token_accuracy": 0.8347434401512146, "num_tokens": 264927516.0, "step": 6978 }, { "epoch": 0.8878005342831701, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.463029384613037, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8678631782531738, "num_tokens": 264966758.0, "step": 6979 }, { "epoch": 0.8879277445617606, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.067084312438965, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8524233102798462, "num_tokens": 265002267.0, "step": 6980 }, { "epoch": 0.8880549548403511, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.58976149559021, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8695294857025146, "num_tokens": 265034957.0, "step": 6981 }, { "epoch": 0.8881821651189417, "ewc_loss": 0.057861328125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.9908840656280518, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8624662160873413, "num_tokens": 265075587.0, "step": 6982 }, { "epoch": 0.8883093753975321, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.791618585586548, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8733454346656799, "num_tokens": 265111385.0, "step": 6983 }, { "epoch": 0.8884365856761226, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0078954696655273, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8713189959526062, "num_tokens": 265153674.0, "step": 6984 }, { "epoch": 0.8885637959547131, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.013515472412109, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8745220899581909, "num_tokens": 265190168.0, "step": 6985 }, { "epoch": 0.8886910062333037, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.196399211883545, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8735960125923157, "num_tokens": 265230145.0, "step": 6986 }, { "epoch": 0.8888182165118942, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.110288619995117, "learning_rate": 1e-05, "loss": 0.5247, "mean_token_accuracy": 0.8494905233383179, "num_tokens": 265263229.0, "step": 6987 }, { "epoch": 0.8889454267904847, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.460946559906006, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8539109230041504, "num_tokens": 265305426.0, "step": 6988 }, { "epoch": 0.8890726370690751, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9549808502197266, "learning_rate": 1e-05, "loss": 0.5121, "mean_token_accuracy": 0.8529173731803894, "num_tokens": 265345677.0, "step": 6989 }, { "epoch": 0.8891998473476657, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5466794967651367, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8705629110336304, "num_tokens": 265383987.0, "step": 6990 }, { "epoch": 0.8893270576262562, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.745468854904175, "learning_rate": 1e-05, "loss": 0.5352, "mean_token_accuracy": 0.8435885310173035, "num_tokens": 265423209.0, "step": 6991 }, { "epoch": 0.8894542679048467, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.508620500564575, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.8518175482749939, "num_tokens": 265463304.0, "step": 6992 }, { "epoch": 0.8895814781834372, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5445849895477295, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.8519427180290222, "num_tokens": 265506991.0, "step": 6993 }, { "epoch": 0.8897086884620278, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.451287269592285, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8633220791816711, "num_tokens": 265539050.0, "step": 6994 }, { "epoch": 0.8898358987406182, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.396390438079834, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8597457408905029, "num_tokens": 265583593.0, "step": 6995 }, { "epoch": 0.8899631090192087, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.385160207748413, "learning_rate": 1e-05, "loss": 0.5305, "mean_token_accuracy": 0.843646764755249, "num_tokens": 265623873.0, "step": 6996 }, { "epoch": 0.8900903192977992, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.316037178039551, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8561681509017944, "num_tokens": 265666191.0, "step": 6997 }, { "epoch": 0.8902175295763898, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.33166241645813, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8594859838485718, "num_tokens": 265703929.0, "step": 6998 }, { "epoch": 0.8903447398549803, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.28334641456604, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8630048632621765, "num_tokens": 265737246.0, "step": 6999 }, { "epoch": 0.8904719501335708, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.294954538345337, "learning_rate": 1e-05, "loss": 0.5182, "mean_token_accuracy": 0.8523663282394409, "num_tokens": 265774549.0, "step": 7000 }, { "epoch": 0.8905991604121613, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.255765914916992, "learning_rate": 1e-05, "loss": 0.5064, "mean_token_accuracy": 0.8548891544342041, "num_tokens": 265806972.0, "step": 7001 }, { "epoch": 0.8907263706907518, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2748470306396484, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8532745838165283, "num_tokens": 265844505.0, "step": 7002 }, { "epoch": 0.8908535809693423, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.220332145690918, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8628893494606018, "num_tokens": 265885350.0, "step": 7003 }, { "epoch": 0.8909807912479328, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.256129741668701, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8654886484146118, "num_tokens": 265925345.0, "step": 7004 }, { "epoch": 0.8911080015265234, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.211812734603882, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8669224381446838, "num_tokens": 265961329.0, "step": 7005 }, { "epoch": 0.8912352118051139, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.270772933959961, "learning_rate": 1e-05, "loss": 0.5601, "mean_token_accuracy": 0.834571361541748, "num_tokens": 266000836.0, "step": 7006 }, { "epoch": 0.8913624220837043, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.188849449157715, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.855758786201477, "num_tokens": 266037178.0, "step": 7007 }, { "epoch": 0.8914896323622948, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2760157585144043, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8693010210990906, "num_tokens": 266077125.0, "step": 7008 }, { "epoch": 0.8916168426408854, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.153667688369751, "learning_rate": 1e-05, "loss": 0.5111, "mean_token_accuracy": 0.8452798128128052, "num_tokens": 266117313.0, "step": 7009 }, { "epoch": 0.8917440529194759, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.308067560195923, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8612009882926941, "num_tokens": 266157432.0, "step": 7010 }, { "epoch": 0.8918712631980664, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.1078267097473145, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8598729372024536, "num_tokens": 266197170.0, "step": 7011 }, { "epoch": 0.891998473476657, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3832435607910156, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.8538616895675659, "num_tokens": 266230383.0, "step": 7012 }, { "epoch": 0.8921256837552475, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 3.0391485691070557, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8627786636352539, "num_tokens": 266268685.0, "step": 7013 }, { "epoch": 0.8922528940338379, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.501201629638672, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8572370409965515, "num_tokens": 266308875.0, "step": 7014 }, { "epoch": 0.8923801043124284, "ewc_loss": 0.056884765625, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.0531158447265625e-05, "grad_norm": 2.9696152210235596, "learning_rate": 1e-05, "loss": 0.5509, "mean_token_accuracy": 0.8350878953933716, "num_tokens": 266346540.0, "step": 7015 }, { "epoch": 0.892507314591019, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.709625244140625, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8614456057548523, "num_tokens": 266383359.0, "step": 7016 }, { "epoch": 0.8926345248696095, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.903639793395996, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8550524115562439, "num_tokens": 266420235.0, "step": 7017 }, { "epoch": 0.8927617351482, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.026207447052002, "learning_rate": 1e-05, "loss": 0.4969, "mean_token_accuracy": 0.8574607968330383, "num_tokens": 266455097.0, "step": 7018 }, { "epoch": 0.8928889454267905, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.0107104778289795, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8648272752761841, "num_tokens": 266488471.0, "step": 7019 }, { "epoch": 0.893016155705381, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.345321178436279, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8563752174377441, "num_tokens": 266528107.0, "step": 7020 }, { "epoch": 0.8931433659839715, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.4604790210723877, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8679231405258179, "num_tokens": 266567859.0, "step": 7021 }, { "epoch": 0.893270576262562, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2958149909973145, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8685241937637329, "num_tokens": 266609722.0, "step": 7022 }, { "epoch": 0.8933977865411525, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.755830764770508, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8724746704101562, "num_tokens": 266650546.0, "step": 7023 }, { "epoch": 0.8935249968197431, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.026676654815674, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.858710527420044, "num_tokens": 266686508.0, "step": 7024 }, { "epoch": 0.8936522070983336, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7129321098327637, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8753739595413208, "num_tokens": 266725528.0, "step": 7025 }, { "epoch": 0.893779417376924, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000164031982421875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8348639011383057, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8749793767929077, "num_tokens": 266754818.0, "step": 7026 }, { "epoch": 0.8939066276555145, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5784029960632324, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8584206104278564, "num_tokens": 266790723.0, "step": 7027 }, { "epoch": 0.8940338379341051, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.685216188430786, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.864587664604187, "num_tokens": 266824889.0, "step": 7028 }, { "epoch": 0.8941610482126956, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4294557571411133, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8603262305259705, "num_tokens": 266866129.0, "step": 7029 }, { "epoch": 0.8942882584912861, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.600398302078247, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.858816385269165, "num_tokens": 266910005.0, "step": 7030 }, { "epoch": 0.8944154687698767, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3080646991729736, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8687728047370911, "num_tokens": 266947366.0, "step": 7031 }, { "epoch": 0.8945426790484671, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.60913348197937, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8577073812484741, "num_tokens": 266982402.0, "step": 7032 }, { "epoch": 0.8946698893270576, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2226898670196533, "learning_rate": 1e-05, "loss": 0.5151, "mean_token_accuracy": 0.850285530090332, "num_tokens": 267015397.0, "step": 7033 }, { "epoch": 0.8947970996056481, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6680264472961426, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8741945028305054, "num_tokens": 267050204.0, "step": 7034 }, { "epoch": 0.8949243098842387, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.125063896179199, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8765280246734619, "num_tokens": 267085127.0, "step": 7035 }, { "epoch": 0.8950515201628292, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.820256471633911, "learning_rate": 1e-05, "loss": 0.5186, "mean_token_accuracy": 0.8515197038650513, "num_tokens": 267122210.0, "step": 7036 }, { "epoch": 0.8951787304414197, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1445963382720947, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8598240613937378, "num_tokens": 267161124.0, "step": 7037 }, { "epoch": 0.8953059407200101, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9885919094085693, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8709249496459961, "num_tokens": 267195339.0, "step": 7038 }, { "epoch": 0.8954331509986007, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.311936855316162, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8557273149490356, "num_tokens": 267228712.0, "step": 7039 }, { "epoch": 0.8955603612771912, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9747228622436523, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8516321778297424, "num_tokens": 267262796.0, "step": 7040 }, { "epoch": 0.8956875715557817, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.4662699699401855, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.854263186454773, "num_tokens": 267304101.0, "step": 7041 }, { "epoch": 0.8958147818343722, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7934165000915527, "learning_rate": 1e-05, "loss": 0.5383, "mean_token_accuracy": 0.8455957770347595, "num_tokens": 267348943.0, "step": 7042 }, { "epoch": 0.8959419921129628, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5084686279296875, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8673154711723328, "num_tokens": 267390436.0, "step": 7043 }, { "epoch": 0.8960692023915532, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.578230857849121, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8729712963104248, "num_tokens": 267428069.0, "step": 7044 }, { "epoch": 0.8961964126701437, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4933085441589355, "learning_rate": 1e-05, "loss": 0.5132, "mean_token_accuracy": 0.8472387790679932, "num_tokens": 267468029.0, "step": 7045 }, { "epoch": 0.8963236229487342, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.384948492050171, "learning_rate": 1e-05, "loss": 0.4985, "mean_token_accuracy": 0.8553961515426636, "num_tokens": 267504013.0, "step": 7046 }, { "epoch": 0.8964508332273248, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4700968265533447, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8695049285888672, "num_tokens": 267537696.0, "step": 7047 }, { "epoch": 0.8965780435059153, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2461228370666504, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8691593408584595, "num_tokens": 267568971.0, "step": 7048 }, { "epoch": 0.8967052537845058, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4608354568481445, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8577069044113159, "num_tokens": 267607829.0, "step": 7049 }, { "epoch": 0.8968324640630962, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1459662914276123, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8645660877227783, "num_tokens": 267649338.0, "step": 7050 }, { "epoch": 0.8969596743416868, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.502471923828125, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8686326742172241, "num_tokens": 267685053.0, "step": 7051 }, { "epoch": 0.8970868846202773, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0747017860412598, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8678838610649109, "num_tokens": 267722291.0, "step": 7052 }, { "epoch": 0.8972140948988678, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.6140940189361572, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8631953001022339, "num_tokens": 267760533.0, "step": 7053 }, { "epoch": 0.8973413051774584, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.019277572631836, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8641107082366943, "num_tokens": 267796655.0, "step": 7054 }, { "epoch": 0.8974685154560489, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7700750827789307, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8654392957687378, "num_tokens": 267835950.0, "step": 7055 }, { "epoch": 0.8975957257346393, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0346758365631104, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8655110597610474, "num_tokens": 267871796.0, "step": 7056 }, { "epoch": 0.8977229360132298, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9570865631103516, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8650588989257812, "num_tokens": 267910562.0, "step": 7057 }, { "epoch": 0.8978501462918204, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.1689884662628174, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8635951280593872, "num_tokens": 267948962.0, "step": 7058 }, { "epoch": 0.8979773565704109, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.050345420837402, "learning_rate": 1e-05, "loss": 0.5199, "mean_token_accuracy": 0.8539721965789795, "num_tokens": 267993158.0, "step": 7059 }, { "epoch": 0.8981045668490014, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.396338939666748, "learning_rate": 1e-05, "loss": 0.5128, "mean_token_accuracy": 0.8528621196746826, "num_tokens": 268033366.0, "step": 7060 }, { "epoch": 0.898231777127592, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.945742130279541, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8629406690597534, "num_tokens": 268074125.0, "step": 7061 }, { "epoch": 0.8983589874061825, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5033204555511475, "learning_rate": 1e-05, "loss": 0.503, "mean_token_accuracy": 0.8579120635986328, "num_tokens": 268112132.0, "step": 7062 }, { "epoch": 0.8984861976847729, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7559359073638916, "learning_rate": 1e-05, "loss": 0.5509, "mean_token_accuracy": 0.8393961191177368, "num_tokens": 268148161.0, "step": 7063 }, { "epoch": 0.8986134079633634, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4840805530548096, "learning_rate": 1e-05, "loss": 0.4023, "mean_token_accuracy": 0.8856149911880493, "num_tokens": 268184977.0, "step": 7064 }, { "epoch": 0.898740618241954, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5740926265716553, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.861611545085907, "num_tokens": 268219355.0, "step": 7065 }, { "epoch": 0.8988678285205445, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4264163970947266, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8650397658348083, "num_tokens": 268260533.0, "step": 7066 }, { "epoch": 0.898995038799135, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4451534748077393, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.856156587600708, "num_tokens": 268294415.0, "step": 7067 }, { "epoch": 0.8991222490777255, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.365981340408325, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.869243860244751, "num_tokens": 268331038.0, "step": 7068 }, { "epoch": 0.899249459356316, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.36548113822937, "learning_rate": 1e-05, "loss": 0.5195, "mean_token_accuracy": 0.8447924852371216, "num_tokens": 268369438.0, "step": 7069 }, { "epoch": 0.8993766696349065, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3104772567749023, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8633794784545898, "num_tokens": 268403808.0, "step": 7070 }, { "epoch": 0.899503879913497, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3068761825561523, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8748968839645386, "num_tokens": 268442575.0, "step": 7071 }, { "epoch": 0.8996310901920875, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.271657705307007, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.859592616558075, "num_tokens": 268475387.0, "step": 7072 }, { "epoch": 0.8997583004706781, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2928218841552734, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8608145713806152, "num_tokens": 268513514.0, "step": 7073 }, { "epoch": 0.8998855107492686, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2213940620422363, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8663270473480225, "num_tokens": 268553690.0, "step": 7074 }, { "epoch": 0.900012721027859, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.296736478805542, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8698862195014954, "num_tokens": 268593285.0, "step": 7075 }, { "epoch": 0.9001399313064495, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1847915649414062, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8634582757949829, "num_tokens": 268631929.0, "step": 7076 }, { "epoch": 0.9002671415850401, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.333296298980713, "learning_rate": 1e-05, "loss": 0.5419, "mean_token_accuracy": 0.835323691368103, "num_tokens": 268669882.0, "step": 7077 }, { "epoch": 0.9003943518636306, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.129552125930786, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8471198678016663, "num_tokens": 268712709.0, "step": 7078 }, { "epoch": 0.9005215621422211, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3896594047546387, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8595482110977173, "num_tokens": 268751514.0, "step": 7079 }, { "epoch": 0.9006487724208116, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.05627703666687, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.847232460975647, "num_tokens": 268796841.0, "step": 7080 }, { "epoch": 0.9007759826994021, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.501990795135498, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8712742924690247, "num_tokens": 268841536.0, "step": 7081 }, { "epoch": 0.9009031929779926, "ewc_loss": 0.05712890625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.076957702636719e-05, "grad_norm": 2.9742650985717773, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8603963851928711, "num_tokens": 268884636.0, "step": 7082 }, { "epoch": 0.9010304032565831, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7062430381774902, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8576043844223022, "num_tokens": 268924114.0, "step": 7083 }, { "epoch": 0.9011576135351737, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.923327684402466, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8685888051986694, "num_tokens": 268966476.0, "step": 7084 }, { "epoch": 0.9012848238137642, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 4.009256839752197, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8715932369232178, "num_tokens": 268998224.0, "step": 7085 }, { "epoch": 0.9014120340923547, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.078042507171631, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.855421245098114, "num_tokens": 269043552.0, "step": 7086 }, { "epoch": 0.9015392443709451, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.229953765869141, "learning_rate": 1e-05, "loss": 0.5295, "mean_token_accuracy": 0.8483874797821045, "num_tokens": 269085180.0, "step": 7087 }, { "epoch": 0.9016664546495357, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.431432008743286, "learning_rate": 1e-05, "loss": 0.524, "mean_token_accuracy": 0.8497889637947083, "num_tokens": 269121694.0, "step": 7088 }, { "epoch": 0.9017936649281262, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.158045768737793, "learning_rate": 1e-05, "loss": 0.5544, "mean_token_accuracy": 0.8449611663818359, "num_tokens": 269164896.0, "step": 7089 }, { "epoch": 0.9019208752067167, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.632025957107544, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8821573853492737, "num_tokens": 269201095.0, "step": 7090 }, { "epoch": 0.9020480854853072, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.939868211746216, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8645629286766052, "num_tokens": 269243055.0, "step": 7091 }, { "epoch": 0.9021752957638978, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6048197746276855, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8702722787857056, "num_tokens": 269279902.0, "step": 7092 }, { "epoch": 0.9023025060424882, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.752009868621826, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8770877718925476, "num_tokens": 269315773.0, "step": 7093 }, { "epoch": 0.9024297163210787, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5090558528900146, "learning_rate": 1e-05, "loss": 0.54, "mean_token_accuracy": 0.8441020846366882, "num_tokens": 269353127.0, "step": 7094 }, { "epoch": 0.9025569265996692, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6183149814605713, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8634384274482727, "num_tokens": 269387023.0, "step": 7095 }, { "epoch": 0.9026841368782598, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.405520439147949, "learning_rate": 1e-05, "loss": 0.5424, "mean_token_accuracy": 0.8425730466842651, "num_tokens": 269419983.0, "step": 7096 }, { "epoch": 0.9028113471568503, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.532667636871338, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8676953315734863, "num_tokens": 269460580.0, "step": 7097 }, { "epoch": 0.9029385574354408, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2872941493988037, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8770225644111633, "num_tokens": 269497711.0, "step": 7098 }, { "epoch": 0.9030657677140312, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.516578197479248, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.862274706363678, "num_tokens": 269538111.0, "step": 7099 }, { "epoch": 0.9031929779926218, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.196897506713867, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8811469078063965, "num_tokens": 269570695.0, "step": 7100 }, { "epoch": 0.9033201882712123, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.567864418029785, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8638555407524109, "num_tokens": 269605849.0, "step": 7101 }, { "epoch": 0.9034473985498028, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1142828464508057, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.857930064201355, "num_tokens": 269646445.0, "step": 7102 }, { "epoch": 0.9035746088283934, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.683823585510254, "learning_rate": 1e-05, "loss": 0.5296, "mean_token_accuracy": 0.847050666809082, "num_tokens": 269686806.0, "step": 7103 }, { "epoch": 0.9037018191069839, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.076531410217285, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8521518707275391, "num_tokens": 269727311.0, "step": 7104 }, { "epoch": 0.9038290293855743, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.843820333480835, "learning_rate": 1e-05, "loss": 0.5532, "mean_token_accuracy": 0.8420079946517944, "num_tokens": 269765066.0, "step": 7105 }, { "epoch": 0.9039562396641648, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1317906379699707, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8539921045303345, "num_tokens": 269807065.0, "step": 7106 }, { "epoch": 0.9040834499427554, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9710288047790527, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8763283491134644, "num_tokens": 269839874.0, "step": 7107 }, { "epoch": 0.9042106602213459, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.308159828186035, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8708447217941284, "num_tokens": 269877647.0, "step": 7108 }, { "epoch": 0.9043378704999364, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.938689947128296, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8627522587776184, "num_tokens": 269912403.0, "step": 7109 }, { "epoch": 0.9044650807785269, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.462193012237549, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8693901300430298, "num_tokens": 269946086.0, "step": 7110 }, { "epoch": 0.9045922910571175, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.739671230316162, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.873315691947937, "num_tokens": 269985820.0, "step": 7111 }, { "epoch": 0.9047195013357079, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.498708724975586, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8699321746826172, "num_tokens": 270021377.0, "step": 7112 }, { "epoch": 0.9048467116142984, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5289363861083984, "learning_rate": 1e-05, "loss": 0.5315, "mean_token_accuracy": 0.8481981754302979, "num_tokens": 270056928.0, "step": 7113 }, { "epoch": 0.9049739218928889, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.466836929321289, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8724981546401978, "num_tokens": 270093720.0, "step": 7114 }, { "epoch": 0.9051011321714795, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3320322036743164, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.8541509509086609, "num_tokens": 270140151.0, "step": 7115 }, { "epoch": 0.90522834245007, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4661002159118652, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8648648262023926, "num_tokens": 270173970.0, "step": 7116 }, { "epoch": 0.9053555527286605, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2088747024536133, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8634372353553772, "num_tokens": 270207014.0, "step": 7117 }, { "epoch": 0.905482763007251, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4898009300231934, "learning_rate": 1e-05, "loss": 0.5253, "mean_token_accuracy": 0.8474690914154053, "num_tokens": 270246635.0, "step": 7118 }, { "epoch": 0.9056099732858415, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.103973388671875, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.858018696308136, "num_tokens": 270289902.0, "step": 7119 }, { "epoch": 0.905737183564432, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.575690269470215, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.8480045795440674, "num_tokens": 270321173.0, "step": 7120 }, { "epoch": 0.9058643938430225, "ewc_loss": 0.05859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.036303997039795, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8716745376586914, "num_tokens": 270360281.0, "step": 7121 }, { "epoch": 0.9059916041216131, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.710475444793701, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8693061470985413, "num_tokens": 270396609.0, "step": 7122 }, { "epoch": 0.9061188144002036, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0284006595611572, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8492647409439087, "num_tokens": 270436365.0, "step": 7123 }, { "epoch": 0.906246024678794, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8819386959075928, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8662572503089905, "num_tokens": 270471314.0, "step": 7124 }, { "epoch": 0.9063732349573845, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.121509313583374, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.852637767791748, "num_tokens": 270510169.0, "step": 7125 }, { "epoch": 0.9065004452359751, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9911651611328125, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8693035840988159, "num_tokens": 270549787.0, "step": 7126 }, { "epoch": 0.9066276555145656, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3148202896118164, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8699386119842529, "num_tokens": 270584059.0, "step": 7127 }, { "epoch": 0.9067548657931561, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.94227933883667, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8688119649887085, "num_tokens": 270618059.0, "step": 7128 }, { "epoch": 0.9068820760717466, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4352433681488037, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.86141037940979, "num_tokens": 270656968.0, "step": 7129 }, { "epoch": 0.9070092863503371, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.793205738067627, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8617634177207947, "num_tokens": 270692088.0, "step": 7130 }, { "epoch": 0.9071364966289276, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.463183879852295, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8683748245239258, "num_tokens": 270729480.0, "step": 7131 }, { "epoch": 0.9072637069075181, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.612079620361328, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8709129095077515, "num_tokens": 270769062.0, "step": 7132 }, { "epoch": 0.9073909171861086, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4074184894561768, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8572306632995605, "num_tokens": 270808142.0, "step": 7133 }, { "epoch": 0.9075181274646992, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4879331588745117, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8581424355506897, "num_tokens": 270842232.0, "step": 7134 }, { "epoch": 0.9076453377432897, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.354097366333008, "learning_rate": 1e-05, "loss": 0.5177, "mean_token_accuracy": 0.8507635593414307, "num_tokens": 270881588.0, "step": 7135 }, { "epoch": 0.9077725480218801, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.403860569000244, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8566901683807373, "num_tokens": 270921122.0, "step": 7136 }, { "epoch": 0.9078997583004706, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2965762615203857, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.8570749163627625, "num_tokens": 270959037.0, "step": 7137 }, { "epoch": 0.9080269685790612, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.355165719985962, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.86482173204422, "num_tokens": 271002731.0, "step": 7138 }, { "epoch": 0.9081541788576517, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.236253023147583, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8648080229759216, "num_tokens": 271038487.0, "step": 7139 }, { "epoch": 0.9082813891362422, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3739142417907715, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8596560955047607, "num_tokens": 271068811.0, "step": 7140 }, { "epoch": 0.9084085994148328, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.166459560394287, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8746258616447449, "num_tokens": 271111019.0, "step": 7141 }, { "epoch": 0.9085358096934232, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4116880893707275, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8586956262588501, "num_tokens": 271149317.0, "step": 7142 }, { "epoch": 0.9086630199720137, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.098193645477295, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.8803572654724121, "num_tokens": 271190664.0, "step": 7143 }, { "epoch": 0.9087902302506042, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5070230960845947, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8669372797012329, "num_tokens": 271225071.0, "step": 7144 }, { "epoch": 0.9089174405291948, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.0368618965148926, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8623415231704712, "num_tokens": 271264228.0, "step": 7145 }, { "epoch": 0.9090446508077853, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6610147953033447, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8627192974090576, "num_tokens": 271300888.0, "step": 7146 }, { "epoch": 0.9091718610863758, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.9945452213287354, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8620607256889343, "num_tokens": 271339629.0, "step": 7147 }, { "epoch": 0.9092990713649662, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8691024780273438, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8640136122703552, "num_tokens": 271379116.0, "step": 7148 }, { "epoch": 0.9094262816435568, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.0806996822357178, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.850896418094635, "num_tokens": 271420829.0, "step": 7149 }, { "epoch": 0.9095534919221473, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.02528715133667, "learning_rate": 1e-05, "loss": 0.488, "mean_token_accuracy": 0.8594260215759277, "num_tokens": 271457934.0, "step": 7150 }, { "epoch": 0.9096807022007378, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.295363426208496, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8623815774917603, "num_tokens": 271497441.0, "step": 7151 }, { "epoch": 0.9098079124793284, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9939286708831787, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.8554197549819946, "num_tokens": 271532358.0, "step": 7152 }, { "epoch": 0.9099351227579189, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.468991756439209, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8677334785461426, "num_tokens": 271570103.0, "step": 7153 }, { "epoch": 0.9100623330365093, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.818986654281616, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8664906024932861, "num_tokens": 271604542.0, "step": 7154 }, { "epoch": 0.9101895433150998, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5019338130950928, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8516062498092651, "num_tokens": 271646901.0, "step": 7155 }, { "epoch": 0.9103167535936904, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6117889881134033, "learning_rate": 1e-05, "loss": 0.5296, "mean_token_accuracy": 0.8434505462646484, "num_tokens": 271684922.0, "step": 7156 }, { "epoch": 0.9104439638722809, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4540185928344727, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8794995546340942, "num_tokens": 271721206.0, "step": 7157 }, { "epoch": 0.9105711741508714, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4418721199035645, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8739508390426636, "num_tokens": 271757978.0, "step": 7158 }, { "epoch": 0.9106983844294619, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.421978235244751, "learning_rate": 1e-05, "loss": 0.5632, "mean_token_accuracy": 0.836293637752533, "num_tokens": 271792208.0, "step": 7159 }, { "epoch": 0.9108255947080524, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3310601711273193, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.856289803981781, "num_tokens": 271828853.0, "step": 7160 }, { "epoch": 0.9109528049866429, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3766987323760986, "learning_rate": 1e-05, "loss": 0.5094, "mean_token_accuracy": 0.8477203845977783, "num_tokens": 271868596.0, "step": 7161 }, { "epoch": 0.9110800152652334, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2465906143188477, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8677293062210083, "num_tokens": 271911292.0, "step": 7162 }, { "epoch": 0.9112072255438239, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.350212812423706, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8697284460067749, "num_tokens": 271954512.0, "step": 7163 }, { "epoch": 0.9113344358224145, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.206599235534668, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8567394614219666, "num_tokens": 271994229.0, "step": 7164 }, { "epoch": 0.911461646101005, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3517231941223145, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8677091598510742, "num_tokens": 272031712.0, "step": 7165 }, { "epoch": 0.9115888563795955, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1575748920440674, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8578979969024658, "num_tokens": 272070269.0, "step": 7166 }, { "epoch": 0.9117160666581859, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.378612756729126, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8613332509994507, "num_tokens": 272115407.0, "step": 7167 }, { "epoch": 0.9118432769367765, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.099748373031616, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8618941307067871, "num_tokens": 272161463.0, "step": 7168 }, { "epoch": 0.911970487215367, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4585938453674316, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8673871159553528, "num_tokens": 272196712.0, "step": 7169 }, { "epoch": 0.9120976974939575, "ewc_loss": 0.0576171875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0440168380737305, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8569158315658569, "num_tokens": 272235617.0, "step": 7170 }, { "epoch": 0.912224907772548, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5932722091674805, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8656073212623596, "num_tokens": 272266464.0, "step": 7171 }, { "epoch": 0.9123521180511386, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.9833242893218994, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.849550724029541, "num_tokens": 272301069.0, "step": 7172 }, { "epoch": 0.912479328329729, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.798759698867798, "learning_rate": 1e-05, "loss": 0.5367, "mean_token_accuracy": 0.8453733921051025, "num_tokens": 272333119.0, "step": 7173 }, { "epoch": 0.9126065386083195, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.9839799404144287, "learning_rate": 1e-05, "loss": 0.5033, "mean_token_accuracy": 0.8494091629981995, "num_tokens": 272365560.0, "step": 7174 }, { "epoch": 0.9127337488869101, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.032345294952393, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8805792927742004, "num_tokens": 272401290.0, "step": 7175 }, { "epoch": 0.9128609591655006, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.138739824295044, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8660533428192139, "num_tokens": 272442925.0, "step": 7176 }, { "epoch": 0.9129881694440911, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.193435192108154, "learning_rate": 1e-05, "loss": 0.5127, "mean_token_accuracy": 0.8552436828613281, "num_tokens": 272477192.0, "step": 7177 }, { "epoch": 0.9131153797226816, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.4792637825012207, "learning_rate": 1e-05, "loss": 0.5345, "mean_token_accuracy": 0.8450549840927124, "num_tokens": 272506987.0, "step": 7178 }, { "epoch": 0.9132425900012721, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0501790046691895, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8685009479522705, "num_tokens": 272544399.0, "step": 7179 }, { "epoch": 0.9133698002798626, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5927672386169434, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.861370861530304, "num_tokens": 272577639.0, "step": 7180 }, { "epoch": 0.9134970105584531, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.840867280960083, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8675140142440796, "num_tokens": 272615358.0, "step": 7181 }, { "epoch": 0.9136242208370436, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5456817150115967, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8595684170722961, "num_tokens": 272652927.0, "step": 7182 }, { "epoch": 0.9137514311156342, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6751351356506348, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8556603193283081, "num_tokens": 272682509.0, "step": 7183 }, { "epoch": 0.9138786413942247, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.458251476287842, "learning_rate": 1e-05, "loss": 0.526, "mean_token_accuracy": 0.8464917540550232, "num_tokens": 272723295.0, "step": 7184 }, { "epoch": 0.9140058516728151, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5345237255096436, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8481345772743225, "num_tokens": 272764985.0, "step": 7185 }, { "epoch": 0.9141330619514056, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3538897037506104, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.864025354385376, "num_tokens": 272807926.0, "step": 7186 }, { "epoch": 0.9142602722299962, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4736647605895996, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8662002086639404, "num_tokens": 272844249.0, "step": 7187 }, { "epoch": 0.9143874825085867, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2828917503356934, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8603447675704956, "num_tokens": 272878029.0, "step": 7188 }, { "epoch": 0.9145146927871772, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.459160327911377, "learning_rate": 1e-05, "loss": 0.5093, "mean_token_accuracy": 0.849172055721283, "num_tokens": 272913185.0, "step": 7189 }, { "epoch": 0.9146419030657678, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.194624423980713, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8668546676635742, "num_tokens": 272946835.0, "step": 7190 }, { "epoch": 0.9147691133443582, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.492699384689331, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.8547271490097046, "num_tokens": 272983467.0, "step": 7191 }, { "epoch": 0.9148963236229487, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1172995567321777, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8631932139396667, "num_tokens": 273022464.0, "step": 7192 }, { "epoch": 0.9150235339015392, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.58811354637146, "learning_rate": 1e-05, "loss": 0.527, "mean_token_accuracy": 0.8445330858230591, "num_tokens": 273059746.0, "step": 7193 }, { "epoch": 0.9151507441801298, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.046093702316284, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8714815378189087, "num_tokens": 273100815.0, "step": 7194 }, { "epoch": 0.9152779544587203, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7629458904266357, "learning_rate": 1e-05, "loss": 0.5115, "mean_token_accuracy": 0.8514220118522644, "num_tokens": 273139429.0, "step": 7195 }, { "epoch": 0.9154051647373108, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0622987747192383, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8689028024673462, "num_tokens": 273174159.0, "step": 7196 }, { "epoch": 0.9155323750159012, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9326775074005127, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8696876764297485, "num_tokens": 273211434.0, "step": 7197 }, { "epoch": 0.9156595852944918, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.202792167663574, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.853841245174408, "num_tokens": 273243463.0, "step": 7198 }, { "epoch": 0.9157867955730823, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.986759901046753, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.865032970905304, "num_tokens": 273283515.0, "step": 7199 }, { "epoch": 0.9159140058516728, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.578519821166992, "learning_rate": 1e-05, "loss": 0.5136, "mean_token_accuracy": 0.8540492057800293, "num_tokens": 273322927.0, "step": 7200 }, { "epoch": 0.9160412161302633, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6985695362091064, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8699393272399902, "num_tokens": 273359157.0, "step": 7201 }, { "epoch": 0.9161684264088539, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.426825761795044, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8659083843231201, "num_tokens": 273394385.0, "step": 7202 }, { "epoch": 0.9162956366874443, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.505704402923584, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8539424538612366, "num_tokens": 273430286.0, "step": 7203 }, { "epoch": 0.9164228469660348, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.410142183303833, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8701252937316895, "num_tokens": 273467821.0, "step": 7204 }, { "epoch": 0.9165500572446253, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016498565673828125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3519203662872314, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8559733629226685, "num_tokens": 273507659.0, "step": 7205 }, { "epoch": 0.9166772675232159, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.405013084411621, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.8486189842224121, "num_tokens": 273546445.0, "step": 7206 }, { "epoch": 0.9168044778018064, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.247509717941284, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8559524416923523, "num_tokens": 273580354.0, "step": 7207 }, { "epoch": 0.9169316880803969, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.40389084815979, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8601225018501282, "num_tokens": 273619194.0, "step": 7208 }, { "epoch": 0.9170588983589874, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.156921863555908, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8680548071861267, "num_tokens": 273662888.0, "step": 7209 }, { "epoch": 0.9171861086375779, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.452648401260376, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8655931949615479, "num_tokens": 273697823.0, "step": 7210 }, { "epoch": 0.9173133189161684, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.081493616104126, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.8502360582351685, "num_tokens": 273737541.0, "step": 7211 }, { "epoch": 0.9174405291947589, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5456743240356445, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.8548030257225037, "num_tokens": 273777565.0, "step": 7212 }, { "epoch": 0.9175677394733495, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.027317523956299, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8492689728736877, "num_tokens": 273814269.0, "step": 7213 }, { "epoch": 0.91769494975194, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.682957410812378, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8619229197502136, "num_tokens": 273855828.0, "step": 7214 }, { "epoch": 0.9178221600305305, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.009648323059082, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8602471351623535, "num_tokens": 273887514.0, "step": 7215 }, { "epoch": 0.9179493703091209, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8800594806671143, "learning_rate": 1e-05, "loss": 0.5091, "mean_token_accuracy": 0.8522099256515503, "num_tokens": 273923134.0, "step": 7216 }, { "epoch": 0.9180765805877115, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.0615689754486084, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8663730621337891, "num_tokens": 273962089.0, "step": 7217 }, { "epoch": 0.918203790866302, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.058883190155029, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8632090091705322, "num_tokens": 273997150.0, "step": 7218 }, { "epoch": 0.9183310011448925, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.278168201446533, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8783091902732849, "num_tokens": 274032144.0, "step": 7219 }, { "epoch": 0.918458211423483, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.0833001136779785, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8771851658821106, "num_tokens": 274063344.0, "step": 7220 }, { "epoch": 0.9185854217020736, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5477137565612793, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8684189915657043, "num_tokens": 274102242.0, "step": 7221 }, { "epoch": 0.918712631980664, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8565962314605713, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8628783226013184, "num_tokens": 274146560.0, "step": 7222 }, { "epoch": 0.9188398422592545, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.527657985687256, "learning_rate": 1e-05, "loss": 0.5176, "mean_token_accuracy": 0.8488839268684387, "num_tokens": 274180793.0, "step": 7223 }, { "epoch": 0.918967052537845, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6521193981170654, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8642032146453857, "num_tokens": 274221207.0, "step": 7224 }, { "epoch": 0.9190942628164356, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.446887731552124, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8550254106521606, "num_tokens": 274265413.0, "step": 7225 }, { "epoch": 0.9192214730950261, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.511533260345459, "learning_rate": 1e-05, "loss": 0.5539, "mean_token_accuracy": 0.8370794057846069, "num_tokens": 274305875.0, "step": 7226 }, { "epoch": 0.9193486833736166, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.391570568084717, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8567858934402466, "num_tokens": 274344223.0, "step": 7227 }, { "epoch": 0.919475893652207, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.405790328979492, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.858873188495636, "num_tokens": 274382671.0, "step": 7228 }, { "epoch": 0.9196031039307976, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.333050489425659, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8611571192741394, "num_tokens": 274413475.0, "step": 7229 }, { "epoch": 0.9197303142093881, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3469512462615967, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8785644173622131, "num_tokens": 274453714.0, "step": 7230 }, { "epoch": 0.9198575244879786, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2786483764648438, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8551808595657349, "num_tokens": 274490388.0, "step": 7231 }, { "epoch": 0.9199847347665692, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3194546699523926, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8649550080299377, "num_tokens": 274529939.0, "step": 7232 }, { "epoch": 0.9201119450451597, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.222014904022217, "learning_rate": 1e-05, "loss": 0.492, "mean_token_accuracy": 0.8592115640640259, "num_tokens": 274570967.0, "step": 7233 }, { "epoch": 0.9202391553237501, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.344318151473999, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.8647493124008179, "num_tokens": 274601977.0, "step": 7234 }, { "epoch": 0.9203663656023406, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1610615253448486, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8679341673851013, "num_tokens": 274638509.0, "step": 7235 }, { "epoch": 0.9204935758809312, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3799140453338623, "learning_rate": 1e-05, "loss": 0.3968, "mean_token_accuracy": 0.8826697468757629, "num_tokens": 274674325.0, "step": 7236 }, { "epoch": 0.9206207861595217, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1004817485809326, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8629888296127319, "num_tokens": 274709686.0, "step": 7237 }, { "epoch": 0.9207479964381122, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4743335247039795, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8562423586845398, "num_tokens": 274748917.0, "step": 7238 }, { "epoch": 0.9208752067167028, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.0164802074432373, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.850637674331665, "num_tokens": 274794119.0, "step": 7239 }, { "epoch": 0.9210024169952932, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.7699389457702637, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8510351181030273, "num_tokens": 274827907.0, "step": 7240 }, { "epoch": 0.9211296272738837, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.941662549972534, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.882292628288269, "num_tokens": 274865821.0, "step": 7241 }, { "epoch": 0.9212568375524742, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001659393310546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.008296489715576, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.865312933921814, "num_tokens": 274903550.0, "step": 7242 }, { "epoch": 0.9213840478310648, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.073348045349121, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8637126684188843, "num_tokens": 274944212.0, "step": 7243 }, { "epoch": 0.9215112581096553, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.258055210113525, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8629717826843262, "num_tokens": 274980175.0, "step": 7244 }, { "epoch": 0.9216384683882458, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.457420587539673, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8562695384025574, "num_tokens": 275020018.0, "step": 7245 }, { "epoch": 0.9217656786668362, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1679511070251465, "learning_rate": 1e-05, "loss": 0.5036, "mean_token_accuracy": 0.8572136759757996, "num_tokens": 275063538.0, "step": 7246 }, { "epoch": 0.9218928889454268, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6721785068511963, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8577711582183838, "num_tokens": 275099884.0, "step": 7247 }, { "epoch": 0.9220200992240173, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.923062801361084, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8549617528915405, "num_tokens": 275140292.0, "step": 7248 }, { "epoch": 0.9221473095026078, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6306703090667725, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8638185262680054, "num_tokens": 275178480.0, "step": 7249 }, { "epoch": 0.9222745197811983, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7184722423553467, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8560319542884827, "num_tokens": 275215855.0, "step": 7250 }, { "epoch": 0.9224017300597889, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5211048126220703, "learning_rate": 1e-05, "loss": 0.5301, "mean_token_accuracy": 0.8465942144393921, "num_tokens": 275254895.0, "step": 7251 }, { "epoch": 0.9225289403383793, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.580503463745117, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8615604639053345, "num_tokens": 275292352.0, "step": 7252 }, { "epoch": 0.9226561506169698, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.393007278442383, "learning_rate": 1e-05, "loss": 0.509, "mean_token_accuracy": 0.8555281758308411, "num_tokens": 275335867.0, "step": 7253 }, { "epoch": 0.9227833608955603, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5070106983184814, "learning_rate": 1e-05, "loss": 0.5033, "mean_token_accuracy": 0.8562148809432983, "num_tokens": 275376551.0, "step": 7254 }, { "epoch": 0.9229105711741509, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2908787727355957, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8552566766738892, "num_tokens": 275412568.0, "step": 7255 }, { "epoch": 0.9230377814527414, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5000908374786377, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8612204790115356, "num_tokens": 275450534.0, "step": 7256 }, { "epoch": 0.9231649917313319, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.184385299682617, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8571450710296631, "num_tokens": 275493165.0, "step": 7257 }, { "epoch": 0.9232922020099223, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.559098958969116, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8630092144012451, "num_tokens": 275533935.0, "step": 7258 }, { "epoch": 0.9234194122885129, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.102473020553589, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8557605743408203, "num_tokens": 275570937.0, "step": 7259 }, { "epoch": 0.9235466225671034, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.694078207015991, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8679922819137573, "num_tokens": 275606036.0, "step": 7260 }, { "epoch": 0.9236738328456939, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0524637699127197, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8651513457298279, "num_tokens": 275645849.0, "step": 7261 }, { "epoch": 0.9238010431242845, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.881044626235962, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8744967579841614, "num_tokens": 275682192.0, "step": 7262 }, { "epoch": 0.923928253402875, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1345345973968506, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.853564977645874, "num_tokens": 275722074.0, "step": 7263 }, { "epoch": 0.9240554636814655, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.0244317054748535, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.855802059173584, "num_tokens": 275764966.0, "step": 7264 }, { "epoch": 0.9241826739600559, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3323588371276855, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8730937838554382, "num_tokens": 275805085.0, "step": 7265 }, { "epoch": 0.9243098842386465, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.980370044708252, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8666031360626221, "num_tokens": 275837533.0, "step": 7266 }, { "epoch": 0.924437094517237, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5078282356262207, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8607074618339539, "num_tokens": 275877168.0, "step": 7267 }, { "epoch": 0.9245643047958275, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7609500885009766, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8522156476974487, "num_tokens": 275917861.0, "step": 7268 }, { "epoch": 0.924691515074418, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.537101984024048, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.849023699760437, "num_tokens": 275952433.0, "step": 7269 }, { "epoch": 0.9248187253530086, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.527674674987793, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8616547584533691, "num_tokens": 275993124.0, "step": 7270 }, { "epoch": 0.924945935631599, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.488662004470825, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8689978122711182, "num_tokens": 276032967.0, "step": 7271 }, { "epoch": 0.9250731459101895, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.349318504333496, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8650416135787964, "num_tokens": 276068840.0, "step": 7272 }, { "epoch": 0.92520035618878, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4700913429260254, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8568881750106812, "num_tokens": 276109196.0, "step": 7273 }, { "epoch": 0.9253275664673706, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.219594955444336, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.869247555732727, "num_tokens": 276151544.0, "step": 7274 }, { "epoch": 0.9254547767459611, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.478085994720459, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8585643768310547, "num_tokens": 276192849.0, "step": 7275 }, { "epoch": 0.9255819870245516, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1442291736602783, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8730517625808716, "num_tokens": 276226543.0, "step": 7276 }, { "epoch": 0.925709197303142, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5271666049957275, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.867953896522522, "num_tokens": 276267199.0, "step": 7277 }, { "epoch": 0.9258364075817326, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.0720629692077637, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8563435673713684, "num_tokens": 276306786.0, "step": 7278 }, { "epoch": 0.9259636178603231, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.651496410369873, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.8612923622131348, "num_tokens": 276343479.0, "step": 7279 }, { "epoch": 0.9260908281389136, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.041883945465088, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8586241006851196, "num_tokens": 276378067.0, "step": 7280 }, { "epoch": 0.9262180384175042, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.812079906463623, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8615339994430542, "num_tokens": 276407610.0, "step": 7281 }, { "epoch": 0.9263452486960947, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.062758445739746, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.8522635698318481, "num_tokens": 276443664.0, "step": 7282 }, { "epoch": 0.9264724589746851, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.962893009185791, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8658175468444824, "num_tokens": 276480446.0, "step": 7283 }, { "epoch": 0.9265996692532756, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.207077980041504, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8684800267219543, "num_tokens": 276515012.0, "step": 7284 }, { "epoch": 0.9267268795318662, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.016225337982178, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8810902833938599, "num_tokens": 276555804.0, "step": 7285 }, { "epoch": 0.9268540898104567, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3941733837127686, "learning_rate": 1e-05, "loss": 0.5176, "mean_token_accuracy": 0.8487445116043091, "num_tokens": 276596290.0, "step": 7286 }, { "epoch": 0.9269813000890472, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9049181938171387, "learning_rate": 1e-05, "loss": 0.5176, "mean_token_accuracy": 0.8524153232574463, "num_tokens": 276641767.0, "step": 7287 }, { "epoch": 0.9271085103676378, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.472234010696411, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8754163980484009, "num_tokens": 276678375.0, "step": 7288 }, { "epoch": 0.9272357206462282, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7303175926208496, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8614393472671509, "num_tokens": 276715835.0, "step": 7289 }, { "epoch": 0.9273629309248187, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4648523330688477, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8586702346801758, "num_tokens": 276750323.0, "step": 7290 }, { "epoch": 0.9274901412034092, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001678466796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5607683658599854, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8629330396652222, "num_tokens": 276786867.0, "step": 7291 }, { "epoch": 0.9276173514819998, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.42145037651062, "learning_rate": 1e-05, "loss": 0.5522, "mean_token_accuracy": 0.8389682769775391, "num_tokens": 276831263.0, "step": 7292 }, { "epoch": 0.9277445617605903, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4209539890289307, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8663675785064697, "num_tokens": 276867573.0, "step": 7293 }, { "epoch": 0.9278717720391808, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.354423999786377, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8813656568527222, "num_tokens": 276900414.0, "step": 7294 }, { "epoch": 0.9279989823177712, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.356031656265259, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8555630445480347, "num_tokens": 276938336.0, "step": 7295 }, { "epoch": 0.9281261925963618, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.29089617729187, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8695571422576904, "num_tokens": 276981277.0, "step": 7296 }, { "epoch": 0.9282534028749523, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.322848320007324, "learning_rate": 1e-05, "loss": 0.5204, "mean_token_accuracy": 0.8550841808319092, "num_tokens": 277023346.0, "step": 7297 }, { "epoch": 0.9283806131535428, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2658209800720215, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8629257678985596, "num_tokens": 277054821.0, "step": 7298 }, { "epoch": 0.9285078234321333, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3101794719696045, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8635439872741699, "num_tokens": 277092124.0, "step": 7299 }, { "epoch": 0.9286350337107239, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.215660810470581, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8605650663375854, "num_tokens": 277130119.0, "step": 7300 }, { "epoch": 0.9287622439893143, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.320827007293701, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8791604042053223, "num_tokens": 277167777.0, "step": 7301 }, { "epoch": 0.9288894542679048, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.159964084625244, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8700767755508423, "num_tokens": 277207273.0, "step": 7302 }, { "epoch": 0.9290166645464953, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3652279376983643, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8709802031517029, "num_tokens": 277248996.0, "step": 7303 }, { "epoch": 0.9291438748250859, "ewc_loss": 0.058837890625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.123758316040039, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.8468281030654907, "num_tokens": 277286212.0, "step": 7304 }, { "epoch": 0.9292710851036764, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4297358989715576, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8437142372131348, "num_tokens": 277330482.0, "step": 7305 }, { "epoch": 0.9293982953822669, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.0570907592773438, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8686777353286743, "num_tokens": 277364965.0, "step": 7306 }, { "epoch": 0.9295255056608573, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.564319372177124, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8632981777191162, "num_tokens": 277392227.0, "step": 7307 }, { "epoch": 0.9296527159394479, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.967271566390991, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8706662654876709, "num_tokens": 277435658.0, "step": 7308 }, { "epoch": 0.9297799262180384, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7887935638427734, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8602348566055298, "num_tokens": 277472513.0, "step": 7309 }, { "epoch": 0.9299071364966289, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.9527111053466797, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.8498725891113281, "num_tokens": 277517740.0, "step": 7310 }, { "epoch": 0.9300343467752195, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0864176750183105, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8770747184753418, "num_tokens": 277547304.0, "step": 7311 }, { "epoch": 0.93016155705381, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.208862543106079, "learning_rate": 1e-05, "loss": 0.524, "mean_token_accuracy": 0.8495200872421265, "num_tokens": 277581380.0, "step": 7312 }, { "epoch": 0.9302887673324005, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.194343566894531, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.868840217590332, "num_tokens": 277621615.0, "step": 7313 }, { "epoch": 0.9304159776109909, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4777188301086426, "learning_rate": 1e-05, "loss": 0.5178, "mean_token_accuracy": 0.8510957956314087, "num_tokens": 277655157.0, "step": 7314 }, { "epoch": 0.9305431878895815, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.098850250244141, "learning_rate": 1e-05, "loss": 0.5341, "mean_token_accuracy": 0.8476302623748779, "num_tokens": 277688634.0, "step": 7315 }, { "epoch": 0.930670398168172, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6090445518493652, "learning_rate": 1e-05, "loss": 0.5391, "mean_token_accuracy": 0.8495772480964661, "num_tokens": 277736123.0, "step": 7316 }, { "epoch": 0.9307976084467625, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.878788471221924, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8608007431030273, "num_tokens": 277776197.0, "step": 7317 }, { "epoch": 0.930924818725353, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.559063673019409, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8726294040679932, "num_tokens": 277820477.0, "step": 7318 }, { "epoch": 0.9310520290039436, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7012417316436768, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8634785413742065, "num_tokens": 277864660.0, "step": 7319 }, { "epoch": 0.931179239282534, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.467698335647583, "learning_rate": 1e-05, "loss": 0.5188, "mean_token_accuracy": 0.8478054404258728, "num_tokens": 277905702.0, "step": 7320 }, { "epoch": 0.9313064495611245, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.572834014892578, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8654672503471375, "num_tokens": 277946033.0, "step": 7321 }, { "epoch": 0.931433659839715, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3715555667877197, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8622534275054932, "num_tokens": 277981556.0, "step": 7322 }, { "epoch": 0.9315608701183056, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.508105754852295, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8746938705444336, "num_tokens": 278021455.0, "step": 7323 }, { "epoch": 0.9316880803968961, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2742042541503906, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8683645725250244, "num_tokens": 278057840.0, "step": 7324 }, { "epoch": 0.9318152906754866, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.515510082244873, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8569551110267639, "num_tokens": 278093186.0, "step": 7325 }, { "epoch": 0.931942500954077, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1885626316070557, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.8611272573471069, "num_tokens": 278127235.0, "step": 7326 }, { "epoch": 0.9320697112326676, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.56382155418396, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8637871146202087, "num_tokens": 278165236.0, "step": 7327 }, { "epoch": 0.9321969215112581, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1130475997924805, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8490939140319824, "num_tokens": 278205153.0, "step": 7328 }, { "epoch": 0.9323241317898486, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6714136600494385, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8750373125076294, "num_tokens": 278240808.0, "step": 7329 }, { "epoch": 0.9324513420684392, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.071744680404663, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8538098931312561, "num_tokens": 278281364.0, "step": 7330 }, { "epoch": 0.9325785523470297, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8377389907836914, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8676854372024536, "num_tokens": 278316021.0, "step": 7331 }, { "epoch": 0.9327057626256201, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.13936448097229, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8563648462295532, "num_tokens": 278357744.0, "step": 7332 }, { "epoch": 0.9328329729042106, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9554760456085205, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8614546656608582, "num_tokens": 278390826.0, "step": 7333 }, { "epoch": 0.9329601831828012, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.312871217727661, "learning_rate": 1e-05, "loss": 0.527, "mean_token_accuracy": 0.8456100225448608, "num_tokens": 278424395.0, "step": 7334 }, { "epoch": 0.9330873934613917, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8903706073760986, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8748024702072144, "num_tokens": 278462180.0, "step": 7335 }, { "epoch": 0.9332146037399822, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.435206413269043, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8574142456054688, "num_tokens": 278499349.0, "step": 7336 }, { "epoch": 0.9333418140185727, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7346177101135254, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.850528359413147, "num_tokens": 278530782.0, "step": 7337 }, { "epoch": 0.9334690242971632, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4833643436431885, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8561787009239197, "num_tokens": 278563262.0, "step": 7338 }, { "epoch": 0.9335962345757537, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5163793563842773, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8621623516082764, "num_tokens": 278598882.0, "step": 7339 }, { "epoch": 0.9337234448543442, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.479919672012329, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8628213405609131, "num_tokens": 278628535.0, "step": 7340 }, { "epoch": 0.9338506551329347, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3266658782958984, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.857758104801178, "num_tokens": 278667732.0, "step": 7341 }, { "epoch": 0.9339778654115253, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4594194889068604, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8642856478691101, "num_tokens": 278706925.0, "step": 7342 }, { "epoch": 0.9341050756901158, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2629311084747314, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8739482164382935, "num_tokens": 278740491.0, "step": 7343 }, { "epoch": 0.9342322859687062, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4442708492279053, "learning_rate": 1e-05, "loss": 0.5078, "mean_token_accuracy": 0.8511651158332825, "num_tokens": 278776739.0, "step": 7344 }, { "epoch": 0.9343594962472968, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1495416164398193, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8712514042854309, "num_tokens": 278813019.0, "step": 7345 }, { "epoch": 0.9344867065258873, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4778265953063965, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8615677356719971, "num_tokens": 278853807.0, "step": 7346 }, { "epoch": 0.9346139168044778, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.079516887664795, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8791791200637817, "num_tokens": 278891643.0, "step": 7347 }, { "epoch": 0.9347411270830683, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5975332260131836, "learning_rate": 1e-05, "loss": 0.5796, "mean_token_accuracy": 0.8419739603996277, "num_tokens": 278929915.0, "step": 7348 }, { "epoch": 0.9348683373616589, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0400826930999756, "learning_rate": 1e-05, "loss": 0.5194, "mean_token_accuracy": 0.8441965579986572, "num_tokens": 278964400.0, "step": 7349 }, { "epoch": 0.9349955476402493, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.727200508117676, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.8634809255599976, "num_tokens": 279003322.0, "step": 7350 }, { "epoch": 0.9351227579188398, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0270533561706543, "learning_rate": 1e-05, "loss": 0.506, "mean_token_accuracy": 0.8507095575332642, "num_tokens": 279045630.0, "step": 7351 }, { "epoch": 0.9352499681974303, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9008853435516357, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8613787889480591, "num_tokens": 279085330.0, "step": 7352 }, { "epoch": 0.9353771784760209, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.116018295288086, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8678927421569824, "num_tokens": 279125499.0, "step": 7353 }, { "epoch": 0.9355043887546114, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.049431800842285, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8646494150161743, "num_tokens": 279154772.0, "step": 7354 }, { "epoch": 0.9356315990332019, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.322838306427002, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.857524037361145, "num_tokens": 279196947.0, "step": 7355 }, { "epoch": 0.9357588093117923, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.000498294830322, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8781838417053223, "num_tokens": 279236872.0, "step": 7356 }, { "epoch": 0.9358860195903829, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.478060722351074, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8692513108253479, "num_tokens": 279275456.0, "step": 7357 }, { "epoch": 0.9360132298689734, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8405423164367676, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.8581384420394897, "num_tokens": 279308406.0, "step": 7358 }, { "epoch": 0.9361404401475639, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.492799758911133, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.8593810796737671, "num_tokens": 279349522.0, "step": 7359 }, { "epoch": 0.9362676504261545, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6444907188415527, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8728850483894348, "num_tokens": 279388242.0, "step": 7360 }, { "epoch": 0.936394860704745, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4560937881469727, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8753776550292969, "num_tokens": 279421239.0, "step": 7361 }, { "epoch": 0.9365220709833355, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4945740699768066, "learning_rate": 1e-05, "loss": 0.5517, "mean_token_accuracy": 0.8378349542617798, "num_tokens": 279464823.0, "step": 7362 }, { "epoch": 0.9366492812619259, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.395298480987549, "learning_rate": 1e-05, "loss": 0.5143, "mean_token_accuracy": 0.8500155210494995, "num_tokens": 279507645.0, "step": 7363 }, { "epoch": 0.9367764915405165, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.390524387359619, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8584139943122864, "num_tokens": 279539905.0, "step": 7364 }, { "epoch": 0.936903701819107, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.333735704421997, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8700543642044067, "num_tokens": 279577929.0, "step": 7365 }, { "epoch": 0.9370309120976975, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3294644355773926, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8645192384719849, "num_tokens": 279618421.0, "step": 7366 }, { "epoch": 0.937158122376288, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.287104368209839, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8685719966888428, "num_tokens": 279656563.0, "step": 7367 }, { "epoch": 0.9372853326548786, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3081042766571045, "learning_rate": 1e-05, "loss": 0.4941, "mean_token_accuracy": 0.853797197341919, "num_tokens": 279693159.0, "step": 7368 }, { "epoch": 0.937412542933469, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.233546733856201, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8756522536277771, "num_tokens": 279731492.0, "step": 7369 }, { "epoch": 0.9375397532120595, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3043713569641113, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.867127001285553, "num_tokens": 279774906.0, "step": 7370 }, { "epoch": 0.93766696349065, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2149369716644287, "learning_rate": 1e-05, "loss": 0.5384, "mean_token_accuracy": 0.8408002257347107, "num_tokens": 279809797.0, "step": 7371 }, { "epoch": 0.9377941737692406, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3063740730285645, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8667210936546326, "num_tokens": 279850107.0, "step": 7372 }, { "epoch": 0.9379213840478311, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.17204213142395, "learning_rate": 1e-05, "loss": 0.393, "mean_token_accuracy": 0.8855258226394653, "num_tokens": 279885115.0, "step": 7373 }, { "epoch": 0.9380485943264216, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3487133979797363, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8694334030151367, "num_tokens": 279921343.0, "step": 7374 }, { "epoch": 0.938175804605012, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1155202388763428, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8729526996612549, "num_tokens": 279953059.0, "step": 7375 }, { "epoch": 0.9383030148836026, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00016880035400390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4317946434020996, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8708153963088989, "num_tokens": 279987993.0, "step": 7376 }, { "epoch": 0.9384302251621931, "ewc_loss": 0.058349609375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.053323745727539, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8549548387527466, "num_tokens": 280024620.0, "step": 7377 }, { "epoch": 0.9385574354407836, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5446009635925293, "learning_rate": 1e-05, "loss": 0.5355, "mean_token_accuracy": 0.8437952995300293, "num_tokens": 280069011.0, "step": 7378 }, { "epoch": 0.9386846457193742, "ewc_loss": 0.05810546875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.100799560546875e-05, "grad_norm": 2.963892936706543, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8817751407623291, "num_tokens": 280105877.0, "step": 7379 }, { "epoch": 0.9388118559979647, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.760275363922119, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.87359219789505, "num_tokens": 280143155.0, "step": 7380 }, { "epoch": 0.9389390662765551, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.935133695602417, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8713390231132507, "num_tokens": 280181490.0, "step": 7381 }, { "epoch": 0.9390662765551456, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.068694114685059, "learning_rate": 1e-05, "loss": 0.5229, "mean_token_accuracy": 0.8515151739120483, "num_tokens": 280221959.0, "step": 7382 }, { "epoch": 0.9391934868337362, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.108858823776245, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8815177083015442, "num_tokens": 280265224.0, "step": 7383 }, { "epoch": 0.9393206971123267, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.289198875427246, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8666241765022278, "num_tokens": 280302039.0, "step": 7384 }, { "epoch": 0.9394479073909172, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5145843029022217, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8644929528236389, "num_tokens": 280338559.0, "step": 7385 }, { "epoch": 0.9395751176695077, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.16046667098999, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8740354776382446, "num_tokens": 280376731.0, "step": 7386 }, { "epoch": 0.9397023279480982, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6875874996185303, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8606840372085571, "num_tokens": 280414201.0, "step": 7387 }, { "epoch": 0.9398295382266887, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.917570114135742, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8639423251152039, "num_tokens": 280456490.0, "step": 7388 }, { "epoch": 0.9399567485052792, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.614349126815796, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8608803749084473, "num_tokens": 280499593.0, "step": 7389 }, { "epoch": 0.9400839587838697, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.744354009628296, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.856145977973938, "num_tokens": 280535722.0, "step": 7390 }, { "epoch": 0.9402111690624603, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5075738430023193, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8678252696990967, "num_tokens": 280569666.0, "step": 7391 }, { "epoch": 0.9403383793410508, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6031246185302734, "learning_rate": 1e-05, "loss": 0.3865, "mean_token_accuracy": 0.8898721933364868, "num_tokens": 280607359.0, "step": 7392 }, { "epoch": 0.9404655896196412, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3943538665771484, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8599194288253784, "num_tokens": 280641054.0, "step": 7393 }, { "epoch": 0.9405927998982317, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.537766456604004, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8526886701583862, "num_tokens": 280683798.0, "step": 7394 }, { "epoch": 0.9407200101768223, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2782111167907715, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8532744646072388, "num_tokens": 280729407.0, "step": 7395 }, { "epoch": 0.9408472204554128, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.550633192062378, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8656327724456787, "num_tokens": 280764327.0, "step": 7396 }, { "epoch": 0.9409744307340033, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1904385089874268, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8598579168319702, "num_tokens": 280802056.0, "step": 7397 }, { "epoch": 0.9411016410125939, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.6065304279327393, "learning_rate": 1e-05, "loss": 0.3903, "mean_token_accuracy": 0.8846685886383057, "num_tokens": 280837269.0, "step": 7398 }, { "epoch": 0.9412288512911843, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1261661052703857, "learning_rate": 1e-05, "loss": 0.5332, "mean_token_accuracy": 0.8419003486633301, "num_tokens": 280873914.0, "step": 7399 }, { "epoch": 0.9413560615697748, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.736654281616211, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8644819855690002, "num_tokens": 280911436.0, "step": 7400 }, { "epoch": 0.9414832718483653, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.086864709854126, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8734570741653442, "num_tokens": 280952395.0, "step": 7401 }, { "epoch": 0.9416104821269559, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.9114842414855957, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8468146324157715, "num_tokens": 280990455.0, "step": 7402 }, { "epoch": 0.9417376924055464, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2118096351623535, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.865126371383667, "num_tokens": 281020584.0, "step": 7403 }, { "epoch": 0.9418649026841369, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9735963344573975, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8745383024215698, "num_tokens": 281059907.0, "step": 7404 }, { "epoch": 0.9419921129627273, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3695247173309326, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8668687343597412, "num_tokens": 281097425.0, "step": 7405 }, { "epoch": 0.9421193232413179, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8732643127441406, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.860710620880127, "num_tokens": 281146696.0, "step": 7406 }, { "epoch": 0.9422465335199084, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.483332633972168, "learning_rate": 1e-05, "loss": 0.5129, "mean_token_accuracy": 0.8496230840682983, "num_tokens": 281186166.0, "step": 7407 }, { "epoch": 0.9423737437984989, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6660561561584473, "learning_rate": 1e-05, "loss": 0.5813, "mean_token_accuracy": 0.8321850299835205, "num_tokens": 281230372.0, "step": 7408 }, { "epoch": 0.9425009540770894, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.508026599884033, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8636965751647949, "num_tokens": 281256035.0, "step": 7409 }, { "epoch": 0.94262816435568, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4331979751586914, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8764491081237793, "num_tokens": 281290792.0, "step": 7410 }, { "epoch": 0.9427553746342705, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4828734397888184, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8680912256240845, "num_tokens": 281329949.0, "step": 7411 }, { "epoch": 0.9428825849128609, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2614641189575195, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8707002401351929, "num_tokens": 281366383.0, "step": 7412 }, { "epoch": 0.9430097951914514, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4925553798675537, "learning_rate": 1e-05, "loss": 0.4975, "mean_token_accuracy": 0.8535517454147339, "num_tokens": 281408537.0, "step": 7413 }, { "epoch": 0.943137005470042, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1427903175354004, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8568643927574158, "num_tokens": 281447745.0, "step": 7414 }, { "epoch": 0.9432642157486325, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.556215524673462, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8684956431388855, "num_tokens": 281484430.0, "step": 7415 }, { "epoch": 0.943391426027223, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.069572925567627, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8815776109695435, "num_tokens": 281519720.0, "step": 7416 }, { "epoch": 0.9435186363058136, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6889381408691406, "learning_rate": 1e-05, "loss": 0.5164, "mean_token_accuracy": 0.8477947115898132, "num_tokens": 281556039.0, "step": 7417 }, { "epoch": 0.943645846584404, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.031022787094116, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8547149300575256, "num_tokens": 281594781.0, "step": 7418 }, { "epoch": 0.9437730568629945, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8678719997406006, "learning_rate": 1e-05, "loss": 0.5179, "mean_token_accuracy": 0.8508570194244385, "num_tokens": 281632743.0, "step": 7419 }, { "epoch": 0.943900267141585, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1005518436431885, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8630925416946411, "num_tokens": 281669706.0, "step": 7420 }, { "epoch": 0.9440274774201756, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.004621982574463, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8676220178604126, "num_tokens": 281706108.0, "step": 7421 }, { "epoch": 0.9441546876987661, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2746031284332275, "learning_rate": 1e-05, "loss": 0.4985, "mean_token_accuracy": 0.8574355840682983, "num_tokens": 281743437.0, "step": 7422 }, { "epoch": 0.9442818979773566, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.009486675262451, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.876262366771698, "num_tokens": 281779652.0, "step": 7423 }, { "epoch": 0.944409108255947, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.43326473236084, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.867433488368988, "num_tokens": 281815841.0, "step": 7424 }, { "epoch": 0.9445363185345376, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8869524002075195, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8621710538864136, "num_tokens": 281853991.0, "step": 7425 }, { "epoch": 0.9446635288131281, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.51033616065979, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8573651313781738, "num_tokens": 281883264.0, "step": 7426 }, { "epoch": 0.9447907390917186, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6948142051696777, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8776129484176636, "num_tokens": 281920266.0, "step": 7427 }, { "epoch": 0.9449179493703092, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4520010948181152, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8637228012084961, "num_tokens": 281959949.0, "step": 7428 }, { "epoch": 0.9450451596488997, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.553032875061035, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8695318698883057, "num_tokens": 281992255.0, "step": 7429 }, { "epoch": 0.9451723699274901, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3987059593200684, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.863206148147583, "num_tokens": 282028000.0, "step": 7430 }, { "epoch": 0.9452995802060806, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4555656909942627, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8650562763214111, "num_tokens": 282064332.0, "step": 7431 }, { "epoch": 0.9454267904846712, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3261468410491943, "learning_rate": 1e-05, "loss": 0.4949, "mean_token_accuracy": 0.8524057865142822, "num_tokens": 282102185.0, "step": 7432 }, { "epoch": 0.9455540007632617, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.415224313735962, "learning_rate": 1e-05, "loss": 0.5758, "mean_token_accuracy": 0.8268481492996216, "num_tokens": 282138708.0, "step": 7433 }, { "epoch": 0.9456812110418522, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2580409049987793, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8743335604667664, "num_tokens": 282168223.0, "step": 7434 }, { "epoch": 0.9458084213204427, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.402088165283203, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8728146553039551, "num_tokens": 282205226.0, "step": 7435 }, { "epoch": 0.9459356315990332, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1958913803100586, "learning_rate": 1e-05, "loss": 0.5067, "mean_token_accuracy": 0.8500376343727112, "num_tokens": 282239131.0, "step": 7436 }, { "epoch": 0.9460628418776237, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.433485507965088, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8638559579849243, "num_tokens": 282276175.0, "step": 7437 }, { "epoch": 0.9461900521562142, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1409285068511963, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.855240523815155, "num_tokens": 282308598.0, "step": 7438 }, { "epoch": 0.9463172624348047, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.504952907562256, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8812232613563538, "num_tokens": 282342193.0, "step": 7439 }, { "epoch": 0.9464444727133953, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.062554121017456, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8641268014907837, "num_tokens": 282380531.0, "step": 7440 }, { "epoch": 0.9465716829919858, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6602888107299805, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8584884405136108, "num_tokens": 282410713.0, "step": 7441 }, { "epoch": 0.9466988932705762, "ewc_loss": 0.059814453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.018056631088257, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8659330606460571, "num_tokens": 282451881.0, "step": 7442 }, { "epoch": 0.9468261035491667, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8609886169433594, "learning_rate": 1e-05, "loss": 0.5207, "mean_token_accuracy": 0.8474235534667969, "num_tokens": 282486031.0, "step": 7443 }, { "epoch": 0.9469533138277573, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.079566240310669, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8748842477798462, "num_tokens": 282523753.0, "step": 7444 }, { "epoch": 0.9470805241063478, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.020345211029053, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8575435876846313, "num_tokens": 282563322.0, "step": 7445 }, { "epoch": 0.9472077343849383, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.2831337451934814, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8602012395858765, "num_tokens": 282602574.0, "step": 7446 }, { "epoch": 0.9473349446635289, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.017149925231934, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8690859079360962, "num_tokens": 282636608.0, "step": 7447 }, { "epoch": 0.9474621549421193, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4579415321350098, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8773709535598755, "num_tokens": 282675320.0, "step": 7448 }, { "epoch": 0.9475893652207098, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.849844217300415, "learning_rate": 1e-05, "loss": 0.5132, "mean_token_accuracy": 0.858615517616272, "num_tokens": 282712537.0, "step": 7449 }, { "epoch": 0.9477165754993003, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5100576877593994, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8679804801940918, "num_tokens": 282744620.0, "step": 7450 }, { "epoch": 0.9478437857778909, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6304049491882324, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8704887628555298, "num_tokens": 282786480.0, "step": 7451 }, { "epoch": 0.9479709960564814, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4752907752990723, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.871086597442627, "num_tokens": 282819143.0, "step": 7452 }, { "epoch": 0.9480982063350719, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4483346939086914, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8651990294456482, "num_tokens": 282859658.0, "step": 7453 }, { "epoch": 0.9482254166136623, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4310617446899414, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8616838455200195, "num_tokens": 282896190.0, "step": 7454 }, { "epoch": 0.9483526268922529, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3220345973968506, "learning_rate": 1e-05, "loss": 0.5006, "mean_token_accuracy": 0.853552520275116, "num_tokens": 282937458.0, "step": 7455 }, { "epoch": 0.9484798371708434, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4004037380218506, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8645656704902649, "num_tokens": 282970669.0, "step": 7456 }, { "epoch": 0.9486070474494339, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2477657794952393, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8592514991760254, "num_tokens": 283012372.0, "step": 7457 }, { "epoch": 0.9487342577280244, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3982994556427, "learning_rate": 1e-05, "loss": 0.4973, "mean_token_accuracy": 0.8504799604415894, "num_tokens": 283046165.0, "step": 7458 }, { "epoch": 0.948861468006615, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1800410747528076, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8606328964233398, "num_tokens": 283086539.0, "step": 7459 }, { "epoch": 0.9489886782852054, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.416688919067383, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8596543669700623, "num_tokens": 283123643.0, "step": 7460 }, { "epoch": 0.9491158885637959, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1180636882781982, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8807962536811829, "num_tokens": 283162354.0, "step": 7461 }, { "epoch": 0.9492430988423864, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.490560531616211, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8691234588623047, "num_tokens": 283196814.0, "step": 7462 }, { "epoch": 0.949370309120977, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.050570011138916, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.856182336807251, "num_tokens": 283238051.0, "step": 7463 }, { "epoch": 0.9494975193995675, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.622401475906372, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8531923294067383, "num_tokens": 283278025.0, "step": 7464 }, { "epoch": 0.949624729678158, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0005745887756348, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8673216104507446, "num_tokens": 283311279.0, "step": 7465 }, { "epoch": 0.9497519399567486, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.81465220451355, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8556244373321533, "num_tokens": 283351350.0, "step": 7466 }, { "epoch": 0.949879150235339, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.011945962905884, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8674372434616089, "num_tokens": 283387374.0, "step": 7467 }, { "epoch": 0.9500063605139295, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.03249979019165, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8564094305038452, "num_tokens": 283431305.0, "step": 7468 }, { "epoch": 0.95013357079252, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.199993371963501, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8629425764083862, "num_tokens": 283464941.0, "step": 7469 }, { "epoch": 0.9502607810711106, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.131577014923096, "learning_rate": 1e-05, "loss": 0.5586, "mean_token_accuracy": 0.845079779624939, "num_tokens": 283507254.0, "step": 7470 }, { "epoch": 0.9503879913497011, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.45184326171875, "learning_rate": 1e-05, "loss": 0.5151, "mean_token_accuracy": 0.8501023054122925, "num_tokens": 283545117.0, "step": 7471 }, { "epoch": 0.9505152016282916, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.025308609008789, "learning_rate": 1e-05, "loss": 0.522, "mean_token_accuracy": 0.8517712950706482, "num_tokens": 283577786.0, "step": 7472 }, { "epoch": 0.950642411906882, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5638842582702637, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8693253993988037, "num_tokens": 283613486.0, "step": 7473 }, { "epoch": 0.9507696221854726, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.807596445083618, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8628186583518982, "num_tokens": 283654480.0, "step": 7474 }, { "epoch": 0.9508968324640631, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5428106784820557, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8564850687980652, "num_tokens": 283686089.0, "step": 7475 }, { "epoch": 0.9510240427426536, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.735945463180542, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8626677393913269, "num_tokens": 283726483.0, "step": 7476 }, { "epoch": 0.9511512530212441, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.465073347091675, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.867570161819458, "num_tokens": 283756170.0, "step": 7477 }, { "epoch": 0.9512784632998347, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.513091564178467, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8645224571228027, "num_tokens": 283794950.0, "step": 7478 }, { "epoch": 0.9514056735784251, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.368042469024658, "learning_rate": 1e-05, "loss": 0.5517, "mean_token_accuracy": 0.8404452204704285, "num_tokens": 283835596.0, "step": 7479 }, { "epoch": 0.9515328838570156, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4460525512695312, "learning_rate": 1e-05, "loss": 0.5022, "mean_token_accuracy": 0.8534877300262451, "num_tokens": 283875380.0, "step": 7480 }, { "epoch": 0.9516600941356061, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3048858642578125, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8570547699928284, "num_tokens": 283910026.0, "step": 7481 }, { "epoch": 0.9517873044141967, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.408721923828125, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8666367530822754, "num_tokens": 283950533.0, "step": 7482 }, { "epoch": 0.9519145146927872, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.244419574737549, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8634476661682129, "num_tokens": 283980939.0, "step": 7483 }, { "epoch": 0.9520417249713777, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.40889835357666, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.863929033279419, "num_tokens": 284018923.0, "step": 7484 }, { "epoch": 0.9521689352499682, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1692798137664795, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8794611096382141, "num_tokens": 284059752.0, "step": 7485 }, { "epoch": 0.9522961455285587, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.446981430053711, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8702541589736938, "num_tokens": 284103783.0, "step": 7486 }, { "epoch": 0.9524233558071492, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.112672805786133, "learning_rate": 1e-05, "loss": 0.5433, "mean_token_accuracy": 0.8419967889785767, "num_tokens": 284144595.0, "step": 7487 }, { "epoch": 0.9525505660857397, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5510408878326416, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8581091165542603, "num_tokens": 284182373.0, "step": 7488 }, { "epoch": 0.9526777763643303, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0541157722473145, "learning_rate": 1e-05, "loss": 0.5348, "mean_token_accuracy": 0.8403048515319824, "num_tokens": 284222537.0, "step": 7489 }, { "epoch": 0.9528049866429208, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6960484981536865, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8562639355659485, "num_tokens": 284264780.0, "step": 7490 }, { "epoch": 0.9529321969215112, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0261902809143066, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8762699365615845, "num_tokens": 284298663.0, "step": 7491 }, { "epoch": 0.9530594072001017, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9187734127044678, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8676429390907288, "num_tokens": 284330516.0, "step": 7492 }, { "epoch": 0.9531866174786923, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.125696897506714, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8500183820724487, "num_tokens": 284370414.0, "step": 7493 }, { "epoch": 0.9533138277572828, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00017070770263671875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.026534557342529, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8839749693870544, "num_tokens": 284409079.0, "step": 7494 }, { "epoch": 0.9534410380358733, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.314777374267578, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8712804913520813, "num_tokens": 284451937.0, "step": 7495 }, { "epoch": 0.9535682483144639, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.015365123748779, "learning_rate": 1e-05, "loss": 0.5039, "mean_token_accuracy": 0.85793137550354, "num_tokens": 284490174.0, "step": 7496 }, { "epoch": 0.9536954585930543, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.484553813934326, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8706099987030029, "num_tokens": 284530578.0, "step": 7497 }, { "epoch": 0.9538226688716448, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8435652256011963, "learning_rate": 1e-05, "loss": 0.5208, "mean_token_accuracy": 0.8486182689666748, "num_tokens": 284562412.0, "step": 7498 }, { "epoch": 0.9539498791502353, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5168282985687256, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8632458448410034, "num_tokens": 284601846.0, "step": 7499 }, { "epoch": 0.9540770894288259, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.623082399368286, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8569542169570923, "num_tokens": 284640918.0, "step": 7500 }, { "epoch": 0.9542042997074164, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.47499942779541, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8713127374649048, "num_tokens": 284672268.0, "step": 7501 }, { "epoch": 0.9543315099860069, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.457184314727783, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.869780421257019, "num_tokens": 284708638.0, "step": 7502 }, { "epoch": 0.9544587202645973, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4372167587280273, "learning_rate": 1e-05, "loss": 0.5487, "mean_token_accuracy": 0.8408335447311401, "num_tokens": 284747670.0, "step": 7503 }, { "epoch": 0.9545859305431879, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.348219633102417, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8561144471168518, "num_tokens": 284786561.0, "step": 7504 }, { "epoch": 0.9547131408217784, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3816750049591064, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8662846088409424, "num_tokens": 284828631.0, "step": 7505 }, { "epoch": 0.9548403511003689, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2721104621887207, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8635854125022888, "num_tokens": 284865782.0, "step": 7506 }, { "epoch": 0.9549675613789594, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3509953022003174, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8684722781181335, "num_tokens": 284904000.0, "step": 7507 }, { "epoch": 0.95509477165755, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2200300693511963, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8666685819625854, "num_tokens": 284946432.0, "step": 7508 }, { "epoch": 0.9552219819361404, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.347468137741089, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8641903400421143, "num_tokens": 284985534.0, "step": 7509 }, { "epoch": 0.9553491922147309, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1835968494415283, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8754143118858337, "num_tokens": 285023710.0, "step": 7510 }, { "epoch": 0.9554764024933214, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.364975929260254, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8501152396202087, "num_tokens": 285064182.0, "step": 7511 }, { "epoch": 0.955603612771912, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1444575786590576, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8665292859077454, "num_tokens": 285100896.0, "step": 7512 }, { "epoch": 0.9557308230505025, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.465769052505493, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8628291487693787, "num_tokens": 285139347.0, "step": 7513 }, { "epoch": 0.955858033329093, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 3.041712999343872, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8610494136810303, "num_tokens": 285183047.0, "step": 7514 }, { "epoch": 0.9559852436076836, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.598581314086914, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.86259925365448, "num_tokens": 285225407.0, "step": 7515 }, { "epoch": 0.956112453886274, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.9694018363952637, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8681185841560364, "num_tokens": 285258465.0, "step": 7516 }, { "epoch": 0.9562396641648645, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8269267082214355, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8637344837188721, "num_tokens": 285297221.0, "step": 7517 }, { "epoch": 0.956366874443455, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.9593448638916016, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8672782778739929, "num_tokens": 285338145.0, "step": 7518 }, { "epoch": 0.9564940847220456, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1250996589660645, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8657569885253906, "num_tokens": 285381647.0, "step": 7519 }, { "epoch": 0.9566212950006361, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2011382579803467, "learning_rate": 1e-05, "loss": 0.5297, "mean_token_accuracy": 0.8432145714759827, "num_tokens": 285424468.0, "step": 7520 }, { "epoch": 0.9567485052792266, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.260195255279541, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8665881156921387, "num_tokens": 285470415.0, "step": 7521 }, { "epoch": 0.956875715557817, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.5566463470458984, "learning_rate": 1e-05, "loss": 0.5122, "mean_token_accuracy": 0.8502748608589172, "num_tokens": 285513299.0, "step": 7522 }, { "epoch": 0.9570029258364076, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.098498344421387, "learning_rate": 1e-05, "loss": 0.5363, "mean_token_accuracy": 0.8468589782714844, "num_tokens": 285555855.0, "step": 7523 }, { "epoch": 0.9571301361149981, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.686720132827759, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.866258442401886, "num_tokens": 285589670.0, "step": 7524 }, { "epoch": 0.9572573463935886, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8566722869873047, "learning_rate": 1e-05, "loss": 0.5244, "mean_token_accuracy": 0.8510980606079102, "num_tokens": 285627257.0, "step": 7525 }, { "epoch": 0.9573845566721791, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5864360332489014, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8671095967292786, "num_tokens": 285670100.0, "step": 7526 }, { "epoch": 0.9575117669507697, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6834657192230225, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8670278787612915, "num_tokens": 285706902.0, "step": 7527 }, { "epoch": 0.9576389772293601, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.48189640045166, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.857812225818634, "num_tokens": 285746952.0, "step": 7528 }, { "epoch": 0.9577661875079506, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5696287155151367, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8760759234428406, "num_tokens": 285781532.0, "step": 7529 }, { "epoch": 0.9578933977865411, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.380108594894409, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8574968576431274, "num_tokens": 285818725.0, "step": 7530 }, { "epoch": 0.9580206080651317, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5056464672088623, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8541112542152405, "num_tokens": 285855502.0, "step": 7531 }, { "epoch": 0.9581478183437222, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.282444715499878, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8597341179847717, "num_tokens": 285893815.0, "step": 7532 }, { "epoch": 0.9582750286223127, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5093963146209717, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8589868545532227, "num_tokens": 285928709.0, "step": 7533 }, { "epoch": 0.9584022389009031, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.181391477584839, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8785050511360168, "num_tokens": 285959358.0, "step": 7534 }, { "epoch": 0.9585294491794937, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5753183364868164, "learning_rate": 1e-05, "loss": 0.5294, "mean_token_accuracy": 0.845016360282898, "num_tokens": 285997643.0, "step": 7535 }, { "epoch": 0.9586566594580842, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1070516109466553, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8666121363639832, "num_tokens": 286032866.0, "step": 7536 }, { "epoch": 0.9587838697366747, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.7061679363250732, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8626297116279602, "num_tokens": 286066585.0, "step": 7537 }, { "epoch": 0.9589110800152653, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.069965124130249, "learning_rate": 1e-05, "loss": 0.5141, "mean_token_accuracy": 0.8473043441772461, "num_tokens": 286106151.0, "step": 7538 }, { "epoch": 0.9590382902938558, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.873802661895752, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8676572442054749, "num_tokens": 286142409.0, "step": 7539 }, { "epoch": 0.9591655005724462, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1526052951812744, "learning_rate": 1e-05, "loss": 0.5189, "mean_token_accuracy": 0.8511936068534851, "num_tokens": 286179408.0, "step": 7540 }, { "epoch": 0.9592927108510367, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.998905658721924, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8617473840713501, "num_tokens": 286212767.0, "step": 7541 }, { "epoch": 0.9594199211296273, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.337026596069336, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8609321117401123, "num_tokens": 286250176.0, "step": 7542 }, { "epoch": 0.9595471314082178, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9371793270111084, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8596912622451782, "num_tokens": 286292060.0, "step": 7543 }, { "epoch": 0.9596743416868083, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000171661376953125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4758412837982178, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.868683397769928, "num_tokens": 286333047.0, "step": 7544 }, { "epoch": 0.9598015519653988, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.732081890106201, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8653745651245117, "num_tokens": 286373465.0, "step": 7545 }, { "epoch": 0.9599287622439893, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5093512535095215, "learning_rate": 1e-05, "loss": 0.5258, "mean_token_accuracy": 0.8511184453964233, "num_tokens": 286414431.0, "step": 7546 }, { "epoch": 0.9600559725225798, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5217039585113525, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8597820997238159, "num_tokens": 286449431.0, "step": 7547 }, { "epoch": 0.9601831828011703, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4832570552825928, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8644859790802002, "num_tokens": 286486563.0, "step": 7548 }, { "epoch": 0.9603103930797608, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.330575704574585, "learning_rate": 1e-05, "loss": 0.4981, "mean_token_accuracy": 0.853975772857666, "num_tokens": 286528691.0, "step": 7549 }, { "epoch": 0.9604376033583514, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.46504282951355, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8590549230575562, "num_tokens": 286568726.0, "step": 7550 }, { "epoch": 0.9605648136369419, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.2037150859832764, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8709962964057922, "num_tokens": 286608159.0, "step": 7551 }, { "epoch": 0.9606920239155323, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5012688636779785, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.8592901229858398, "num_tokens": 286645494.0, "step": 7552 }, { "epoch": 0.9608192341941229, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.121589422225952, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8712712526321411, "num_tokens": 286679819.0, "step": 7553 }, { "epoch": 0.9609464444727134, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5782814025878906, "learning_rate": 1e-05, "loss": 0.5164, "mean_token_accuracy": 0.8473391532897949, "num_tokens": 286714116.0, "step": 7554 }, { "epoch": 0.9610736547513039, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.056117296218872, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.874305248260498, "num_tokens": 286746905.0, "step": 7555 }, { "epoch": 0.9612008650298944, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7042438983917236, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8833293914794922, "num_tokens": 286784724.0, "step": 7556 }, { "epoch": 0.961328075308485, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0348153114318848, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8671667575836182, "num_tokens": 286825672.0, "step": 7557 }, { "epoch": 0.9614552855870754, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8908910751342773, "learning_rate": 1e-05, "loss": 0.533, "mean_token_accuracy": 0.8497388362884521, "num_tokens": 286862072.0, "step": 7558 }, { "epoch": 0.9615824958656659, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1109507083892822, "learning_rate": 1e-05, "loss": 0.4726, "mean_token_accuracy": 0.8641053438186646, "num_tokens": 286903111.0, "step": 7559 }, { "epoch": 0.9617097061442564, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.023820400238037, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8672263622283936, "num_tokens": 286947957.0, "step": 7560 }, { "epoch": 0.961836916422847, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3129425048828125, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8623297214508057, "num_tokens": 286983254.0, "step": 7561 }, { "epoch": 0.9619641267014375, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.001227378845215, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8826526999473572, "num_tokens": 287015294.0, "step": 7562 }, { "epoch": 0.962091336980028, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.456973075866699, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8670936822891235, "num_tokens": 287053439.0, "step": 7563 }, { "epoch": 0.9622185472586186, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.860344886779785, "learning_rate": 1e-05, "loss": 0.5548, "mean_token_accuracy": 0.8397179245948792, "num_tokens": 287096970.0, "step": 7564 }, { "epoch": 0.962345757537209, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.491980791091919, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8555074334144592, "num_tokens": 287139179.0, "step": 7565 }, { "epoch": 0.9624729678157995, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.670454263687134, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8556508421897888, "num_tokens": 287177375.0, "step": 7566 }, { "epoch": 0.96260017809439, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.458214282989502, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8658369779586792, "num_tokens": 287211296.0, "step": 7567 }, { "epoch": 0.9627273883729806, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.519559860229492, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8687053918838501, "num_tokens": 287246913.0, "step": 7568 }, { "epoch": 0.9628545986515711, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3942089080810547, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.8588224053382874, "num_tokens": 287283639.0, "step": 7569 }, { "epoch": 0.9629818089301616, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.455714464187622, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8696658611297607, "num_tokens": 287322694.0, "step": 7570 }, { "epoch": 0.963109019208752, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3156261444091797, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8530488014221191, "num_tokens": 287361103.0, "step": 7571 }, { "epoch": 0.9632362294873426, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.4005918502807617, "learning_rate": 1e-05, "loss": 0.5182, "mean_token_accuracy": 0.8479154109954834, "num_tokens": 287400666.0, "step": 7572 }, { "epoch": 0.9633634397659331, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.248650312423706, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8783073425292969, "num_tokens": 287434729.0, "step": 7573 }, { "epoch": 0.9634906500445236, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3922367095947266, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8744628429412842, "num_tokens": 287472194.0, "step": 7574 }, { "epoch": 0.9636178603231141, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1830737590789795, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8567469120025635, "num_tokens": 287510139.0, "step": 7575 }, { "epoch": 0.9637450706017047, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.438415288925171, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8519954681396484, "num_tokens": 287549529.0, "step": 7576 }, { "epoch": 0.9638722808802951, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1066348552703857, "learning_rate": 1e-05, "loss": 0.5064, "mean_token_accuracy": 0.8456945419311523, "num_tokens": 287596525.0, "step": 7577 }, { "epoch": 0.9639994911588856, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.537722110748291, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8790168166160583, "num_tokens": 287632284.0, "step": 7578 }, { "epoch": 0.9641267014374761, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0460245609283447, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8675509095191956, "num_tokens": 287671867.0, "step": 7579 }, { "epoch": 0.9642539117160667, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001735687255859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6921074390411377, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8530453443527222, "num_tokens": 287706715.0, "step": 7580 }, { "epoch": 0.9643811219946572, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001735687255859375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0224099159240723, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8651019334793091, "num_tokens": 287746269.0, "step": 7581 }, { "epoch": 0.9645083322732477, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.895050287246704, "learning_rate": 1e-05, "loss": 0.5131, "mean_token_accuracy": 0.8538437485694885, "num_tokens": 287786194.0, "step": 7582 }, { "epoch": 0.9646355425518381, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00017261505126953125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1358861923217773, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8546388149261475, "num_tokens": 287820873.0, "step": 7583 }, { "epoch": 0.9647627528304287, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.025067329406738, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8628378510475159, "num_tokens": 287847356.0, "step": 7584 }, { "epoch": 0.9648899631090192, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001735687255859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3288772106170654, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8573046326637268, "num_tokens": 287879611.0, "step": 7585 }, { "epoch": 0.9650171733876097, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001735687255859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.95539927482605, "learning_rate": 1e-05, "loss": 0.5368, "mean_token_accuracy": 0.8476582765579224, "num_tokens": 287920865.0, "step": 7586 }, { "epoch": 0.9651443836662003, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001735687255859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.443647861480713, "learning_rate": 1e-05, "loss": 0.5017, "mean_token_accuracy": 0.8565516471862793, "num_tokens": 287963997.0, "step": 7587 }, { "epoch": 0.9652715939447908, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.796157121658325, "learning_rate": 1e-05, "loss": 0.5036, "mean_token_accuracy": 0.8531105518341064, "num_tokens": 287997283.0, "step": 7588 }, { "epoch": 0.9653988042233812, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4761006832122803, "learning_rate": 1e-05, "loss": 0.5553, "mean_token_accuracy": 0.8364279866218567, "num_tokens": 288037876.0, "step": 7589 }, { "epoch": 0.9655260145019717, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6028225421905518, "learning_rate": 1e-05, "loss": 0.536, "mean_token_accuracy": 0.847241222858429, "num_tokens": 288078835.0, "step": 7590 }, { "epoch": 0.9656532247805623, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4411051273345947, "learning_rate": 1e-05, "loss": 0.4949, "mean_token_accuracy": 0.8555195331573486, "num_tokens": 288118342.0, "step": 7591 }, { "epoch": 0.9657804350591528, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4351868629455566, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8655312061309814, "num_tokens": 288159629.0, "step": 7592 }, { "epoch": 0.9659076453377433, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.410332202911377, "learning_rate": 1e-05, "loss": 0.4951, "mean_token_accuracy": 0.8563873767852783, "num_tokens": 288199887.0, "step": 7593 }, { "epoch": 0.9660348556163338, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3318586349487305, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.8528861403465271, "num_tokens": 288233958.0, "step": 7594 }, { "epoch": 0.9661620658949243, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.381718635559082, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8576621413230896, "num_tokens": 288274735.0, "step": 7595 }, { "epoch": 0.9662892761735148, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2522082328796387, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8654147386550903, "num_tokens": 288309699.0, "step": 7596 }, { "epoch": 0.9664164864521053, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3745791912078857, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8613731861114502, "num_tokens": 288344397.0, "step": 7597 }, { "epoch": 0.9665436967306958, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1894619464874268, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.867314875125885, "num_tokens": 288385612.0, "step": 7598 }, { "epoch": 0.9666709070092864, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4090936183929443, "learning_rate": 1e-05, "loss": 0.5225, "mean_token_accuracy": 0.8474146723747253, "num_tokens": 288420570.0, "step": 7599 }, { "epoch": 0.9667981172878769, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.13986873626709, "learning_rate": 1e-05, "loss": 0.5244, "mean_token_accuracy": 0.8490263223648071, "num_tokens": 288456917.0, "step": 7600 }, { "epoch": 0.9669253275664673, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4567391872406006, "learning_rate": 1e-05, "loss": 0.5095, "mean_token_accuracy": 0.8474463820457458, "num_tokens": 288498228.0, "step": 7601 }, { "epoch": 0.9670525378450578, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0715367794036865, "learning_rate": 1e-05, "loss": 0.5335, "mean_token_accuracy": 0.8414413928985596, "num_tokens": 288534082.0, "step": 7602 }, { "epoch": 0.9671797481236484, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5589776039123535, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8569350838661194, "num_tokens": 288576142.0, "step": 7603 }, { "epoch": 0.9673069584022389, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00017452239990234375, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0127274990081787, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8548977375030518, "num_tokens": 288609707.0, "step": 7604 }, { "epoch": 0.9674341686808294, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017547607421875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.731029510498047, "learning_rate": 1e-05, "loss": 0.5431, "mean_token_accuracy": 0.845535159111023, "num_tokens": 288648228.0, "step": 7605 }, { "epoch": 0.96756137895942, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00017547607421875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.9752771854400635, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8663530349731445, "num_tokens": 288689845.0, "step": 7606 }, { "epoch": 0.9676885892380104, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017547607421875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9691524505615234, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8738498091697693, "num_tokens": 288729571.0, "step": 7607 }, { "epoch": 0.9678157995166009, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.096083164215088, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8754415512084961, "num_tokens": 288764070.0, "step": 7608 }, { "epoch": 0.9679430097951914, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.158096790313721, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8629651069641113, "num_tokens": 288805557.0, "step": 7609 }, { "epoch": 0.968070220073782, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.3775901794433594, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8695220947265625, "num_tokens": 288840567.0, "step": 7610 }, { "epoch": 0.9681974303523725, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.109958648681641, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8603000640869141, "num_tokens": 288881952.0, "step": 7611 }, { "epoch": 0.968324640630963, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.583674430847168, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8595651388168335, "num_tokens": 288919208.0, "step": 7612 }, { "epoch": 0.9684518509095535, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8997950553894043, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8604332208633423, "num_tokens": 288961622.0, "step": 7613 }, { "epoch": 0.968579061188144, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5746636390686035, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8700072169303894, "num_tokens": 288999799.0, "step": 7614 }, { "epoch": 0.9687062714667345, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.693657636642456, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8713662624359131, "num_tokens": 289040434.0, "step": 7615 }, { "epoch": 0.968833481745325, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.505465030670166, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8600205183029175, "num_tokens": 289072847.0, "step": 7616 }, { "epoch": 0.9689606920239155, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.545952558517456, "learning_rate": 1e-05, "loss": 0.5351, "mean_token_accuracy": 0.847299337387085, "num_tokens": 289114838.0, "step": 7617 }, { "epoch": 0.9690879023025061, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.415102005004883, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8558520674705505, "num_tokens": 289151937.0, "step": 7618 }, { "epoch": 0.9692151125810966, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4511141777038574, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8617499470710754, "num_tokens": 289188630.0, "step": 7619 }, { "epoch": 0.969342322859687, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.322575092315674, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8645921945571899, "num_tokens": 289229412.0, "step": 7620 }, { "epoch": 0.9694695331382776, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4084603786468506, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.871483325958252, "num_tokens": 289265195.0, "step": 7621 }, { "epoch": 0.9695967434168681, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2537145614624023, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8522574305534363, "num_tokens": 289306203.0, "step": 7622 }, { "epoch": 0.9697239536954586, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4189178943634033, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8570566177368164, "num_tokens": 289345196.0, "step": 7623 }, { "epoch": 0.9698511639740491, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.179795503616333, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.871487557888031, "num_tokens": 289377221.0, "step": 7624 }, { "epoch": 0.9699783742526397, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4729578495025635, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8560656309127808, "num_tokens": 289413196.0, "step": 7625 }, { "epoch": 0.9701055845312301, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0945541858673096, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8551850914955139, "num_tokens": 289452774.0, "step": 7626 }, { "epoch": 0.9702327948098206, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5891737937927246, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8624557852745056, "num_tokens": 289489806.0, "step": 7627 }, { "epoch": 0.9703600050884111, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.02629017829895, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8605940937995911, "num_tokens": 289528362.0, "step": 7628 }, { "epoch": 0.9704872153670017, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.785830497741699, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8618050813674927, "num_tokens": 289562725.0, "step": 7629 }, { "epoch": 0.9706144256455922, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.021620035171509, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8558626174926758, "num_tokens": 289605968.0, "step": 7630 }, { "epoch": 0.9707416359241827, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.016798496246338, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8537371158599854, "num_tokens": 289639065.0, "step": 7631 }, { "epoch": 0.9708688462027731, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.207219362258911, "learning_rate": 1e-05, "loss": 0.5486, "mean_token_accuracy": 0.8394927978515625, "num_tokens": 289679616.0, "step": 7632 }, { "epoch": 0.9709960564813637, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.106349945068359, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8618991374969482, "num_tokens": 289713189.0, "step": 7633 }, { "epoch": 0.9711232667599542, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.449638843536377, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8658249378204346, "num_tokens": 289751406.0, "step": 7634 }, { "epoch": 0.9712504770385447, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.963716983795166, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8620410561561584, "num_tokens": 289795735.0, "step": 7635 }, { "epoch": 0.9713776873171353, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5702340602874756, "learning_rate": 1e-05, "loss": 0.5257, "mean_token_accuracy": 0.8483391404151917, "num_tokens": 289833574.0, "step": 7636 }, { "epoch": 0.9715048975957258, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.737459421157837, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8723282217979431, "num_tokens": 289871922.0, "step": 7637 }, { "epoch": 0.9716321078743162, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.524101972579956, "learning_rate": 1e-05, "loss": 0.5127, "mean_token_accuracy": 0.8546890020370483, "num_tokens": 289913517.0, "step": 7638 }, { "epoch": 0.9717593181529067, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5380337238311768, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8599692583084106, "num_tokens": 289957046.0, "step": 7639 }, { "epoch": 0.9718865284314973, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.457448720932007, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8613415956497192, "num_tokens": 289997983.0, "step": 7640 }, { "epoch": 0.9720137387100878, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4057509899139404, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8633850812911987, "num_tokens": 290035923.0, "step": 7641 }, { "epoch": 0.9721409489886783, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3963963985443115, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8700709342956543, "num_tokens": 290075152.0, "step": 7642 }, { "epoch": 0.9722681592672688, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.332895040512085, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8751641511917114, "num_tokens": 290108360.0, "step": 7643 }, { "epoch": 0.9723953695458593, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3455309867858887, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.870072603225708, "num_tokens": 290148698.0, "step": 7644 }, { "epoch": 0.9725225798244498, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.295099973678589, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8696132302284241, "num_tokens": 290185847.0, "step": 7645 }, { "epoch": 0.9726497901030403, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.299596071243286, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8625695705413818, "num_tokens": 290228763.0, "step": 7646 }, { "epoch": 0.9727770003816308, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.286285400390625, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8574562668800354, "num_tokens": 290262040.0, "step": 7647 }, { "epoch": 0.9729042106602214, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.269620895385742, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8673057556152344, "num_tokens": 290302205.0, "step": 7648 }, { "epoch": 0.9730314209388119, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.273074150085449, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8671764731407166, "num_tokens": 290340132.0, "step": 7649 }, { "epoch": 0.9731586312174023, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2433178424835205, "learning_rate": 1e-05, "loss": 0.3859, "mean_token_accuracy": 0.8880760073661804, "num_tokens": 290375808.0, "step": 7650 }, { "epoch": 0.9732858414959928, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.27592396736145, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8578079342842102, "num_tokens": 290415538.0, "step": 7651 }, { "epoch": 0.9734130517745834, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2181596755981445, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8690983653068542, "num_tokens": 290453363.0, "step": 7652 }, { "epoch": 0.9735402620531739, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.3122854232788086, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8546292781829834, "num_tokens": 290483869.0, "step": 7653 }, { "epoch": 0.9736674723317644, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.178044319152832, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8700568675994873, "num_tokens": 290521772.0, "step": 7654 }, { "epoch": 0.973794682610355, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3453335762023926, "learning_rate": 1e-05, "loss": 0.5168, "mean_token_accuracy": 0.8465372920036316, "num_tokens": 290563226.0, "step": 7655 }, { "epoch": 0.9739218928889454, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1114890575408936, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8611451983451843, "num_tokens": 290602209.0, "step": 7656 }, { "epoch": 0.9740491031675359, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4200475215911865, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8520187735557556, "num_tokens": 290645554.0, "step": 7657 }, { "epoch": 0.9741763134461264, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 3.035968065261841, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8698021173477173, "num_tokens": 290683982.0, "step": 7658 }, { "epoch": 0.974303523724717, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.577622890472412, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8525853753089905, "num_tokens": 290717720.0, "step": 7659 }, { "epoch": 0.9744307340033075, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 2.957251787185669, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8593262434005737, "num_tokens": 290759711.0, "step": 7660 }, { "epoch": 0.974557944281898, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8202035427093506, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8647197484970093, "num_tokens": 290795767.0, "step": 7661 }, { "epoch": 0.9746851545604885, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.9225008487701416, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8689398169517517, "num_tokens": 290839148.0, "step": 7662 }, { "epoch": 0.974812364839079, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.182767868041992, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8650623559951782, "num_tokens": 290878325.0, "step": 7663 }, { "epoch": 0.9749395751176695, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.1853885650634766, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8589917421340942, "num_tokens": 290913053.0, "step": 7664 }, { "epoch": 0.97506678539626, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.389589309692383, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8685307502746582, "num_tokens": 290954984.0, "step": 7665 }, { "epoch": 0.9751939956748505, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.630622148513794, "learning_rate": 1e-05, "loss": 0.5148, "mean_token_accuracy": 0.8528823852539062, "num_tokens": 290996052.0, "step": 7666 }, { "epoch": 0.9753212059534411, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.216836452484131, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.8497490882873535, "num_tokens": 291034021.0, "step": 7667 }, { "epoch": 0.9754484162320316, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7842190265655518, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8727003335952759, "num_tokens": 291064216.0, "step": 7668 }, { "epoch": 0.975575626510622, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9581284523010254, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8697658181190491, "num_tokens": 291100119.0, "step": 7669 }, { "epoch": 0.9757028367892125, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6669270992279053, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8601747155189514, "num_tokens": 291136709.0, "step": 7670 }, { "epoch": 0.9758300470678031, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.794358253479004, "learning_rate": 1e-05, "loss": 0.5376, "mean_token_accuracy": 0.8449028134346008, "num_tokens": 291175038.0, "step": 7671 }, { "epoch": 0.9759572573463936, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5334908962249756, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8635971546173096, "num_tokens": 291206950.0, "step": 7672 }, { "epoch": 0.9760844676249841, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.733458995819092, "learning_rate": 1e-05, "loss": 0.5058, "mean_token_accuracy": 0.8544070720672607, "num_tokens": 291247847.0, "step": 7673 }, { "epoch": 0.9762116779035747, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3782131671905518, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8751440644264221, "num_tokens": 291282262.0, "step": 7674 }, { "epoch": 0.9763388881821651, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.684312582015991, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8624361157417297, "num_tokens": 291315090.0, "step": 7675 }, { "epoch": 0.9764660984607556, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2603495121002197, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8685781955718994, "num_tokens": 291356380.0, "step": 7676 }, { "epoch": 0.9765933087393461, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7311246395111084, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8615556955337524, "num_tokens": 291401133.0, "step": 7677 }, { "epoch": 0.9767205190179367, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1960484981536865, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8638100028038025, "num_tokens": 291439999.0, "step": 7678 }, { "epoch": 0.9768477292965272, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8560690879821777, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8666468858718872, "num_tokens": 291476784.0, "step": 7679 }, { "epoch": 0.9769749395751177, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.201986074447632, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8728659749031067, "num_tokens": 291517677.0, "step": 7680 }, { "epoch": 0.9771021498537081, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9814157485961914, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8662670850753784, "num_tokens": 291551838.0, "step": 7681 }, { "epoch": 0.9772293601322987, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3445918560028076, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8714936971664429, "num_tokens": 291591192.0, "step": 7682 }, { "epoch": 0.9773565704108892, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.980449676513672, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.8620363473892212, "num_tokens": 291626885.0, "step": 7683 }, { "epoch": 0.9774837806894797, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5000596046447754, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8645930290222168, "num_tokens": 291666421.0, "step": 7684 }, { "epoch": 0.9776109909680702, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7774012088775635, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.8599477410316467, "num_tokens": 291704973.0, "step": 7685 }, { "epoch": 0.9777382012466608, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.533547878265381, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8573422431945801, "num_tokens": 291750279.0, "step": 7686 }, { "epoch": 0.9778654115252512, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.558821439743042, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8591581583023071, "num_tokens": 291784390.0, "step": 7687 }, { "epoch": 0.9779926218038417, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.509705066680908, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8707160949707031, "num_tokens": 291826385.0, "step": 7688 }, { "epoch": 0.9781198320824323, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.372096538543701, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8536549210548401, "num_tokens": 291858129.0, "step": 7689 }, { "epoch": 0.9782470423610228, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00017642974853515625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.483219623565674, "learning_rate": 1e-05, "loss": 0.5459, "mean_token_accuracy": 0.8482346534729004, "num_tokens": 291902763.0, "step": 7690 }, { "epoch": 0.9783742526396133, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.221921920776367, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8750808238983154, "num_tokens": 291940686.0, "step": 7691 }, { "epoch": 0.9785014629182038, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4958317279815674, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8631682991981506, "num_tokens": 291976849.0, "step": 7692 }, { "epoch": 0.9786286731967943, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1249966621398926, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8773356676101685, "num_tokens": 292014105.0, "step": 7693 }, { "epoch": 0.9787558834753848, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.572249174118042, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8651645183563232, "num_tokens": 292055083.0, "step": 7694 }, { "epoch": 0.9788830937539753, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0573978424072266, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8654758930206299, "num_tokens": 292089713.0, "step": 7695 }, { "epoch": 0.9790103040325658, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.7005867958068848, "learning_rate": 1e-05, "loss": 0.505, "mean_token_accuracy": 0.851962685585022, "num_tokens": 292128139.0, "step": 7696 }, { "epoch": 0.9791375143111564, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.035318374633789, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.860011875629425, "num_tokens": 292164674.0, "step": 7697 }, { "epoch": 0.9792647245897469, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8810086250305176, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8649103045463562, "num_tokens": 292197175.0, "step": 7698 }, { "epoch": 0.9793919348683373, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1087164878845215, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8488502502441406, "num_tokens": 292237051.0, "step": 7699 }, { "epoch": 0.9795191451469278, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.029281139373779, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8718502521514893, "num_tokens": 292272817.0, "step": 7700 }, { "epoch": 0.9796463554255184, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.2931301593780518, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8614452481269836, "num_tokens": 292310239.0, "step": 7701 }, { "epoch": 0.9797735657041089, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.02078104019165, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8610131740570068, "num_tokens": 292351090.0, "step": 7702 }, { "epoch": 0.9799007759826994, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4747042655944824, "learning_rate": 1e-05, "loss": 0.543, "mean_token_accuracy": 0.8439946174621582, "num_tokens": 292390500.0, "step": 7703 }, { "epoch": 0.98002798626129, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.862379789352417, "learning_rate": 1e-05, "loss": 0.5109, "mean_token_accuracy": 0.8547186851501465, "num_tokens": 292423483.0, "step": 7704 }, { "epoch": 0.9801551965398804, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.509894609451294, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8767156600952148, "num_tokens": 292460791.0, "step": 7705 }, { "epoch": 0.9802824068184709, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6543490886688232, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8777982592582703, "num_tokens": 292502214.0, "step": 7706 }, { "epoch": 0.9804096170970614, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.477627754211426, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8489726781845093, "num_tokens": 292541845.0, "step": 7707 }, { "epoch": 0.980536827375652, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.500281572341919, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8741132616996765, "num_tokens": 292573932.0, "step": 7708 }, { "epoch": 0.9806640376542425, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4081828594207764, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8687015771865845, "num_tokens": 292615932.0, "step": 7709 }, { "epoch": 0.980791247932833, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.421039581298828, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.850292444229126, "num_tokens": 292648680.0, "step": 7710 }, { "epoch": 0.9809184582114235, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3364484310150146, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8752568364143372, "num_tokens": 292687987.0, "step": 7711 }, { "epoch": 0.981045668490014, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3713347911834717, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8596569299697876, "num_tokens": 292732243.0, "step": 7712 }, { "epoch": 0.9811728787686045, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2767374515533447, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8684374094009399, "num_tokens": 292764869.0, "step": 7713 }, { "epoch": 0.981300089047195, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.338444471359253, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8720791339874268, "num_tokens": 292806401.0, "step": 7714 }, { "epoch": 0.9814272993257855, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2462446689605713, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8696484565734863, "num_tokens": 292843165.0, "step": 7715 }, { "epoch": 0.9815545096043761, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001773834228515625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3395724296569824, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.861525297164917, "num_tokens": 292885092.0, "step": 7716 }, { "epoch": 0.9816817198829666, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1969666481018066, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.854183554649353, "num_tokens": 292920595.0, "step": 7717 }, { "epoch": 0.981808930161557, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.388296127319336, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8541319370269775, "num_tokens": 292954079.0, "step": 7718 }, { "epoch": 0.9819361404401475, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1413230895996094, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8681162595748901, "num_tokens": 292985554.0, "step": 7719 }, { "epoch": 0.9820633507187381, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4535269737243652, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8637937307357788, "num_tokens": 293021832.0, "step": 7720 }, { "epoch": 0.9821905609973286, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 3.0780580043792725, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8613480925559998, "num_tokens": 293054220.0, "step": 7721 }, { "epoch": 0.9823177712759191, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5586535930633545, "learning_rate": 1e-05, "loss": 0.5092, "mean_token_accuracy": 0.8528757095336914, "num_tokens": 293099348.0, "step": 7722 }, { "epoch": 0.9824449815545097, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.007681131362915, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8591796159744263, "num_tokens": 293134801.0, "step": 7723 }, { "epoch": 0.9825721918331001, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7625207901000977, "learning_rate": 1e-05, "loss": 0.5435, "mean_token_accuracy": 0.8398202657699585, "num_tokens": 293172023.0, "step": 7724 }, { "epoch": 0.9826994021116906, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 2.9820165634155273, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8622257113456726, "num_tokens": 293210019.0, "step": 7725 }, { "epoch": 0.9828266123902811, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.017166614532471, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8636160492897034, "num_tokens": 293249614.0, "step": 7726 }, { "epoch": 0.9829538226688717, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1303913593292236, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8717832565307617, "num_tokens": 293289860.0, "step": 7727 }, { "epoch": 0.9830810329474622, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.184341907501221, "learning_rate": 1e-05, "loss": 0.5389, "mean_token_accuracy": 0.8502436876296997, "num_tokens": 293330337.0, "step": 7728 }, { "epoch": 0.9832082432260527, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.440699815750122, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.864173173904419, "num_tokens": 293370062.0, "step": 7729 }, { "epoch": 0.9833354535046431, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.089267253875732, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8614041805267334, "num_tokens": 293409544.0, "step": 7730 }, { "epoch": 0.9834626637832337, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6116316318511963, "learning_rate": 1e-05, "loss": 0.5391, "mean_token_accuracy": 0.8496674299240112, "num_tokens": 293453924.0, "step": 7731 }, { "epoch": 0.9835898740618242, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8595805168151855, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8670297861099243, "num_tokens": 293493759.0, "step": 7732 }, { "epoch": 0.9837170843404147, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.575439691543579, "learning_rate": 1e-05, "loss": 0.5482, "mean_token_accuracy": 0.8410419225692749, "num_tokens": 293530155.0, "step": 7733 }, { "epoch": 0.9838442946190052, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.660130262374878, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8597264289855957, "num_tokens": 293573083.0, "step": 7734 }, { "epoch": 0.9839715048975958, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4872474670410156, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8715567588806152, "num_tokens": 293608312.0, "step": 7735 }, { "epoch": 0.9840987151761862, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5173020362854004, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8606354594230652, "num_tokens": 293645036.0, "step": 7736 }, { "epoch": 0.9842259254547767, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4109268188476562, "learning_rate": 1e-05, "loss": 0.563, "mean_token_accuracy": 0.8384281992912292, "num_tokens": 293679063.0, "step": 7737 }, { "epoch": 0.9843531357333672, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.416564464569092, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8648974895477295, "num_tokens": 293720239.0, "step": 7738 }, { "epoch": 0.9844803460119578, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3204269409179688, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8783469200134277, "num_tokens": 293757154.0, "step": 7739 }, { "epoch": 0.9846075562905483, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.388901472091675, "learning_rate": 1e-05, "loss": 0.5088, "mean_token_accuracy": 0.850744366645813, "num_tokens": 293797181.0, "step": 7740 }, { "epoch": 0.9847347665691388, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2456812858581543, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8601212501525879, "num_tokens": 293838130.0, "step": 7741 }, { "epoch": 0.9848619768477292, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4056143760681152, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8572707772254944, "num_tokens": 293877341.0, "step": 7742 }, { "epoch": 0.9849891871263198, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1670846939086914, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8666422367095947, "num_tokens": 293913764.0, "step": 7743 }, { "epoch": 0.9851163974049103, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4657039642333984, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8566825985908508, "num_tokens": 293954142.0, "step": 7744 }, { "epoch": 0.9852436076835008, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.0900766849517822, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8553091883659363, "num_tokens": 293999016.0, "step": 7745 }, { "epoch": 0.9853708179620914, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.596181631088257, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8638617396354675, "num_tokens": 294033545.0, "step": 7746 }, { "epoch": 0.9854980282406819, "ewc_loss": 0.06005859375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.010777235031128, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.876612663269043, "num_tokens": 294076763.0, "step": 7747 }, { "epoch": 0.9856252385192723, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.799557685852051, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8837990760803223, "num_tokens": 294114661.0, "step": 7748 }, { "epoch": 0.9857524487978628, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0312461853027344, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8578769564628601, "num_tokens": 294148819.0, "step": 7749 }, { "epoch": 0.9858796590764534, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.02249813079834, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8663860559463501, "num_tokens": 294190820.0, "step": 7750 }, { "epoch": 0.9860068693550439, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.188499689102173, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8617786169052124, "num_tokens": 294230814.0, "step": 7751 }, { "epoch": 0.9861340796336344, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.142125129699707, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.860349178314209, "num_tokens": 294271811.0, "step": 7752 }, { "epoch": 0.986261289912225, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4602837562561035, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.8594892024993896, "num_tokens": 294313901.0, "step": 7753 }, { "epoch": 0.9863885001908154, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.010547637939453, "learning_rate": 1e-05, "loss": 0.5121, "mean_token_accuracy": 0.8514552712440491, "num_tokens": 294353275.0, "step": 7754 }, { "epoch": 0.9865157104694059, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5761024951934814, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8804632425308228, "num_tokens": 294390569.0, "step": 7755 }, { "epoch": 0.9866429207479964, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7937498092651367, "learning_rate": 1e-05, "loss": 0.5122, "mean_token_accuracy": 0.8504459857940674, "num_tokens": 294433050.0, "step": 7756 }, { "epoch": 0.986770131026587, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.547027111053467, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8712481260299683, "num_tokens": 294471198.0, "step": 7757 }, { "epoch": 0.9868973413051775, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5979909896850586, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8661450147628784, "num_tokens": 294511855.0, "step": 7758 }, { "epoch": 0.987024551583768, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.477736473083496, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8675289154052734, "num_tokens": 294547056.0, "step": 7759 }, { "epoch": 0.9871517618623584, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4617435932159424, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8589392900466919, "num_tokens": 294584516.0, "step": 7760 }, { "epoch": 0.987278972140949, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.400418996810913, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8605039119720459, "num_tokens": 294622253.0, "step": 7761 }, { "epoch": 0.9874061824195395, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3717222213745117, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8709715604782104, "num_tokens": 294664842.0, "step": 7762 }, { "epoch": 0.98753339269813, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.338564872741699, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8558658957481384, "num_tokens": 294707135.0, "step": 7763 }, { "epoch": 0.9876606029767205, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3349454402923584, "learning_rate": 1e-05, "loss": 0.5065, "mean_token_accuracy": 0.8487751483917236, "num_tokens": 294751356.0, "step": 7764 }, { "epoch": 0.9877878132553111, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.297492027282715, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.855332612991333, "num_tokens": 294789444.0, "step": 7765 }, { "epoch": 0.9879150235339016, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.329413414001465, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.8532629609107971, "num_tokens": 294823247.0, "step": 7766 }, { "epoch": 0.988042233812492, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2379555702209473, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8514337539672852, "num_tokens": 294862932.0, "step": 7767 }, { "epoch": 0.9881694440910825, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.334923505783081, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8621251583099365, "num_tokens": 294902686.0, "step": 7768 }, { "epoch": 0.9882966543696731, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1831753253936768, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8690424561500549, "num_tokens": 294942028.0, "step": 7769 }, { "epoch": 0.9884238646482636, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3805880546569824, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8619847297668457, "num_tokens": 294977699.0, "step": 7770 }, { "epoch": 0.9885510749268541, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1334288120269775, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8606451153755188, "num_tokens": 295014441.0, "step": 7771 }, { "epoch": 0.9886782852054447, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4523236751556396, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8710982799530029, "num_tokens": 295053569.0, "step": 7772 }, { "epoch": 0.9888054954840351, "ewc_loss": 0.05908203125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.070160388946533, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8593605160713196, "num_tokens": 295083607.0, "step": 7773 }, { "epoch": 0.9889327057626256, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.57782244682312, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8514436483383179, "num_tokens": 295118074.0, "step": 7774 }, { "epoch": 0.9890599160412161, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.172325134277344e-05, "grad_norm": 2.991812229156494, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8641726970672607, "num_tokens": 295157926.0, "step": 7775 }, { "epoch": 0.9891871263198067, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.777928352355957, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8595845699310303, "num_tokens": 295201297.0, "step": 7776 }, { "epoch": 0.9893143365983972, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.96968936920166, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8563961982727051, "num_tokens": 295238178.0, "step": 7777 }, { "epoch": 0.9894415468769877, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.063726425170898, "learning_rate": 1e-05, "loss": 0.4794, "mean_token_accuracy": 0.8620275259017944, "num_tokens": 295273843.0, "step": 7778 }, { "epoch": 0.9895687571555781, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.14411997795105, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.866080641746521, "num_tokens": 295308703.0, "step": 7779 }, { "epoch": 0.9896959674341687, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.245485782623291, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8602152466773987, "num_tokens": 295347710.0, "step": 7780 }, { "epoch": 0.9898231777127592, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4888296127319336, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8704292178153992, "num_tokens": 295382255.0, "step": 7781 }, { "epoch": 0.9899503879913497, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.137884140014648, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8664125204086304, "num_tokens": 295418185.0, "step": 7782 }, { "epoch": 0.9900775982699402, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000179290771484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6659843921661377, "learning_rate": 1e-05, "loss": 0.55, "mean_token_accuracy": 0.8432735204696655, "num_tokens": 295455124.0, "step": 7783 }, { "epoch": 0.9902048085485308, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8980154991149902, "learning_rate": 1e-05, "loss": 0.5234, "mean_token_accuracy": 0.8495078682899475, "num_tokens": 295498058.0, "step": 7784 }, { "epoch": 0.9903320188271212, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.602898359298706, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8770274519920349, "num_tokens": 295534623.0, "step": 7785 }, { "epoch": 0.9904592291057117, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.704946756362915, "learning_rate": 1e-05, "loss": 0.5102, "mean_token_accuracy": 0.8534789085388184, "num_tokens": 295580064.0, "step": 7786 }, { "epoch": 0.9905864393843022, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4974558353424072, "learning_rate": 1e-05, "loss": 0.523, "mean_token_accuracy": 0.8496159315109253, "num_tokens": 295620720.0, "step": 7787 }, { "epoch": 0.9907136496628928, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5814547538757324, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8680328130722046, "num_tokens": 295662157.0, "step": 7788 }, { "epoch": 0.9908408599414833, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.391244411468506, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8612692356109619, "num_tokens": 295700215.0, "step": 7789 }, { "epoch": 0.9909680702200738, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5045361518859863, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8842676877975464, "num_tokens": 295738363.0, "step": 7790 }, { "epoch": 0.9910952804986642, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2978646755218506, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8742860555648804, "num_tokens": 295773911.0, "step": 7791 }, { "epoch": 0.9912224907772548, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.499983549118042, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8609665632247925, "num_tokens": 295809341.0, "step": 7792 }, { "epoch": 0.9913497010558453, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.191927433013916, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8520569205284119, "num_tokens": 295855952.0, "step": 7793 }, { "epoch": 0.9914769113344358, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.571103572845459, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8650846481323242, "num_tokens": 295892047.0, "step": 7794 }, { "epoch": 0.9916041216130264, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.1259114742279053, "learning_rate": 1e-05, "loss": 0.5244, "mean_token_accuracy": 0.8432166576385498, "num_tokens": 295926259.0, "step": 7795 }, { "epoch": 0.9917313318916169, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.687854051589966, "learning_rate": 1e-05, "loss": 0.49, "mean_token_accuracy": 0.8564144372940063, "num_tokens": 295965584.0, "step": 7796 }, { "epoch": 0.9918585421702073, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.074439525604248, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.870620846748352, "num_tokens": 296002754.0, "step": 7797 }, { "epoch": 0.9919857524487978, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.9045653343200684, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8601843118667603, "num_tokens": 296042161.0, "step": 7798 }, { "epoch": 0.9921129627273884, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1668343544006348, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.85185307264328, "num_tokens": 296079153.0, "step": 7799 }, { "epoch": 0.9922401730059789, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9822187423706055, "learning_rate": 1e-05, "loss": 0.5159, "mean_token_accuracy": 0.8518321514129639, "num_tokens": 296124064.0, "step": 7800 }, { "epoch": 0.9923673832845694, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3091959953308105, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8680386543273926, "num_tokens": 296166035.0, "step": 7801 }, { "epoch": 0.9924945935631599, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9567337036132812, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8776338696479797, "num_tokens": 296206566.0, "step": 7802 }, { "epoch": 0.9926218038417504, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.470791816711426, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8683610558509827, "num_tokens": 296243587.0, "step": 7803 }, { "epoch": 0.9927490141203409, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7778382301330566, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.8585947751998901, "num_tokens": 296282704.0, "step": 7804 }, { "epoch": 0.9928762243989314, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5142557621002197, "learning_rate": 1e-05, "loss": 0.5124, "mean_token_accuracy": 0.8460330367088318, "num_tokens": 296317964.0, "step": 7805 }, { "epoch": 0.993003434677522, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.568643808364868, "learning_rate": 1e-05, "loss": 0.5091, "mean_token_accuracy": 0.8532900810241699, "num_tokens": 296352861.0, "step": 7806 }, { "epoch": 0.9931306449561125, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4817872047424316, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8607853651046753, "num_tokens": 296391302.0, "step": 7807 }, { "epoch": 0.993257855234703, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3755555152893066, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.8564064502716064, "num_tokens": 296428026.0, "step": 7808 }, { "epoch": 0.9933850655132934, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.456712007522583, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.8615243434906006, "num_tokens": 296468866.0, "step": 7809 }, { "epoch": 0.993512275791884, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.242064952850342, "learning_rate": 1e-05, "loss": 0.4751, "mean_token_accuracy": 0.8624173402786255, "num_tokens": 296514806.0, "step": 7810 }, { "epoch": 0.9936394860704745, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4718587398529053, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8673509359359741, "num_tokens": 296552328.0, "step": 7811 }, { "epoch": 0.993766696349065, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.2177250385284424, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8561170101165771, "num_tokens": 296589908.0, "step": 7812 }, { "epoch": 0.9938939066276555, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4796197414398193, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8504315614700317, "num_tokens": 296631147.0, "step": 7813 }, { "epoch": 0.9940211169062461, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.118621349334717, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8535164594650269, "num_tokens": 296673633.0, "step": 7814 }, { "epoch": 0.9941483271848366, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5290918350219727, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8604024648666382, "num_tokens": 296714782.0, "step": 7815 }, { "epoch": 0.994275537463427, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0703325271606445, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8575960397720337, "num_tokens": 296751556.0, "step": 7816 }, { "epoch": 0.9944027477420175, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.662773847579956, "learning_rate": 1e-05, "loss": 0.5396, "mean_token_accuracy": 0.8418384790420532, "num_tokens": 296788319.0, "step": 7817 }, { "epoch": 0.9945299580206081, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0337934494018555, "learning_rate": 1e-05, "loss": 0.4941, "mean_token_accuracy": 0.8544137477874756, "num_tokens": 296823513.0, "step": 7818 }, { "epoch": 0.9946571682991986, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8271803855895996, "learning_rate": 1e-05, "loss": 0.5036, "mean_token_accuracy": 0.8553643226623535, "num_tokens": 296862059.0, "step": 7819 }, { "epoch": 0.9947843785777891, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0715065002441406, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8654916882514954, "num_tokens": 296898051.0, "step": 7820 }, { "epoch": 0.9949115888563796, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9902608394622803, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8663453459739685, "num_tokens": 296935266.0, "step": 7821 }, { "epoch": 0.9950387991349701, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2169711589813232, "learning_rate": 1e-05, "loss": 0.5265, "mean_token_accuracy": 0.8477627635002136, "num_tokens": 296971653.0, "step": 7822 }, { "epoch": 0.9951660094135606, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.048578262329102, "learning_rate": 1e-05, "loss": 0.5727, "mean_token_accuracy": 0.8363163471221924, "num_tokens": 297015825.0, "step": 7823 }, { "epoch": 0.9952932196921511, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.407069444656372, "learning_rate": 1e-05, "loss": 0.5257, "mean_token_accuracy": 0.8486090302467346, "num_tokens": 297059115.0, "step": 7824 }, { "epoch": 0.9954204299707416, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9482388496398926, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8552974462509155, "num_tokens": 297092634.0, "step": 7825 }, { "epoch": 0.9955476402493322, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5055177211761475, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8608174324035645, "num_tokens": 297137154.0, "step": 7826 }, { "epoch": 0.9956748505279227, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7585067749023438, "learning_rate": 1e-05, "loss": 0.5238, "mean_token_accuracy": 0.8504283428192139, "num_tokens": 297175029.0, "step": 7827 }, { "epoch": 0.9958020608065131, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5025484561920166, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8641226887702942, "num_tokens": 297207397.0, "step": 7828 }, { "epoch": 0.9959292710851037, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.567298412322998, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8572268486022949, "num_tokens": 297250114.0, "step": 7829 }, { "epoch": 0.9960564813636942, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.434999465942383, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8663047552108765, "num_tokens": 297289765.0, "step": 7830 }, { "epoch": 0.9961836916422847, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.441601276397705, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8818078637123108, "num_tokens": 297327359.0, "step": 7831 }, { "epoch": 0.9963109019208752, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3940465450286865, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8657015562057495, "num_tokens": 297360851.0, "step": 7832 }, { "epoch": 0.9964381121994658, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3626108169555664, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8587310314178467, "num_tokens": 297396957.0, "step": 7833 }, { "epoch": 0.9965653224780562, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.349412441253662, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.861667275428772, "num_tokens": 297427900.0, "step": 7834 }, { "epoch": 0.9966925327566467, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.301933526992798, "learning_rate": 1e-05, "loss": 0.5094, "mean_token_accuracy": 0.8489209413528442, "num_tokens": 297464910.0, "step": 7835 }, { "epoch": 0.9968197430352372, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.307891845703125, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8739458322525024, "num_tokens": 297497585.0, "step": 7836 }, { "epoch": 0.9969469533138278, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2951056957244873, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.852365255355835, "num_tokens": 297537074.0, "step": 7837 }, { "epoch": 0.9970741635924183, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2940127849578857, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8703925609588623, "num_tokens": 297570848.0, "step": 7838 }, { "epoch": 0.9972013738710088, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2650105953216553, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8506778478622437, "num_tokens": 297607520.0, "step": 7839 }, { "epoch": 0.9973285841495992, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2597312927246094, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.865494430065155, "num_tokens": 297648442.0, "step": 7840 }, { "epoch": 0.9974557944281898, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2664525508880615, "learning_rate": 1e-05, "loss": 0.5292, "mean_token_accuracy": 0.8412332534790039, "num_tokens": 297686163.0, "step": 7841 }, { "epoch": 0.9975830047067803, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.233774185180664, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8723987340927124, "num_tokens": 297724191.0, "step": 7842 }, { "epoch": 0.9977102149853708, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2780113220214844, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8546286225318909, "num_tokens": 297766205.0, "step": 7843 }, { "epoch": 0.9978374252639614, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.214628219604492, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8704129457473755, "num_tokens": 297802153.0, "step": 7844 }, { "epoch": 0.9979646355425519, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.291186809539795, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8684651851654053, "num_tokens": 297841252.0, "step": 7845 }, { "epoch": 0.9980918458211423, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.176746368408203, "learning_rate": 1e-05, "loss": 0.3766, "mean_token_accuracy": 0.8897227048873901, "num_tokens": 297875460.0, "step": 7846 }, { "epoch": 0.9982190560997328, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.364711046218872, "learning_rate": 1e-05, "loss": 0.4824, "mean_token_accuracy": 0.8585326075553894, "num_tokens": 297908534.0, "step": 7847 }, { "epoch": 0.9983462663783234, "ewc_loss": 0.060546875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.1177051067352295, "learning_rate": 1e-05, "loss": 0.5286, "mean_token_accuracy": 0.8388142585754395, "num_tokens": 297950798.0, "step": 7848 }, { "epoch": 0.9984734766569139, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4293785095214844, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8673956990242004, "num_tokens": 297989696.0, "step": 7849 }, { "epoch": 0.9986006869355044, "ewc_loss": 0.059326171875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.124641418457031e-05, "grad_norm": 3.028251886367798, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8574538230895996, "num_tokens": 298033702.0, "step": 7850 }, { "epoch": 0.9987278972140949, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.577951431274414, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8773743510246277, "num_tokens": 298071867.0, "step": 7851 }, { "epoch": 0.9988551074926854, "ewc_loss": 0.0595703125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.1484832763671875e-05, "grad_norm": 2.954484462738037, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8503586053848267, "num_tokens": 298108726.0, "step": 7852 }, { "epoch": 0.9989823177712759, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.841660499572754, "learning_rate": 1e-05, "loss": 0.5117, "mean_token_accuracy": 0.8561753034591675, "num_tokens": 298145174.0, "step": 7853 }, { "epoch": 0.9991095280498664, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.922790765762329, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8733165860176086, "num_tokens": 298181300.0, "step": 7854 }, { "epoch": 0.9992367383284569, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.201963901519775, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.861993670463562, "num_tokens": 298220592.0, "step": 7855 }, { "epoch": 0.9993639486070475, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.1798434257507324, "learning_rate": 1e-05, "loss": 0.5094, "mean_token_accuracy": 0.8563361763954163, "num_tokens": 298260796.0, "step": 7856 }, { "epoch": 0.999491158885638, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.426165580749512, "learning_rate": 1e-05, "loss": 0.5418, "mean_token_accuracy": 0.844744086265564, "num_tokens": 298300640.0, "step": 7857 }, { "epoch": 0.9996183691642284, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.6664271354675293, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8758251667022705, "num_tokens": 298336433.0, "step": 7858 }, { "epoch": 0.9997455794428189, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.232496738433838, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8735431432723999, "num_tokens": 298370355.0, "step": 7859 }, { "epoch": 0.9998727897214095, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.8239426612854004, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8613348007202148, "num_tokens": 298406262.0, "step": 7860 }, { "epoch": 1.0, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9668822288513184, "learning_rate": 1e-05, "loss": 0.5664, "mean_token_accuracy": 0.8420770168304443, "num_tokens": 298443310.0, "step": 7861 }, { "epoch": 1.0001272102785905, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6845829486846924, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8633538484573364, "num_tokens": 298482309.0, "step": 7862 }, { "epoch": 1.000254420557181, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7848165035247803, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8627301454544067, "num_tokens": 298522555.0, "step": 7863 }, { "epoch": 1.0003816308357716, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5277199745178223, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8823200464248657, "num_tokens": 298559559.0, "step": 7864 }, { "epoch": 1.0005088411143621, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.689375400543213, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.867716908454895, "num_tokens": 298596604.0, "step": 7865 }, { "epoch": 1.0006360513929526, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3810713291168213, "learning_rate": 1e-05, "loss": 0.397, "mean_token_accuracy": 0.8842303156852722, "num_tokens": 298634590.0, "step": 7866 }, { "epoch": 1.0007632616715432, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6701533794403076, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8771764039993286, "num_tokens": 298670714.0, "step": 7867 }, { "epoch": 1.0008904719501335, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.262601137161255, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8684399127960205, "num_tokens": 298713924.0, "step": 7868 }, { "epoch": 1.001017682228724, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7406463623046875, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8711600303649902, "num_tokens": 298749537.0, "step": 7869 }, { "epoch": 1.0011448925073145, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1877505779266357, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8617538213729858, "num_tokens": 298789887.0, "step": 7870 }, { "epoch": 1.001272102785905, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8766603469848633, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8719327449798584, "num_tokens": 298826790.0, "step": 7871 }, { "epoch": 1.0013993130644956, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2266085147857666, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8620660305023193, "num_tokens": 298860427.0, "step": 7872 }, { "epoch": 1.0015265233430861, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.999211311340332, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8696659803390503, "num_tokens": 298892627.0, "step": 7873 }, { "epoch": 1.0016537336216766, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3520026206970215, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8571876883506775, "num_tokens": 298930513.0, "step": 7874 }, { "epoch": 1.0017809439002672, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9597768783569336, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.888483464717865, "num_tokens": 298971234.0, "step": 7875 }, { "epoch": 1.0019081541788577, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.475102186203003, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8733510971069336, "num_tokens": 299013224.0, "step": 7876 }, { "epoch": 1.0020353644574482, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8262016773223877, "learning_rate": 1e-05, "loss": 0.5265, "mean_token_accuracy": 0.8440201878547668, "num_tokens": 299055093.0, "step": 7877 }, { "epoch": 1.0021625747360388, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.625145196914673, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8811851739883423, "num_tokens": 299085368.0, "step": 7878 }, { "epoch": 1.0022897850146293, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5366334915161133, "learning_rate": 1e-05, "loss": 0.4956, "mean_token_accuracy": 0.8621137738227844, "num_tokens": 299128279.0, "step": 7879 }, { "epoch": 1.0024169952932196, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.531532049179077, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8698334693908691, "num_tokens": 299161714.0, "step": 7880 }, { "epoch": 1.0025442055718101, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3310775756835938, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8588905334472656, "num_tokens": 299200565.0, "step": 7881 }, { "epoch": 1.0026714158504006, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.517026424407959, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8690721392631531, "num_tokens": 299242242.0, "step": 7882 }, { "epoch": 1.0027986261289912, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1982247829437256, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.881848931312561, "num_tokens": 299278141.0, "step": 7883 }, { "epoch": 1.0029258364075817, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5607125759124756, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.874150812625885, "num_tokens": 299312413.0, "step": 7884 }, { "epoch": 1.0030530466861722, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.100663661956787, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.877241313457489, "num_tokens": 299348158.0, "step": 7885 }, { "epoch": 1.0031802569647628, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018024444580078125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6534018516540527, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8750075101852417, "num_tokens": 299386963.0, "step": 7886 }, { "epoch": 1.0033074672433533, "ewc_loss": 0.060791015625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0414774417877197, "learning_rate": 1e-05, "loss": 0.4157, "mean_token_accuracy": 0.8768369555473328, "num_tokens": 299424929.0, "step": 7887 }, { "epoch": 1.0034346775219438, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8506479263305664, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8633570075035095, "num_tokens": 299461588.0, "step": 7888 }, { "epoch": 1.0035618878005343, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0857834815979004, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8779062032699585, "num_tokens": 299498226.0, "step": 7889 }, { "epoch": 1.0036890980791249, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.019859313964844, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8726841807365417, "num_tokens": 299536806.0, "step": 7890 }, { "epoch": 1.0038163083577154, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2431576251983643, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8653503060340881, "num_tokens": 299583658.0, "step": 7891 }, { "epoch": 1.0039435186363057, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.084905624389648, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8629964590072632, "num_tokens": 299621381.0, "step": 7892 }, { "epoch": 1.0040707289148962, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4757230281829834, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8643542528152466, "num_tokens": 299661055.0, "step": 7893 }, { "epoch": 1.0041979391934868, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9311563968658447, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8733530044555664, "num_tokens": 299702500.0, "step": 7894 }, { "epoch": 1.0043251494720773, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5392537117004395, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8684167861938477, "num_tokens": 299742590.0, "step": 7895 }, { "epoch": 1.0044523597506678, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7462291717529297, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8674699068069458, "num_tokens": 299777716.0, "step": 7896 }, { "epoch": 1.0045795700292584, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5318970680236816, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8816319704055786, "num_tokens": 299808829.0, "step": 7897 }, { "epoch": 1.0047067803078489, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.556703567504883, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8702977299690247, "num_tokens": 299845582.0, "step": 7898 }, { "epoch": 1.0048339905864394, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.440688371658325, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8729773759841919, "num_tokens": 299886373.0, "step": 7899 }, { "epoch": 1.00496120086503, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4478390216827393, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8719726800918579, "num_tokens": 299917464.0, "step": 7900 }, { "epoch": 1.0050884111436205, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.381351947784424, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8699406385421753, "num_tokens": 299953455.0, "step": 7901 }, { "epoch": 1.005215621422211, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3814139366149902, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8600512742996216, "num_tokens": 299993707.0, "step": 7902 }, { "epoch": 1.0053428317008015, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3118889331817627, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8671528100967407, "num_tokens": 300033021.0, "step": 7903 }, { "epoch": 1.0054700419793918, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3535802364349365, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8679971694946289, "num_tokens": 300072497.0, "step": 7904 }, { "epoch": 1.0055972522579824, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2547006607055664, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8781344294548035, "num_tokens": 300113510.0, "step": 7905 }, { "epoch": 1.0057244625365729, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.366770029067993, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8664664030075073, "num_tokens": 300151524.0, "step": 7906 }, { "epoch": 1.0058516728151634, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2058563232421875, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8696183562278748, "num_tokens": 300184281.0, "step": 7907 }, { "epoch": 1.005978883093754, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.401087522506714, "learning_rate": 1e-05, "loss": 0.4941, "mean_token_accuracy": 0.8581772446632385, "num_tokens": 300222772.0, "step": 7908 }, { "epoch": 1.0061060933723445, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.141766309738159, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8625528812408447, "num_tokens": 300257347.0, "step": 7909 }, { "epoch": 1.006233303650935, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.462764263153076, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8637409210205078, "num_tokens": 300300678.0, "step": 7910 }, { "epoch": 1.0063605139295255, "ewc_loss": 0.060302734375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.067007303237915, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8754244446754456, "num_tokens": 300339851.0, "step": 7911 }, { "epoch": 1.006487724208116, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5857903957366943, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.878475546836853, "num_tokens": 300379753.0, "step": 7912 }, { "epoch": 1.0066149344867066, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.013169765472412, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8704087734222412, "num_tokens": 300418235.0, "step": 7913 }, { "epoch": 1.006742144765297, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7888143062591553, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8820354342460632, "num_tokens": 300450237.0, "step": 7914 }, { "epoch": 1.0068693550438876, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 2.99876070022583, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8696029782295227, "num_tokens": 300489416.0, "step": 7915 }, { "epoch": 1.0069965653224782, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.040070533752441, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8775202035903931, "num_tokens": 300526180.0, "step": 7916 }, { "epoch": 1.0071237756010685, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.1748595237731934, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8702167272567749, "num_tokens": 300566327.0, "step": 7917 }, { "epoch": 1.007250985879659, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.183410167694092, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.862582266330719, "num_tokens": 300610136.0, "step": 7918 }, { "epoch": 1.0073781961582495, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.474806785583496, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8677729368209839, "num_tokens": 300647238.0, "step": 7919 }, { "epoch": 1.00750540643684, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.052393436431885, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8728847503662109, "num_tokens": 300684291.0, "step": 7920 }, { "epoch": 1.0076326167154306, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.611316680908203, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8645167350769043, "num_tokens": 300718919.0, "step": 7921 }, { "epoch": 1.0077598269940211, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8327810764312744, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8660786151885986, "num_tokens": 300752353.0, "step": 7922 }, { "epoch": 1.0078870372726116, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5609686374664307, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8739965558052063, "num_tokens": 300789214.0, "step": 7923 }, { "epoch": 1.0080142475512022, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.642383098602295, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8623660802841187, "num_tokens": 300825629.0, "step": 7924 }, { "epoch": 1.0081414578297927, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4629502296447754, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8838735818862915, "num_tokens": 300863908.0, "step": 7925 }, { "epoch": 1.0082686681083832, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.521482229232788, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8648847341537476, "num_tokens": 300905458.0, "step": 7926 }, { "epoch": 1.0083958783869738, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.38193678855896, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8831284642219543, "num_tokens": 300943906.0, "step": 7927 }, { "epoch": 1.0085230886655643, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4547646045684814, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8737964034080505, "num_tokens": 300979371.0, "step": 7928 }, { "epoch": 1.0086502989441546, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3004508018493652, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8682453036308289, "num_tokens": 301017575.0, "step": 7929 }, { "epoch": 1.0087775092227451, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4423105716705322, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8717130422592163, "num_tokens": 301053130.0, "step": 7930 }, { "epoch": 1.0089047195013356, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.233205795288086, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.882027804851532, "num_tokens": 301090740.0, "step": 7931 }, { "epoch": 1.0090319297799262, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.457942247390747, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8708025217056274, "num_tokens": 301133539.0, "step": 7932 }, { "epoch": 1.0091591400585167, "ewc_loss": 0.061767578125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1864001750946045, "learning_rate": 1e-05, "loss": 0.5092, "mean_token_accuracy": 0.8483076095581055, "num_tokens": 301168243.0, "step": 7933 }, { "epoch": 1.0092863503371072, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.501715898513794, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8754754066467285, "num_tokens": 301204444.0, "step": 7934 }, { "epoch": 1.0094135606156978, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.092282295227051, "learning_rate": 1e-05, "loss": 0.3939, "mean_token_accuracy": 0.886088490486145, "num_tokens": 301245768.0, "step": 7935 }, { "epoch": 1.0095407708942883, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6195170879364014, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8568096160888672, "num_tokens": 301282381.0, "step": 7936 }, { "epoch": 1.0096679811728788, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0396926403045654, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8800556659698486, "num_tokens": 301322538.0, "step": 7937 }, { "epoch": 1.0097951914514693, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.809065818786621, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8736352920532227, "num_tokens": 301357527.0, "step": 7938 }, { "epoch": 1.0099224017300599, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.0678224563598633, "learning_rate": 1e-05, "loss": 0.5044, "mean_token_accuracy": 0.8530284762382507, "num_tokens": 301399609.0, "step": 7939 }, { "epoch": 1.0100496120086504, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.997624397277832, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8702967166900635, "num_tokens": 301439271.0, "step": 7940 }, { "epoch": 1.0101768222872407, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.227191209793091, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8705635070800781, "num_tokens": 301476895.0, "step": 7941 }, { "epoch": 1.0103040325658312, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.050418853759766, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8617205023765564, "num_tokens": 301519649.0, "step": 7942 }, { "epoch": 1.0104312428444218, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4379758834838867, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.8606775403022766, "num_tokens": 301557132.0, "step": 7943 }, { "epoch": 1.0105584531230123, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.910688877105713, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8667229413986206, "num_tokens": 301603604.0, "step": 7944 }, { "epoch": 1.0106856634016028, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.515763998031616, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8593810200691223, "num_tokens": 301643797.0, "step": 7945 }, { "epoch": 1.0108128736801933, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.720684766769409, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.875856339931488, "num_tokens": 301676658.0, "step": 7946 }, { "epoch": 1.0109400839587839, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001811981201171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.4949755668640137, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8790885210037231, "num_tokens": 301714749.0, "step": 7947 }, { "epoch": 1.0110672942373744, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.529059410095215, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8731139898300171, "num_tokens": 301754116.0, "step": 7948 }, { "epoch": 1.011194504515965, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4482808113098145, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8805776238441467, "num_tokens": 301791441.0, "step": 7949 }, { "epoch": 1.0113217147945555, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4032840728759766, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8595492839813232, "num_tokens": 301830971.0, "step": 7950 }, { "epoch": 1.011448925073146, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4060332775115967, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8662213683128357, "num_tokens": 301867848.0, "step": 7951 }, { "epoch": 1.0115761353517365, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.312098264694214, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8684252500534058, "num_tokens": 301912700.0, "step": 7952 }, { "epoch": 1.0117033456303268, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.38149094581604, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8750818967819214, "num_tokens": 301945890.0, "step": 7953 }, { "epoch": 1.0118305559089174, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2603089809417725, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8799291253089905, "num_tokens": 301980095.0, "step": 7954 }, { "epoch": 1.0119577661875079, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3790595531463623, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8705652952194214, "num_tokens": 302011855.0, "step": 7955 }, { "epoch": 1.0120849764660984, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.205648899078369, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8680765628814697, "num_tokens": 302048299.0, "step": 7956 }, { "epoch": 1.012212186744689, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3896584510803223, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8681089878082275, "num_tokens": 302082628.0, "step": 7957 }, { "epoch": 1.0123393970232795, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1530873775482178, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8806142210960388, "num_tokens": 302119832.0, "step": 7958 }, { "epoch": 1.01246660730187, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4351022243499756, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.874942421913147, "num_tokens": 302162930.0, "step": 7959 }, { "epoch": 1.0125938175804605, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.097943067550659, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8729060888290405, "num_tokens": 302202250.0, "step": 7960 }, { "epoch": 1.012721027859051, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5410361289978027, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8660544157028198, "num_tokens": 302239446.0, "step": 7961 }, { "epoch": 1.0128482381376416, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.013580322265625, "learning_rate": 1e-05, "loss": 0.393, "mean_token_accuracy": 0.8849515318870544, "num_tokens": 302279949.0, "step": 7962 }, { "epoch": 1.012975448416232, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.729109525680542, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8631936311721802, "num_tokens": 302318128.0, "step": 7963 }, { "epoch": 1.0131026586948226, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.9859251976013184, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8698372840881348, "num_tokens": 302357353.0, "step": 7964 }, { "epoch": 1.0132298689734132, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9714348316192627, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.873655378818512, "num_tokens": 302399843.0, "step": 7965 }, { "epoch": 1.0133570792520035, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.076287269592285, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8691447973251343, "num_tokens": 302440081.0, "step": 7966 }, { "epoch": 1.013484289530594, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.200170993804932, "learning_rate": 1e-05, "loss": 0.4961, "mean_token_accuracy": 0.8597955703735352, "num_tokens": 302479061.0, "step": 7967 }, { "epoch": 1.0136114998091845, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.378474712371826, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8795613050460815, "num_tokens": 302514490.0, "step": 7968 }, { "epoch": 1.013738710087775, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.175626754760742, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8758770227432251, "num_tokens": 302552555.0, "step": 7969 }, { "epoch": 1.0138659203663656, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.620224714279175, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8595433235168457, "num_tokens": 302589296.0, "step": 7970 }, { "epoch": 1.013993130644956, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.96085786819458, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8864749670028687, "num_tokens": 302626085.0, "step": 7971 }, { "epoch": 1.0141203409235466, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6192727088928223, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8727746605873108, "num_tokens": 302662671.0, "step": 7972 }, { "epoch": 1.0142475512021372, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7531094551086426, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8733624219894409, "num_tokens": 302706112.0, "step": 7973 }, { "epoch": 1.0143747614807277, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5303761959075928, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8588193655014038, "num_tokens": 302744942.0, "step": 7974 }, { "epoch": 1.0145019717593182, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5950536727905273, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8841772079467773, "num_tokens": 302784085.0, "step": 7975 }, { "epoch": 1.0146291820379088, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4284932613372803, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8751522302627563, "num_tokens": 302821336.0, "step": 7976 }, { "epoch": 1.0147563923164993, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5169975757598877, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8711321353912354, "num_tokens": 302857245.0, "step": 7977 }, { "epoch": 1.0148836025950896, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3372292518615723, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8734506368637085, "num_tokens": 302894115.0, "step": 7978 }, { "epoch": 1.0150108128736801, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4669618606567383, "learning_rate": 1e-05, "loss": 0.3724, "mean_token_accuracy": 0.89122474193573, "num_tokens": 302930162.0, "step": 7979 }, { "epoch": 1.0151380231522706, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.246502637863159, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8717467188835144, "num_tokens": 302965184.0, "step": 7980 }, { "epoch": 1.0152652334308612, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.495450735092163, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8646098375320435, "num_tokens": 303003895.0, "step": 7981 }, { "epoch": 1.0153924437094517, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.170992612838745, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.870831310749054, "num_tokens": 303038328.0, "step": 7982 }, { "epoch": 1.0155196539880422, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.594599962234497, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8487583994865417, "num_tokens": 303075546.0, "step": 7983 }, { "epoch": 1.0156468642666328, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.085552453994751, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8680088520050049, "num_tokens": 303111825.0, "step": 7984 }, { "epoch": 1.0157740745452233, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.722862958908081, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8789557218551636, "num_tokens": 303147607.0, "step": 7985 }, { "epoch": 1.0159012848238138, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0624771118164062, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8661666512489319, "num_tokens": 303185646.0, "step": 7986 }, { "epoch": 1.0160284951024043, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9201431274414062, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8738865852355957, "num_tokens": 303218723.0, "step": 7987 }, { "epoch": 1.0161557053809949, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1566293239593506, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8662437200546265, "num_tokens": 303259012.0, "step": 7988 }, { "epoch": 1.0162829156595854, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.048532962799072, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8696433901786804, "num_tokens": 303294868.0, "step": 7989 }, { "epoch": 1.0164101259381757, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00018215179443359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.377410650253296, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.871420681476593, "num_tokens": 303331987.0, "step": 7990 }, { "epoch": 1.0165373362167662, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9775283336639404, "learning_rate": 1e-05, "loss": 0.3841, "mean_token_accuracy": 0.8916154503822327, "num_tokens": 303364230.0, "step": 7991 }, { "epoch": 1.0166645464953568, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.515665054321289, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.879754900932312, "num_tokens": 303404169.0, "step": 7992 }, { "epoch": 1.0167917567739473, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7681775093078613, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8747245073318481, "num_tokens": 303445137.0, "step": 7993 }, { "epoch": 1.0169189670525378, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5221452713012695, "learning_rate": 1e-05, "loss": 0.4033, "mean_token_accuracy": 0.8851423263549805, "num_tokens": 303486515.0, "step": 7994 }, { "epoch": 1.0170461773311283, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5812788009643555, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8694469928741455, "num_tokens": 303521514.0, "step": 7995 }, { "epoch": 1.0171733876097189, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.505542755126953, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8685917854309082, "num_tokens": 303555606.0, "step": 7996 }, { "epoch": 1.0173005978883094, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3972699642181396, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.871016263961792, "num_tokens": 303592075.0, "step": 7997 }, { "epoch": 1.0174278081669, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4536983966827393, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8520702123641968, "num_tokens": 303632529.0, "step": 7998 }, { "epoch": 1.0175550184454905, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2924227714538574, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8762832880020142, "num_tokens": 303665419.0, "step": 7999 }, { "epoch": 1.017682228724081, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4444737434387207, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8694263696670532, "num_tokens": 303698778.0, "step": 8000 }, { "epoch": 1.0178094390026715, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1966397762298584, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8704262971878052, "num_tokens": 303738713.0, "step": 8001 }, { "epoch": 1.0179366492812618, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4772417545318604, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8703975677490234, "num_tokens": 303767821.0, "step": 8002 }, { "epoch": 1.0180638595598523, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.120393753051758, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8719338178634644, "num_tokens": 303813088.0, "step": 8003 }, { "epoch": 1.0181910698384429, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.557068347930908, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8543738126754761, "num_tokens": 303849926.0, "step": 8004 }, { "epoch": 1.0183182801170334, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.061509609222412, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8756303787231445, "num_tokens": 303888681.0, "step": 8005 }, { "epoch": 1.018445490395624, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6917812824249268, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8736292123794556, "num_tokens": 303926042.0, "step": 8006 }, { "epoch": 1.0185727006742145, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0352914333343506, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8817339539527893, "num_tokens": 303962014.0, "step": 8007 }, { "epoch": 1.018699910952805, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8754966259002686, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8732423782348633, "num_tokens": 303998856.0, "step": 8008 }, { "epoch": 1.0188271212313955, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1015913486480713, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8756510615348816, "num_tokens": 304034228.0, "step": 8009 }, { "epoch": 1.018954331509986, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0431413650512695, "learning_rate": 1e-05, "loss": 0.5401, "mean_token_accuracy": 0.8515164256095886, "num_tokens": 304070477.0, "step": 8010 }, { "epoch": 1.0190815417885766, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2696518898010254, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8766751289367676, "num_tokens": 304116073.0, "step": 8011 }, { "epoch": 1.019208752067167, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.080042362213135, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8623605370521545, "num_tokens": 304149950.0, "step": 8012 }, { "epoch": 1.0193359623457576, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4647529125213623, "learning_rate": 1e-05, "loss": 0.3981, "mean_token_accuracy": 0.8887541890144348, "num_tokens": 304191824.0, "step": 8013 }, { "epoch": 1.0194631726243482, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.953132152557373, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8837285041809082, "num_tokens": 304223616.0, "step": 8014 }, { "epoch": 1.0195903829029385, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5535595417022705, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8691766858100891, "num_tokens": 304261120.0, "step": 8015 }, { "epoch": 1.019717593181529, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7529985904693604, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.8580620288848877, "num_tokens": 304294796.0, "step": 8016 }, { "epoch": 1.0198448034601195, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.514876365661621, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8612112998962402, "num_tokens": 304333396.0, "step": 8017 }, { "epoch": 1.01997201373871, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5670223236083984, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8726279735565186, "num_tokens": 304371246.0, "step": 8018 }, { "epoch": 1.0200992240173006, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.439120292663574, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8841853737831116, "num_tokens": 304408179.0, "step": 8019 }, { "epoch": 1.020226434295891, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.449730157852173, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.874240517616272, "num_tokens": 304442758.0, "step": 8020 }, { "epoch": 1.0203536445744816, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.376944065093994, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8809640407562256, "num_tokens": 304479568.0, "step": 8021 }, { "epoch": 1.0204808548530722, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3740766048431396, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8709805011749268, "num_tokens": 304521255.0, "step": 8022 }, { "epoch": 1.0206080651316627, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.322500705718994, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8721606135368347, "num_tokens": 304553569.0, "step": 8023 }, { "epoch": 1.0207352754102532, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.345712900161743, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8669095039367676, "num_tokens": 304594632.0, "step": 8024 }, { "epoch": 1.0208624856888437, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2730419635772705, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8688158988952637, "num_tokens": 304635255.0, "step": 8025 }, { "epoch": 1.0209896959674343, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3358442783355713, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8717068433761597, "num_tokens": 304674472.0, "step": 8026 }, { "epoch": 1.0211169062460246, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.23384165763855, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8786770105361938, "num_tokens": 304708950.0, "step": 8027 }, { "epoch": 1.021244116524615, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3389627933502197, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8693919777870178, "num_tokens": 304752849.0, "step": 8028 }, { "epoch": 1.0213713268032056, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1926965713500977, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8689430356025696, "num_tokens": 304791936.0, "step": 8029 }, { "epoch": 1.0214985370817962, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3792245388031006, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8728132247924805, "num_tokens": 304832228.0, "step": 8030 }, { "epoch": 1.0216257473603867, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.147312879562378, "learning_rate": 1e-05, "loss": 0.3798, "mean_token_accuracy": 0.8882408142089844, "num_tokens": 304863169.0, "step": 8031 }, { "epoch": 1.0217529576389772, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4492239952087402, "learning_rate": 1e-05, "loss": 0.5332, "mean_token_accuracy": 0.8385595083236694, "num_tokens": 304904938.0, "step": 8032 }, { "epoch": 1.0218801679175677, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0809497833251953, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8712282776832581, "num_tokens": 304940590.0, "step": 8033 }, { "epoch": 1.0220073781961583, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.577061414718628, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8668838739395142, "num_tokens": 304976945.0, "step": 8034 }, { "epoch": 1.0221345884747488, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.9897947311401367, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.871878981590271, "num_tokens": 305021407.0, "step": 8035 }, { "epoch": 1.0222617987533393, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.831362247467041, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8783747553825378, "num_tokens": 305053141.0, "step": 8036 }, { "epoch": 1.0223890090319299, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 2.9840354919433594, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8768619298934937, "num_tokens": 305086795.0, "step": 8037 }, { "epoch": 1.0225162193105204, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.113687038421631, "learning_rate": 1e-05, "loss": 0.5046, "mean_token_accuracy": 0.8577769994735718, "num_tokens": 305125759.0, "step": 8038 }, { "epoch": 1.0226434295891107, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.176405668258667, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8717268109321594, "num_tokens": 305165169.0, "step": 8039 }, { "epoch": 1.0227706398677012, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.270794868469238, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8884749412536621, "num_tokens": 305203457.0, "step": 8040 }, { "epoch": 1.0228978501462918, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.5357275009155273, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8620876669883728, "num_tokens": 305241021.0, "step": 8041 }, { "epoch": 1.0230250604248823, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.140813827514648, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8507546782493591, "num_tokens": 305277579.0, "step": 8042 }, { "epoch": 1.0231522707034728, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6810288429260254, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8624095916748047, "num_tokens": 305318027.0, "step": 8043 }, { "epoch": 1.0232794809820633, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.922175645828247, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8718647956848145, "num_tokens": 305351249.0, "step": 8044 }, { "epoch": 1.0234066912606539, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6151483058929443, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8728722333908081, "num_tokens": 305387906.0, "step": 8045 }, { "epoch": 1.0235339015392444, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7372052669525146, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.87909334897995, "num_tokens": 305419511.0, "step": 8046 }, { "epoch": 1.023661111817835, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.500659227371216, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8705316185951233, "num_tokens": 305457483.0, "step": 8047 }, { "epoch": 1.0237883220964255, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018310546875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6308488845825195, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8743921518325806, "num_tokens": 305486498.0, "step": 8048 }, { "epoch": 1.023915532375016, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.3858284950256348, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8703681230545044, "num_tokens": 305525250.0, "step": 8049 }, { "epoch": 1.0240427426536065, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5666396617889404, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.867266058921814, "num_tokens": 305566229.0, "step": 8050 }, { "epoch": 1.0241699529321968, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.278780937194824, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8782458305358887, "num_tokens": 305601884.0, "step": 8051 }, { "epoch": 1.0242971632107873, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.585259199142456, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8672257661819458, "num_tokens": 305644495.0, "step": 8052 }, { "epoch": 1.0244243734893779, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1872639656066895, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8758814334869385, "num_tokens": 305680662.0, "step": 8053 }, { "epoch": 1.0245515837679684, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6608476638793945, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8759052753448486, "num_tokens": 305720821.0, "step": 8054 }, { "epoch": 1.024678794046559, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1316661834716797, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8617892265319824, "num_tokens": 305757312.0, "step": 8055 }, { "epoch": 1.0248060043251495, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7949788570404053, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8624045848846436, "num_tokens": 305794372.0, "step": 8056 }, { "epoch": 1.02493321460374, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.129986524581909, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8707963824272156, "num_tokens": 305839975.0, "step": 8057 }, { "epoch": 1.0250604248823305, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.936981678009033, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8791970014572144, "num_tokens": 305877964.0, "step": 8058 }, { "epoch": 1.025187635160921, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.267827033996582, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8708817958831787, "num_tokens": 305911077.0, "step": 8059 }, { "epoch": 1.0253148454395116, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.969933032989502, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8765795826911926, "num_tokens": 305947741.0, "step": 8060 }, { "epoch": 1.025442055718102, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4220545291900635, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.868017315864563, "num_tokens": 305986872.0, "step": 8061 }, { "epoch": 1.0255692659966926, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8308401107788086, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8636614680290222, "num_tokens": 306023437.0, "step": 8062 }, { "epoch": 1.0256964762752832, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.491070032119751, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8809283375740051, "num_tokens": 306063026.0, "step": 8063 }, { "epoch": 1.0258236865538735, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6223483085632324, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8642697930335999, "num_tokens": 306102574.0, "step": 8064 }, { "epoch": 1.025950896832464, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4953904151916504, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8630756139755249, "num_tokens": 306137194.0, "step": 8065 }, { "epoch": 1.0260781071110545, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.444981813430786, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8689727783203125, "num_tokens": 306171539.0, "step": 8066 }, { "epoch": 1.026205317389645, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4669477939605713, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8673710823059082, "num_tokens": 306212385.0, "step": 8067 }, { "epoch": 1.0263325276682356, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.287555456161499, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8722425699234009, "num_tokens": 306247624.0, "step": 8068 }, { "epoch": 1.026459737946826, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4771475791931152, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8568159937858582, "num_tokens": 306286445.0, "step": 8069 }, { "epoch": 1.0265869482254166, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1743998527526855, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8774644136428833, "num_tokens": 306327792.0, "step": 8070 }, { "epoch": 1.0267141585040072, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5309321880340576, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8535921573638916, "num_tokens": 306366811.0, "step": 8071 }, { "epoch": 1.0268413687825977, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.106651544570923, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8665387630462646, "num_tokens": 306403938.0, "step": 8072 }, { "epoch": 1.0269685790611882, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.613259792327881, "learning_rate": 1e-05, "loss": 0.4026, "mean_token_accuracy": 0.8822554349899292, "num_tokens": 306440623.0, "step": 8073 }, { "epoch": 1.0270957893397787, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.047147512435913, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8696985244750977, "num_tokens": 306477623.0, "step": 8074 }, { "epoch": 1.0272229996183693, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.777888774871826, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8692985773086548, "num_tokens": 306513586.0, "step": 8075 }, { "epoch": 1.0273502098969596, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0512990951538086, "learning_rate": 1e-05, "loss": 0.3942, "mean_token_accuracy": 0.8864148259162903, "num_tokens": 306548121.0, "step": 8076 }, { "epoch": 1.02747742017555, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.944528818130493, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8788304328918457, "num_tokens": 306596776.0, "step": 8077 }, { "epoch": 1.0276046304541406, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.1679348945617676, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8639938831329346, "num_tokens": 306635497.0, "step": 8078 }, { "epoch": 1.0277318407327312, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00018405914306640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.10467004776001, "learning_rate": 1e-05, "loss": 0.5368, "mean_token_accuracy": 0.8572832345962524, "num_tokens": 306671933.0, "step": 8079 }, { "epoch": 1.0278590510113217, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.39560604095459, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8730300664901733, "num_tokens": 306709313.0, "step": 8080 }, { "epoch": 1.0279862612899122, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.990570068359375, "learning_rate": 1e-05, "loss": 0.501, "mean_token_accuracy": 0.8539173603057861, "num_tokens": 306748181.0, "step": 8081 }, { "epoch": 1.0281134715685027, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5143604278564453, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8821651935577393, "num_tokens": 306786353.0, "step": 8082 }, { "epoch": 1.0282406818470933, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8146445751190186, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8703007698059082, "num_tokens": 306821660.0, "step": 8083 }, { "epoch": 1.0283678921256838, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5090363025665283, "learning_rate": 1e-05, "loss": 0.3845, "mean_token_accuracy": 0.8924484252929688, "num_tokens": 306858479.0, "step": 8084 }, { "epoch": 1.0284951024042743, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6365301609039307, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8699159622192383, "num_tokens": 306890483.0, "step": 8085 }, { "epoch": 1.0286223126828649, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4589757919311523, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8734396696090698, "num_tokens": 306931865.0, "step": 8086 }, { "epoch": 1.0287495229614554, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5005292892456055, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8666408061981201, "num_tokens": 306966907.0, "step": 8087 }, { "epoch": 1.0288767332400457, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.390544891357422, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8770960569381714, "num_tokens": 307006032.0, "step": 8088 }, { "epoch": 1.0290039435186362, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.42035174369812, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8649533987045288, "num_tokens": 307045621.0, "step": 8089 }, { "epoch": 1.0291311537972267, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.339022397994995, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8600325584411621, "num_tokens": 307078768.0, "step": 8090 }, { "epoch": 1.0292583640758173, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3833651542663574, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8743945360183716, "num_tokens": 307120089.0, "step": 8091 }, { "epoch": 1.0293855743544078, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.269138813018799, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8701651096343994, "num_tokens": 307157478.0, "step": 8092 }, { "epoch": 1.0295127846329983, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3654427528381348, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8712263107299805, "num_tokens": 307199842.0, "step": 8093 }, { "epoch": 1.0296399949115889, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.22196102142334, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8707143664360046, "num_tokens": 307240609.0, "step": 8094 }, { "epoch": 1.0297672051901794, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.397676706314087, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8609675168991089, "num_tokens": 307279112.0, "step": 8095 }, { "epoch": 1.02989441546877, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1800923347473145, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8777029514312744, "num_tokens": 307313718.0, "step": 8096 }, { "epoch": 1.0300216257473604, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4458093643188477, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8702898621559143, "num_tokens": 307353660.0, "step": 8097 }, { "epoch": 1.030148836025951, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1086907386779785, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8627898097038269, "num_tokens": 307393974.0, "step": 8098 }, { "epoch": 1.0302760463045415, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5407280921936035, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.881051778793335, "num_tokens": 307434877.0, "step": 8099 }, { "epoch": 1.0304032565831318, "ewc_loss": 0.061279296875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0455069541931152, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.8810422420501709, "num_tokens": 307473167.0, "step": 8100 }, { "epoch": 1.0305304668617223, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7068865299224854, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8817951679229736, "num_tokens": 307513781.0, "step": 8101 }, { "epoch": 1.0306576771403129, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.002298355102539, "learning_rate": 1e-05, "loss": 0.3799, "mean_token_accuracy": 0.8867150545120239, "num_tokens": 307549524.0, "step": 8102 }, { "epoch": 1.0307848874189034, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.939058542251587, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8698749542236328, "num_tokens": 307585196.0, "step": 8103 }, { "epoch": 1.030912097697494, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.103717803955078, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8590131998062134, "num_tokens": 307623002.0, "step": 8104 }, { "epoch": 1.0310393079760845, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.103217124938965, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8737837076187134, "num_tokens": 307667736.0, "step": 8105 }, { "epoch": 1.031166518254675, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3241302967071533, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8809075951576233, "num_tokens": 307706031.0, "step": 8106 }, { "epoch": 1.0312937285332655, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.106417655944824, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.8591556549072266, "num_tokens": 307746333.0, "step": 8107 }, { "epoch": 1.031420938811856, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.548492908477783, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8572255373001099, "num_tokens": 307784479.0, "step": 8108 }, { "epoch": 1.0315481490904466, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.918717384338379, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8701401948928833, "num_tokens": 307818245.0, "step": 8109 }, { "epoch": 1.031675359369037, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.5638277530670166, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8765503168106079, "num_tokens": 307858730.0, "step": 8110 }, { "epoch": 1.0318025696476276, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.707395315170288, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.883821427822113, "num_tokens": 307897067.0, "step": 8111 }, { "epoch": 1.0319297799262181, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5064988136291504, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8689035177230835, "num_tokens": 307935896.0, "step": 8112 }, { "epoch": 1.0320569902048085, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.560621500015259, "learning_rate": 1e-05, "loss": 0.509, "mean_token_accuracy": 0.851064920425415, "num_tokens": 307970930.0, "step": 8113 }, { "epoch": 1.032184200483399, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4213156700134277, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8597588539123535, "num_tokens": 308015857.0, "step": 8114 }, { "epoch": 1.0323114107619895, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.453321695327759, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8719431757926941, "num_tokens": 308055432.0, "step": 8115 }, { "epoch": 1.03243862104058, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001850128173828125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3413939476013184, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8600068092346191, "num_tokens": 308094922.0, "step": 8116 }, { "epoch": 1.0325658313191706, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4194955825805664, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8728649616241455, "num_tokens": 308132941.0, "step": 8117 }, { "epoch": 1.032693041597761, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2717411518096924, "learning_rate": 1e-05, "loss": 0.4946, "mean_token_accuracy": 0.854505717754364, "num_tokens": 308176749.0, "step": 8118 }, { "epoch": 1.0328202518763516, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.41230845451355, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8607767820358276, "num_tokens": 308215845.0, "step": 8119 }, { "epoch": 1.0329474621549422, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.201143980026245, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.87798011302948, "num_tokens": 308254220.0, "step": 8120 }, { "epoch": 1.0330746724335327, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.452125310897827, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8604004383087158, "num_tokens": 308290059.0, "step": 8121 }, { "epoch": 1.0332018827121232, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.189906597137451, "learning_rate": 1e-05, "loss": 0.5149, "mean_token_accuracy": 0.8454248905181885, "num_tokens": 308320478.0, "step": 8122 }, { "epoch": 1.0333290929907137, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4815328121185303, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8747529983520508, "num_tokens": 308356094.0, "step": 8123 }, { "epoch": 1.0334563032693043, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.095280408859253, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8732746839523315, "num_tokens": 308393120.0, "step": 8124 }, { "epoch": 1.0335835135478946, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.605064868927002, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8813884258270264, "num_tokens": 308424153.0, "step": 8125 }, { "epoch": 1.033710723826485, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.027540445327759, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.874764084815979, "num_tokens": 308462884.0, "step": 8126 }, { "epoch": 1.0338379341050756, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7897450923919678, "learning_rate": 1e-05, "loss": 0.3859, "mean_token_accuracy": 0.8907129764556885, "num_tokens": 308501185.0, "step": 8127 }, { "epoch": 1.0339651443836662, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.04347562789917, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8601630926132202, "num_tokens": 308538770.0, "step": 8128 }, { "epoch": 1.0340923546622567, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.001232624053955, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8729270100593567, "num_tokens": 308575224.0, "step": 8129 }, { "epoch": 1.0342195649408472, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.186217784881592, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8824955224990845, "num_tokens": 308611859.0, "step": 8130 }, { "epoch": 1.0343467752194377, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.114245414733887, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8552437424659729, "num_tokens": 308648212.0, "step": 8131 }, { "epoch": 1.0344739854980283, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4279236793518066, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8685087561607361, "num_tokens": 308689830.0, "step": 8132 }, { "epoch": 1.0346011957766188, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.012257099151611, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8568873405456543, "num_tokens": 308726238.0, "step": 8133 }, { "epoch": 1.0347284060552093, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.581617832183838, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8627849221229553, "num_tokens": 308755090.0, "step": 8134 }, { "epoch": 1.0348556163337999, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7803657054901123, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.856408953666687, "num_tokens": 308794025.0, "step": 8135 }, { "epoch": 1.0349828266123904, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.539191961288452, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8696106672286987, "num_tokens": 308831521.0, "step": 8136 }, { "epoch": 1.0351100368909807, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.585292100906372, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8753513693809509, "num_tokens": 308868113.0, "step": 8137 }, { "epoch": 1.0352372471695712, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4682424068450928, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8697415590286255, "num_tokens": 308905070.0, "step": 8138 }, { "epoch": 1.0353644574481617, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4581081867218018, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8621506094932556, "num_tokens": 308942703.0, "step": 8139 }, { "epoch": 1.0354916677267523, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.402414083480835, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8723137378692627, "num_tokens": 308981328.0, "step": 8140 }, { "epoch": 1.0356188780053428, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3707592487335205, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8834054470062256, "num_tokens": 309017751.0, "step": 8141 }, { "epoch": 1.0357460882839333, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3629560470581055, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8713729977607727, "num_tokens": 309052216.0, "step": 8142 }, { "epoch": 1.0358732985625239, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3171932697296143, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8690229654312134, "num_tokens": 309091323.0, "step": 8143 }, { "epoch": 1.0360005088411144, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3250648975372314, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8705016374588013, "num_tokens": 309125324.0, "step": 8144 }, { "epoch": 1.036127719119705, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2871227264404297, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8701704740524292, "num_tokens": 309166009.0, "step": 8145 }, { "epoch": 1.0362549293982954, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.298342704772949, "learning_rate": 1e-05, "loss": 0.4054, "mean_token_accuracy": 0.8816956877708435, "num_tokens": 309205822.0, "step": 8146 }, { "epoch": 1.036382139676886, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2640745639801025, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8726866245269775, "num_tokens": 309243463.0, "step": 8147 }, { "epoch": 1.0365093499554765, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3689823150634766, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8657340407371521, "num_tokens": 309276904.0, "step": 8148 }, { "epoch": 1.0366365602340668, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.191788673400879, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8698868155479431, "num_tokens": 309314292.0, "step": 8149 }, { "epoch": 1.0367637705126573, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3809397220611572, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8626700639724731, "num_tokens": 309349231.0, "step": 8150 }, { "epoch": 1.0368909807912479, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018596649169921875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.134387493133545, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8467150926589966, "num_tokens": 309387901.0, "step": 8151 }, { "epoch": 1.0370181910698384, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4447576999664307, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8735286593437195, "num_tokens": 309427167.0, "step": 8152 }, { "epoch": 1.037145401348429, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0692527294158936, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8710719347000122, "num_tokens": 309464121.0, "step": 8153 }, { "epoch": 1.0372726116270194, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5581202507019043, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8719601631164551, "num_tokens": 309506759.0, "step": 8154 }, { "epoch": 1.03739982190561, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 2.988687753677368, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8797582387924194, "num_tokens": 309546059.0, "step": 8155 }, { "epoch": 1.0375270321842005, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7749149799346924, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8645947575569153, "num_tokens": 309585076.0, "step": 8156 }, { "epoch": 1.037654242462791, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 2.9533448219299316, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8788954019546509, "num_tokens": 309623354.0, "step": 8157 }, { "epoch": 1.0377814527413816, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.073139190673828, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8559309244155884, "num_tokens": 309668420.0, "step": 8158 }, { "epoch": 1.037908663019972, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.1055238246917725, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8729965686798096, "num_tokens": 309711417.0, "step": 8159 }, { "epoch": 1.0380358732985626, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.325194835662842, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8741751313209534, "num_tokens": 309749938.0, "step": 8160 }, { "epoch": 1.0381630835771531, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.517155408859253, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8747032880783081, "num_tokens": 309783328.0, "step": 8161 }, { "epoch": 1.0382902938557435, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.220718860626221, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8774371147155762, "num_tokens": 309821347.0, "step": 8162 }, { "epoch": 1.038417504134334, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.7266149520874023, "learning_rate": 1e-05, "loss": 0.5095, "mean_token_accuracy": 0.8534907102584839, "num_tokens": 309858430.0, "step": 8163 }, { "epoch": 1.0385447144129245, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.968701124191284, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8698424100875854, "num_tokens": 309896096.0, "step": 8164 }, { "epoch": 1.038671924691515, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.674729347229004, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8686158657073975, "num_tokens": 309929069.0, "step": 8165 }, { "epoch": 1.0387991349701056, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.782071590423584, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.8575394749641418, "num_tokens": 309969706.0, "step": 8166 }, { "epoch": 1.038926345248696, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.55053448677063, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8651794195175171, "num_tokens": 310003990.0, "step": 8167 }, { "epoch": 1.0390535555272866, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6571953296661377, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.849916934967041, "num_tokens": 310036253.0, "step": 8168 }, { "epoch": 1.0391807658058771, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4152567386627197, "learning_rate": 1e-05, "loss": 0.39, "mean_token_accuracy": 0.8848094940185547, "num_tokens": 310073189.0, "step": 8169 }, { "epoch": 1.0393079760844677, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.592413902282715, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8641847372055054, "num_tokens": 310110432.0, "step": 8170 }, { "epoch": 1.0394351863630582, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.307530403137207, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8628560304641724, "num_tokens": 310154830.0, "step": 8171 }, { "epoch": 1.0395623966416487, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6076602935791016, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8555637001991272, "num_tokens": 310187270.0, "step": 8172 }, { "epoch": 1.0396896069202393, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2123231887817383, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8722638487815857, "num_tokens": 310223736.0, "step": 8173 }, { "epoch": 1.0398168171988296, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6928324699401855, "learning_rate": 1e-05, "loss": 0.5213, "mean_token_accuracy": 0.8518866300582886, "num_tokens": 310259262.0, "step": 8174 }, { "epoch": 1.03994402747742, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.131457805633545, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8697388172149658, "num_tokens": 310300438.0, "step": 8175 }, { "epoch": 1.0400712377560106, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.844421148300171, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8791051506996155, "num_tokens": 310336630.0, "step": 8176 }, { "epoch": 1.0401984480346012, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1751537322998047, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8646513223648071, "num_tokens": 310368755.0, "step": 8177 }, { "epoch": 1.0403256583131917, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.968811273574829, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8811926245689392, "num_tokens": 310403520.0, "step": 8178 }, { "epoch": 1.0404528685917822, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.306431531906128, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8673586249351501, "num_tokens": 310433805.0, "step": 8179 }, { "epoch": 1.0405800788703727, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.978355646133423, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8652030229568481, "num_tokens": 310469801.0, "step": 8180 }, { "epoch": 1.0407072891489633, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.4494879245758057, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.8846556544303894, "num_tokens": 310504972.0, "step": 8181 }, { "epoch": 1.0408344994275538, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8377878665924072, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8647515773773193, "num_tokens": 310537851.0, "step": 8182 }, { "epoch": 1.0409617097061443, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5168697834014893, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8697140216827393, "num_tokens": 310572647.0, "step": 8183 }, { "epoch": 1.0410889199847349, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.621436834335327, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8650714755058289, "num_tokens": 310612239.0, "step": 8184 }, { "epoch": 1.0412161302633254, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.493285655975342, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8632521033287048, "num_tokens": 310650442.0, "step": 8185 }, { "epoch": 1.0413433405419157, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.434560775756836, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8662822246551514, "num_tokens": 310687687.0, "step": 8186 }, { "epoch": 1.0414705508205062, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4817869663238525, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8606241941452026, "num_tokens": 310725165.0, "step": 8187 }, { "epoch": 1.0415977610990967, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.289515256881714, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.883995771408081, "num_tokens": 310762532.0, "step": 8188 }, { "epoch": 1.0417249713776873, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4740874767303467, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.860601544380188, "num_tokens": 310800903.0, "step": 8189 }, { "epoch": 1.0418521816562778, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.184739351272583, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8500848412513733, "num_tokens": 310841391.0, "step": 8190 }, { "epoch": 1.0419793919348683, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.516171932220459, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8617444038391113, "num_tokens": 310879027.0, "step": 8191 }, { "epoch": 1.0421066022134589, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.115647315979004, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8733097314834595, "num_tokens": 310918448.0, "step": 8192 }, { "epoch": 1.0422338124920494, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.599735975265503, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8646807670593262, "num_tokens": 310955641.0, "step": 8193 }, { "epoch": 1.04236102277064, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.07257342338562, "learning_rate": 1e-05, "loss": 0.516, "mean_token_accuracy": 0.8457438945770264, "num_tokens": 310996487.0, "step": 8194 }, { "epoch": 1.0424882330492304, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.728691577911377, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8611017465591431, "num_tokens": 311038208.0, "step": 8195 }, { "epoch": 1.042615443327821, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0478975772857666, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8701367378234863, "num_tokens": 311073720.0, "step": 8196 }, { "epoch": 1.0427426536064115, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9063234329223633, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8827542066574097, "num_tokens": 311115488.0, "step": 8197 }, { "epoch": 1.0428698638850018, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.115147113800049, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.869382917881012, "num_tokens": 311159217.0, "step": 8198 }, { "epoch": 1.0429970741635923, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.065396785736084, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8580626249313354, "num_tokens": 311196058.0, "step": 8199 }, { "epoch": 1.0431242844421829, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000186920166015625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3354005813598633, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.875042736530304, "num_tokens": 311231024.0, "step": 8200 }, { "epoch": 1.0432514947207734, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.042417049407959, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8708075284957886, "num_tokens": 311266883.0, "step": 8201 }, { "epoch": 1.043378704999364, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5107760429382324, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8519659042358398, "num_tokens": 311302594.0, "step": 8202 }, { "epoch": 1.0435059152779544, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.860807180404663, "learning_rate": 1e-05, "loss": 0.3926, "mean_token_accuracy": 0.8888150453567505, "num_tokens": 311334384.0, "step": 8203 }, { "epoch": 1.043633125556545, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.517285108566284, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8725811243057251, "num_tokens": 311381350.0, "step": 8204 }, { "epoch": 1.0437603358351355, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.671846628189087, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.874549388885498, "num_tokens": 311418107.0, "step": 8205 }, { "epoch": 1.043887546113726, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4791595935821533, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8701156973838806, "num_tokens": 311457881.0, "step": 8206 }, { "epoch": 1.0440147563923166, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.526697874069214, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8631722331047058, "num_tokens": 311492821.0, "step": 8207 }, { "epoch": 1.044141966670907, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.428436517715454, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8680674433708191, "num_tokens": 311527981.0, "step": 8208 }, { "epoch": 1.0442691769494976, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.405925989151001, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8598670363426208, "num_tokens": 311568853.0, "step": 8209 }, { "epoch": 1.0443963872280881, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3600966930389404, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8807432651519775, "num_tokens": 311613321.0, "step": 8210 }, { "epoch": 1.0445235975066784, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.348665237426758, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8641698360443115, "num_tokens": 311656849.0, "step": 8211 }, { "epoch": 1.044650807785269, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.335658311843872, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8569633960723877, "num_tokens": 311692231.0, "step": 8212 }, { "epoch": 1.0447780180638595, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3088221549987793, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8801435828208923, "num_tokens": 311732793.0, "step": 8213 }, { "epoch": 1.04490522834245, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2975316047668457, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8653504252433777, "num_tokens": 311773320.0, "step": 8214 }, { "epoch": 1.0450324386210406, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2865471839904785, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8660837411880493, "num_tokens": 311822187.0, "step": 8215 }, { "epoch": 1.045159648899631, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.280147075653076, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8683434724807739, "num_tokens": 311860272.0, "step": 8216 }, { "epoch": 1.0452868591782216, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2942113876342773, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.869150698184967, "num_tokens": 311894479.0, "step": 8217 }, { "epoch": 1.0454140694568121, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.257155418395996, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8803279399871826, "num_tokens": 311929401.0, "step": 8218 }, { "epoch": 1.0455412797354027, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.289332389831543, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8618260025978088, "num_tokens": 311969401.0, "step": 8219 }, { "epoch": 1.0456684900139932, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2493252754211426, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8578177690505981, "num_tokens": 312004295.0, "step": 8220 }, { "epoch": 1.0457957002925837, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.297701358795166, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8672144412994385, "num_tokens": 312038487.0, "step": 8221 }, { "epoch": 1.0459229105711743, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.212378740310669, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8696966171264648, "num_tokens": 312076011.0, "step": 8222 }, { "epoch": 1.0460501208497646, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3180835247039795, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8614273071289062, "num_tokens": 312114489.0, "step": 8223 }, { "epoch": 1.046177331128355, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.175753116607666, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8647571802139282, "num_tokens": 312152388.0, "step": 8224 }, { "epoch": 1.0463045414069456, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.360520362854004, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.87464839220047, "num_tokens": 312190269.0, "step": 8225 }, { "epoch": 1.0464317516855361, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.126471996307373, "learning_rate": 1e-05, "loss": 0.4167, "mean_token_accuracy": 0.878358781337738, "num_tokens": 312229909.0, "step": 8226 }, { "epoch": 1.0465589619641267, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4572927951812744, "learning_rate": 1e-05, "loss": 0.5141, "mean_token_accuracy": 0.8531295657157898, "num_tokens": 312267527.0, "step": 8227 }, { "epoch": 1.0466861722427172, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0445139408111572, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8714406490325928, "num_tokens": 312310632.0, "step": 8228 }, { "epoch": 1.0468133825213077, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6129889488220215, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8588624000549316, "num_tokens": 312351570.0, "step": 8229 }, { "epoch": 1.0469405927998983, "ewc_loss": 0.06103515625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.1961669921875e-05, "grad_norm": 2.9597740173339844, "learning_rate": 1e-05, "loss": 0.3788, "mean_token_accuracy": 0.8898736238479614, "num_tokens": 312382765.0, "step": 8230 }, { "epoch": 1.0470678030784888, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.884756565093994, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8773624300956726, "num_tokens": 312415517.0, "step": 8231 }, { "epoch": 1.0471950133570793, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 2.957913875579834, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.875322699546814, "num_tokens": 312450574.0, "step": 8232 }, { "epoch": 1.0473222236356698, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2450737953186035, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.8586564660072327, "num_tokens": 312483054.0, "step": 8233 }, { "epoch": 1.0474494339142604, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.2220442295074463, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.869160532951355, "num_tokens": 312527825.0, "step": 8234 }, { "epoch": 1.0475766441928507, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.507710933685303, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.862991452217102, "num_tokens": 312566971.0, "step": 8235 }, { "epoch": 1.0477038544714412, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.7589919567108154, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8694871664047241, "num_tokens": 312603527.0, "step": 8236 }, { "epoch": 1.0478310647500317, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.197697162628174, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8688706755638123, "num_tokens": 312634281.0, "step": 8237 }, { "epoch": 1.0479582750286223, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.837583303451538, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.863856852054596, "num_tokens": 312668036.0, "step": 8238 }, { "epoch": 1.0480854853072128, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9362905025482178, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8606442213058472, "num_tokens": 312707422.0, "step": 8239 }, { "epoch": 1.0482126955858033, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6904053688049316, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8644485473632812, "num_tokens": 312740715.0, "step": 8240 }, { "epoch": 1.0483399058643939, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.771878242492676, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8713919520378113, "num_tokens": 312773196.0, "step": 8241 }, { "epoch": 1.0484671161429844, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5254523754119873, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.860962986946106, "num_tokens": 312817502.0, "step": 8242 }, { "epoch": 1.048594326421575, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.8099522590637207, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8646251559257507, "num_tokens": 312852810.0, "step": 8243 }, { "epoch": 1.0487215367001654, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3888938426971436, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8781754374504089, "num_tokens": 312889069.0, "step": 8244 }, { "epoch": 1.048848746978756, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7021098136901855, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8647753000259399, "num_tokens": 312929032.0, "step": 8245 }, { "epoch": 1.0489759572573465, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2765746116638184, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8707879781723022, "num_tokens": 312963792.0, "step": 8246 }, { "epoch": 1.0491031675359368, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7709317207336426, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8858109712600708, "num_tokens": 312999532.0, "step": 8247 }, { "epoch": 1.0492303778145273, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2112009525299072, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8688524961471558, "num_tokens": 313040807.0, "step": 8248 }, { "epoch": 1.0493575880931179, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.9004645347595215, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8785916566848755, "num_tokens": 313074895.0, "step": 8249 }, { "epoch": 1.0494847983717084, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2436251640319824, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8724460601806641, "num_tokens": 313115299.0, "step": 8250 }, { "epoch": 1.049612008650299, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.007547378540039, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8684536814689636, "num_tokens": 313157578.0, "step": 8251 }, { "epoch": 1.0497392189288894, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.407510995864868, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8688461780548096, "num_tokens": 313202219.0, "step": 8252 }, { "epoch": 1.04986642920748, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9436285495758057, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8739244937896729, "num_tokens": 313242345.0, "step": 8253 }, { "epoch": 1.0499936394860705, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5098090171813965, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8758594989776611, "num_tokens": 313277052.0, "step": 8254 }, { "epoch": 1.050120849764661, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.770803928375244, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8679440021514893, "num_tokens": 313311291.0, "step": 8255 }, { "epoch": 1.0502480600432516, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5198497772216797, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8729903697967529, "num_tokens": 313349141.0, "step": 8256 }, { "epoch": 1.050375270321842, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001888275146484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.564541816711426, "learning_rate": 1e-05, "loss": 0.4982, "mean_token_accuracy": 0.8566111326217651, "num_tokens": 313387575.0, "step": 8257 }, { "epoch": 1.0505024806004326, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.494711399078369, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8623441457748413, "num_tokens": 313424087.0, "step": 8258 }, { "epoch": 1.0506296908790231, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3926610946655273, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8734079003334045, "num_tokens": 313464517.0, "step": 8259 }, { "epoch": 1.0507569011576134, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4537553787231445, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.859731912612915, "num_tokens": 313504146.0, "step": 8260 }, { "epoch": 1.050884111436204, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2891788482666016, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8655481338500977, "num_tokens": 313541842.0, "step": 8261 }, { "epoch": 1.0510113217147945, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.444890260696411, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8651567101478577, "num_tokens": 313580695.0, "step": 8262 }, { "epoch": 1.051138531993385, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.211944580078125, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8583261370658875, "num_tokens": 313619494.0, "step": 8263 }, { "epoch": 1.0512657422719756, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4692389965057373, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8775989413261414, "num_tokens": 313653234.0, "step": 8264 }, { "epoch": 1.051392952550566, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1342878341674805, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.867280900478363, "num_tokens": 313694395.0, "step": 8265 }, { "epoch": 1.0515201628291566, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5411510467529297, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8557737469673157, "num_tokens": 313733370.0, "step": 8266 }, { "epoch": 1.0516473731077471, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.072613000869751, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8555418848991394, "num_tokens": 313770611.0, "step": 8267 }, { "epoch": 1.0517745833863377, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6730546951293945, "learning_rate": 1e-05, "loss": 0.3856, "mean_token_accuracy": 0.8887014985084534, "num_tokens": 313807262.0, "step": 8268 }, { "epoch": 1.0519017936649282, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.026050090789795, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8637791872024536, "num_tokens": 313844294.0, "step": 8269 }, { "epoch": 1.0520290039435187, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.867007255554199, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8803942799568176, "num_tokens": 313878701.0, "step": 8270 }, { "epoch": 1.0521562142221093, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.0618560314178467, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8796535730361938, "num_tokens": 313917780.0, "step": 8271 }, { "epoch": 1.0522834245006996, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0801777839660645, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8823466300964355, "num_tokens": 313948611.0, "step": 8272 }, { "epoch": 1.05241063477929, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.28255295753479, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8747211694717407, "num_tokens": 313981999.0, "step": 8273 }, { "epoch": 1.0525378450578806, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.117706775665283, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8762221932411194, "num_tokens": 314021199.0, "step": 8274 }, { "epoch": 1.0526650553364711, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5014262199401855, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8816638588905334, "num_tokens": 314058187.0, "step": 8275 }, { "epoch": 1.0527922656150617, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9624435901641846, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.869554877281189, "num_tokens": 314095612.0, "step": 8276 }, { "epoch": 1.0529194758936522, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.59209942817688, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8632745146751404, "num_tokens": 314124895.0, "step": 8277 }, { "epoch": 1.0530466861722427, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.742622137069702, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8793524503707886, "num_tokens": 314167787.0, "step": 8278 }, { "epoch": 1.0531738964508333, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.536388874053955, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8721140027046204, "num_tokens": 314203389.0, "step": 8279 }, { "epoch": 1.0533011067294238, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.558197259902954, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8806043863296509, "num_tokens": 314245161.0, "step": 8280 }, { "epoch": 1.0534283170080143, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.460972785949707, "learning_rate": 1e-05, "loss": 0.5157, "mean_token_accuracy": 0.8471419811248779, "num_tokens": 314288097.0, "step": 8281 }, { "epoch": 1.0535555272866048, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.443530321121216, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8666143417358398, "num_tokens": 314324066.0, "step": 8282 }, { "epoch": 1.0536827375651954, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3924567699432373, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8795838356018066, "num_tokens": 314358430.0, "step": 8283 }, { "epoch": 1.0538099478437857, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.384291648864746, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8668643236160278, "num_tokens": 314394199.0, "step": 8284 }, { "epoch": 1.0539371581223762, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3290915489196777, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8721283674240112, "num_tokens": 314432851.0, "step": 8285 }, { "epoch": 1.0540643684009667, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3505685329437256, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8691107034683228, "num_tokens": 314473843.0, "step": 8286 }, { "epoch": 1.0541915786795573, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.26431941986084, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8849493861198425, "num_tokens": 314518782.0, "step": 8287 }, { "epoch": 1.0543187889581478, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.353977680206299, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8672409653663635, "num_tokens": 314561140.0, "step": 8288 }, { "epoch": 1.0544459992367383, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2282371520996094, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8680959939956665, "num_tokens": 314602605.0, "step": 8289 }, { "epoch": 1.0545732095153288, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3641912937164307, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.854421854019165, "num_tokens": 314647719.0, "step": 8290 }, { "epoch": 1.0547004197939194, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1880831718444824, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8560981750488281, "num_tokens": 314690016.0, "step": 8291 }, { "epoch": 1.05482763007251, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.4171764850616455, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8726891279220581, "num_tokens": 314730432.0, "step": 8292 }, { "epoch": 1.0549548403511004, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1234171390533447, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8774582743644714, "num_tokens": 314764957.0, "step": 8293 }, { "epoch": 1.055082050629691, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.503861427307129, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8737345933914185, "num_tokens": 314803787.0, "step": 8294 }, { "epoch": 1.0552092609082815, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0443716049194336, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8759331703186035, "num_tokens": 314841732.0, "step": 8295 }, { "epoch": 1.0553364711868718, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6607000827789307, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8747282028198242, "num_tokens": 314878212.0, "step": 8296 }, { "epoch": 1.0554636814654623, "ewc_loss": 0.0615234375, "ewc_loss_diag": 0.00018978118896484375, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 2.9915754795074463, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.864803671836853, "num_tokens": 314917130.0, "step": 8297 }, { "epoch": 1.0555908917440529, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8944146633148193, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8772889971733093, "num_tokens": 314961804.0, "step": 8298 }, { "epoch": 1.0557181020226434, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.0191946029663086, "learning_rate": 1e-05, "loss": 0.3872, "mean_token_accuracy": 0.8866068124771118, "num_tokens": 314997157.0, "step": 8299 }, { "epoch": 1.055845312301234, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.176551818847656, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8507940769195557, "num_tokens": 315038370.0, "step": 8300 }, { "epoch": 1.0559725225798244, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.286295175552368, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8826414346694946, "num_tokens": 315081716.0, "step": 8301 }, { "epoch": 1.056099732858415, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.252865314483643, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8498711585998535, "num_tokens": 315124528.0, "step": 8302 }, { "epoch": 1.0562269431370055, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6111819744110107, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8838850259780884, "num_tokens": 315163893.0, "step": 8303 }, { "epoch": 1.056354153415596, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.06532621383667, "learning_rate": 1e-05, "loss": 0.5299, "mean_token_accuracy": 0.8482732772827148, "num_tokens": 315199462.0, "step": 8304 }, { "epoch": 1.0564813636941865, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.674398422241211, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8695164918899536, "num_tokens": 315245075.0, "step": 8305 }, { "epoch": 1.056608573972777, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.830587148666382, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8730759024620056, "num_tokens": 315289291.0, "step": 8306 }, { "epoch": 1.0567357842513676, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.589170455932617, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8773082494735718, "num_tokens": 315327750.0, "step": 8307 }, { "epoch": 1.0568629945299581, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.666464328765869, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8651955127716064, "num_tokens": 315366863.0, "step": 8308 }, { "epoch": 1.0569902048085484, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.476480484008789, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8837249279022217, "num_tokens": 315406935.0, "step": 8309 }, { "epoch": 1.057117415087139, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5667366981506348, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8637669086456299, "num_tokens": 315446483.0, "step": 8310 }, { "epoch": 1.0572446253657295, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.377288818359375, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8657290935516357, "num_tokens": 315482862.0, "step": 8311 }, { "epoch": 1.05737183564432, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5188515186309814, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8735352158546448, "num_tokens": 315522649.0, "step": 8312 }, { "epoch": 1.0574990459229106, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2708323001861572, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8636274337768555, "num_tokens": 315562642.0, "step": 8313 }, { "epoch": 1.057626256201501, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.534130334854126, "learning_rate": 1e-05, "loss": 0.3737, "mean_token_accuracy": 0.8943660259246826, "num_tokens": 315594056.0, "step": 8314 }, { "epoch": 1.0577534664800916, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1819255352020264, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8651087284088135, "num_tokens": 315632474.0, "step": 8315 }, { "epoch": 1.0578806767586821, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6295461654663086, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8711618185043335, "num_tokens": 315662036.0, "step": 8316 }, { "epoch": 1.0580078870372727, "ewc_loss": 0.0625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.102982521057129, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8734710216522217, "num_tokens": 315701974.0, "step": 8317 }, { "epoch": 1.0581350973158632, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.767313003540039, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8675304651260376, "num_tokens": 315745876.0, "step": 8318 }, { "epoch": 1.0582623075944537, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0836739540100098, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8687391877174377, "num_tokens": 315787386.0, "step": 8319 }, { "epoch": 1.058389517873044, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9687321186065674, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8788518905639648, "num_tokens": 315824047.0, "step": 8320 }, { "epoch": 1.0585167281516346, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.228315830230713, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8525977730751038, "num_tokens": 315863890.0, "step": 8321 }, { "epoch": 1.058643938430225, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.037446975708008, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8631409406661987, "num_tokens": 315902855.0, "step": 8322 }, { "epoch": 1.0587711487088156, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4265730381011963, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8725119233131409, "num_tokens": 315942112.0, "step": 8323 }, { "epoch": 1.0588983589874061, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9210216999053955, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8729380369186401, "num_tokens": 315981256.0, "step": 8324 }, { "epoch": 1.0590255692659967, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5315887928009033, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.864276111125946, "num_tokens": 316021192.0, "step": 8325 }, { "epoch": 1.0591527795445872, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.71159291267395, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.867215096950531, "num_tokens": 316056261.0, "step": 8326 }, { "epoch": 1.0592799898231777, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5192649364471436, "learning_rate": 1e-05, "loss": 0.3934, "mean_token_accuracy": 0.8887178301811218, "num_tokens": 316093390.0, "step": 8327 }, { "epoch": 1.0594072001017683, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5071802139282227, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8560313582420349, "num_tokens": 316133771.0, "step": 8328 }, { "epoch": 1.0595344103803588, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.492553234100342, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8693336248397827, "num_tokens": 316172123.0, "step": 8329 }, { "epoch": 1.0596616206589493, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.336149215698242, "learning_rate": 1e-05, "loss": 0.4033, "mean_token_accuracy": 0.882554292678833, "num_tokens": 316211866.0, "step": 8330 }, { "epoch": 1.0597888309375398, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.487154245376587, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8749172687530518, "num_tokens": 316248838.0, "step": 8331 }, { "epoch": 1.0599160412161304, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2277603149414062, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8730698823928833, "num_tokens": 316286551.0, "step": 8332 }, { "epoch": 1.0600432514947207, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.524204730987549, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8680863380432129, "num_tokens": 316321278.0, "step": 8333 }, { "epoch": 1.0601704617733112, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.1530869007110596, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8695361018180847, "num_tokens": 316356282.0, "step": 8334 }, { "epoch": 1.0602976720519017, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.575855255126953, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8680727481842041, "num_tokens": 316398858.0, "step": 8335 }, { "epoch": 1.0604248823304923, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.089629650115967, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8693751096725464, "num_tokens": 316432883.0, "step": 8336 }, { "epoch": 1.0605520926090828, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6953659057617188, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8842350840568542, "num_tokens": 316469305.0, "step": 8337 }, { "epoch": 1.0606793028876733, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0462329387664795, "learning_rate": 1e-05, "loss": 0.383, "mean_token_accuracy": 0.8884819746017456, "num_tokens": 316509072.0, "step": 8338 }, { "epoch": 1.0608065131662638, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8696882724761963, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8662325739860535, "num_tokens": 316549391.0, "step": 8339 }, { "epoch": 1.0609337234448544, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.107156991958618, "learning_rate": 1e-05, "loss": 0.3879, "mean_token_accuracy": 0.8871481418609619, "num_tokens": 316583696.0, "step": 8340 }, { "epoch": 1.061060933723445, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.035408973693848, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8627341389656067, "num_tokens": 316618421.0, "step": 8341 }, { "epoch": 1.0611881440020354, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2781178951263428, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8695551156997681, "num_tokens": 316660284.0, "step": 8342 }, { "epoch": 1.061315354280626, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.058687210083008, "learning_rate": 1e-05, "loss": 0.5186, "mean_token_accuracy": 0.8557427525520325, "num_tokens": 316699753.0, "step": 8343 }, { "epoch": 1.0614425645592165, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.4771223068237305, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8692445755004883, "num_tokens": 316735336.0, "step": 8344 }, { "epoch": 1.0615697748378068, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9602339267730713, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8752137422561646, "num_tokens": 316775327.0, "step": 8345 }, { "epoch": 1.0616969851163973, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5511133670806885, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8543798923492432, "num_tokens": 316817027.0, "step": 8346 }, { "epoch": 1.0618241953949878, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7287118434906006, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8722622394561768, "num_tokens": 316850980.0, "step": 8347 }, { "epoch": 1.0619514056735784, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.4953057765960693, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8869773745536804, "num_tokens": 316886533.0, "step": 8348 }, { "epoch": 1.062078615952169, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5521321296691895, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8799042105674744, "num_tokens": 316927447.0, "step": 8349 }, { "epoch": 1.0622058262307594, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.458289384841919, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8608851432800293, "num_tokens": 316965983.0, "step": 8350 }, { "epoch": 1.06233303650935, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00019073486328125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4315712451934814, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8717193603515625, "num_tokens": 316999678.0, "step": 8351 }, { "epoch": 1.0624602467879405, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.406759738922119, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8601226210594177, "num_tokens": 317038650.0, "step": 8352 }, { "epoch": 1.062587457066531, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.358856439590454, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8736491203308105, "num_tokens": 317075507.0, "step": 8353 }, { "epoch": 1.0627146673451215, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3569562435150146, "learning_rate": 1e-05, "loss": 0.3965, "mean_token_accuracy": 0.8838417530059814, "num_tokens": 317109364.0, "step": 8354 }, { "epoch": 1.062841877623712, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3126111030578613, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.875276505947113, "num_tokens": 317149959.0, "step": 8355 }, { "epoch": 1.0629690879023026, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.336355209350586, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8690860867500305, "num_tokens": 317182604.0, "step": 8356 }, { "epoch": 1.0630962981808931, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2758121490478516, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8798147439956665, "num_tokens": 317224190.0, "step": 8357 }, { "epoch": 1.0632235084594834, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.32181453704834, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8678805232048035, "num_tokens": 317261607.0, "step": 8358 }, { "epoch": 1.063350718738074, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2538070678710938, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8574995994567871, "num_tokens": 317300772.0, "step": 8359 }, { "epoch": 1.0634779290166645, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.315648078918457, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8698331117630005, "num_tokens": 317336981.0, "step": 8360 }, { "epoch": 1.063605139295255, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.228539228439331, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8605144023895264, "num_tokens": 317376522.0, "step": 8361 }, { "epoch": 1.0637323495738455, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00019168853759765625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3328747749328613, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8730123043060303, "num_tokens": 317409924.0, "step": 8362 }, { "epoch": 1.063859559852436, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.190866470336914, "learning_rate": 1e-05, "loss": 0.4028, "mean_token_accuracy": 0.8841395378112793, "num_tokens": 317440762.0, "step": 8363 }, { "epoch": 1.0639867701310266, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.37528133392334, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.873837411403656, "num_tokens": 317471606.0, "step": 8364 }, { "epoch": 1.0641139804096171, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1422104835510254, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8585516214370728, "num_tokens": 317508383.0, "step": 8365 }, { "epoch": 1.0642411906882077, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4351983070373535, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.8905640840530396, "num_tokens": 317546690.0, "step": 8366 }, { "epoch": 1.0643684009667982, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0679798126220703, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8829820156097412, "num_tokens": 317584926.0, "step": 8367 }, { "epoch": 1.0644956112453887, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5807158946990967, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8592337369918823, "num_tokens": 317619941.0, "step": 8368 }, { "epoch": 1.064622821523979, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 2.9957239627838135, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8823097348213196, "num_tokens": 317651006.0, "step": 8369 }, { "epoch": 1.0647500318025696, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.803696632385254, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8724799752235413, "num_tokens": 317688051.0, "step": 8370 }, { "epoch": 1.06487724208116, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001926422119140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.944827079772949, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8775088787078857, "num_tokens": 317727447.0, "step": 8371 }, { "epoch": 1.0650044523597506, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.12906551361084, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8799979090690613, "num_tokens": 317771495.0, "step": 8372 }, { "epoch": 1.0651316626383411, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.1315791606903076, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8818782567977905, "num_tokens": 317808166.0, "step": 8373 }, { "epoch": 1.0652588729169317, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.387632369995117, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8757577538490295, "num_tokens": 317850928.0, "step": 8374 }, { "epoch": 1.0653860831955222, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.572531223297119, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8778214454650879, "num_tokens": 317893156.0, "step": 8375 }, { "epoch": 1.0655132934741127, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.262101650238037, "learning_rate": 1e-05, "loss": 0.5408, "mean_token_accuracy": 0.8462602496147156, "num_tokens": 317937146.0, "step": 8376 }, { "epoch": 1.0656405037527032, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.792578935623169, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8801378011703491, "num_tokens": 317977429.0, "step": 8377 }, { "epoch": 1.0657677140312938, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9932403564453125, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8665290474891663, "num_tokens": 318013445.0, "step": 8378 }, { "epoch": 1.0658949243098843, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7189102172851562, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8624753952026367, "num_tokens": 318054082.0, "step": 8379 }, { "epoch": 1.0660221345884748, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.7893242835998535, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8564894795417786, "num_tokens": 318096125.0, "step": 8380 }, { "epoch": 1.0661493448670654, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5791001319885254, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8723123073577881, "num_tokens": 318130859.0, "step": 8381 }, { "epoch": 1.0662765551456557, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6520578861236572, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8756011724472046, "num_tokens": 318171616.0, "step": 8382 }, { "epoch": 1.0664037654242462, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4480059146881104, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8669846057891846, "num_tokens": 318206741.0, "step": 8383 }, { "epoch": 1.0665309757028367, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5838983058929443, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.864827036857605, "num_tokens": 318252513.0, "step": 8384 }, { "epoch": 1.0666581859814273, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.33441162109375, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8615078926086426, "num_tokens": 318286632.0, "step": 8385 }, { "epoch": 1.0667853962600178, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5763113498687744, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8721858859062195, "num_tokens": 318326923.0, "step": 8386 }, { "epoch": 1.0669126065386083, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.22938871383667, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.8508299589157104, "num_tokens": 318364159.0, "step": 8387 }, { "epoch": 1.0670398168171988, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.64574933052063, "learning_rate": 1e-05, "loss": 0.518, "mean_token_accuracy": 0.848635196685791, "num_tokens": 318402715.0, "step": 8388 }, { "epoch": 1.0671670270957894, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.128324270248413, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8661025762557983, "num_tokens": 318445925.0, "step": 8389 }, { "epoch": 1.06729423737438, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8029720783233643, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8637710809707642, "num_tokens": 318481756.0, "step": 8390 }, { "epoch": 1.0674214476529704, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.116661310195923, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8691087365150452, "num_tokens": 318526885.0, "step": 8391 }, { "epoch": 1.067548657931561, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9882965087890625, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8669271469116211, "num_tokens": 318569396.0, "step": 8392 }, { "epoch": 1.0676758682101515, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2405521869659424, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8858348727226257, "num_tokens": 318606404.0, "step": 8393 }, { "epoch": 1.0678030784887418, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0870842933654785, "learning_rate": 1e-05, "loss": 0.4996, "mean_token_accuracy": 0.856890082359314, "num_tokens": 318642457.0, "step": 8394 }, { "epoch": 1.0679302887673323, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.4790568351745605, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8596941232681274, "num_tokens": 318680498.0, "step": 8395 }, { "epoch": 1.0680574990459228, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.93827748298645, "learning_rate": 1e-05, "loss": 0.4794, "mean_token_accuracy": 0.8639616370201111, "num_tokens": 318720493.0, "step": 8396 }, { "epoch": 1.0681847093245134, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.585930824279785, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8614274859428406, "num_tokens": 318755297.0, "step": 8397 }, { "epoch": 1.068311919603104, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.705007791519165, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8756998181343079, "num_tokens": 318788716.0, "step": 8398 }, { "epoch": 1.0684391298816944, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.562041997909546, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8697933554649353, "num_tokens": 318825130.0, "step": 8399 }, { "epoch": 1.068566340160285, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4765446186065674, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8722187280654907, "num_tokens": 318861216.0, "step": 8400 }, { "epoch": 1.0686935504388755, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.529176712036133, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8732812404632568, "num_tokens": 318896410.0, "step": 8401 }, { "epoch": 1.068820760717466, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3114335536956787, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8842474222183228, "num_tokens": 318933834.0, "step": 8402 }, { "epoch": 1.0689479709960565, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.534163236618042, "learning_rate": 1e-05, "loss": 0.4982, "mean_token_accuracy": 0.851988673210144, "num_tokens": 318973752.0, "step": 8403 }, { "epoch": 1.069075181274647, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.190088987350464, "learning_rate": 1e-05, "loss": 0.3879, "mean_token_accuracy": 0.8871661424636841, "num_tokens": 319009376.0, "step": 8404 }, { "epoch": 1.0692023915532376, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5778064727783203, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8726716041564941, "num_tokens": 319050175.0, "step": 8405 }, { "epoch": 1.0693296018318281, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.125831365585327, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8729586005210876, "num_tokens": 319080808.0, "step": 8406 }, { "epoch": 1.0694568121104184, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6745898723602295, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8759784698486328, "num_tokens": 319123123.0, "step": 8407 }, { "epoch": 1.069584022389009, "ewc_loss": 0.0625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.078507900238037, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8779376745223999, "num_tokens": 319155116.0, "step": 8408 }, { "epoch": 1.0697112326675995, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.832915782928467, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8775972127914429, "num_tokens": 319188087.0, "step": 8409 }, { "epoch": 1.06983844294619, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1006433963775635, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8593478202819824, "num_tokens": 319228820.0, "step": 8410 }, { "epoch": 1.0699656532247805, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.99633526802063, "learning_rate": 1e-05, "loss": 0.5358, "mean_token_accuracy": 0.8480238914489746, "num_tokens": 319266046.0, "step": 8411 }, { "epoch": 1.070092863503371, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.236086368560791, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8742830157279968, "num_tokens": 319305209.0, "step": 8412 }, { "epoch": 1.0702200737819616, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.049628734588623, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8661257028579712, "num_tokens": 319347463.0, "step": 8413 }, { "epoch": 1.0703472840605521, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4258224964141846, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8809295296669006, "num_tokens": 319383965.0, "step": 8414 }, { "epoch": 1.0704744943391427, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9553451538085938, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.878860592842102, "num_tokens": 319418330.0, "step": 8415 }, { "epoch": 1.0706017046177332, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.505568742752075, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8680015802383423, "num_tokens": 319460518.0, "step": 8416 }, { "epoch": 1.0707289148963237, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7782270908355713, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8642145991325378, "num_tokens": 319504044.0, "step": 8417 }, { "epoch": 1.070856125174914, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5027992725372314, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8721292018890381, "num_tokens": 319540707.0, "step": 8418 }, { "epoch": 1.0709833354535045, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.609705686569214, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8756378293037415, "num_tokens": 319580968.0, "step": 8419 }, { "epoch": 1.071110545732095, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.452101707458496, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8599134683609009, "num_tokens": 319619271.0, "step": 8420 }, { "epoch": 1.0712377560106856, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4878528118133545, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8640272617340088, "num_tokens": 319657728.0, "step": 8421 }, { "epoch": 1.0713649662892761, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3834307193756104, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8628394603729248, "num_tokens": 319696807.0, "step": 8422 }, { "epoch": 1.0714921765678667, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4185168743133545, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8598470687866211, "num_tokens": 319734397.0, "step": 8423 }, { "epoch": 1.0716193868464572, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3327479362487793, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8567005395889282, "num_tokens": 319769247.0, "step": 8424 }, { "epoch": 1.0717465971250477, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.393458127975464, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8672717809677124, "num_tokens": 319801147.0, "step": 8425 }, { "epoch": 1.0718738074036382, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.27028751373291, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8776743412017822, "num_tokens": 319834083.0, "step": 8426 }, { "epoch": 1.0720010176822288, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3881590366363525, "learning_rate": 1e-05, "loss": 0.3932, "mean_token_accuracy": 0.8837212324142456, "num_tokens": 319872586.0, "step": 8427 }, { "epoch": 1.0721282279608193, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2293078899383545, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.870209813117981, "num_tokens": 319902756.0, "step": 8428 }, { "epoch": 1.0722554382394098, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.405735969543457, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8740147352218628, "num_tokens": 319943610.0, "step": 8429 }, { "epoch": 1.0723826485180004, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1694371700286865, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.864263117313385, "num_tokens": 319983762.0, "step": 8430 }, { "epoch": 1.0725098587965907, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.471846342086792, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8708856105804443, "num_tokens": 320019560.0, "step": 8431 }, { "epoch": 1.0726370690751812, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1429803371429443, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8646597862243652, "num_tokens": 320052907.0, "step": 8432 }, { "epoch": 1.0727642793537717, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5502915382385254, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8532955646514893, "num_tokens": 320087670.0, "step": 8433 }, { "epoch": 1.0728914896323622, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.054513692855835, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8703285455703735, "num_tokens": 320125768.0, "step": 8434 }, { "epoch": 1.0730186999109528, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6903860569000244, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8788621425628662, "num_tokens": 320163498.0, "step": 8435 }, { "epoch": 1.0731459101895433, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.028167247772217, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.865975022315979, "num_tokens": 320200951.0, "step": 8436 }, { "epoch": 1.0732731204681338, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8993887901306152, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8769519329071045, "num_tokens": 320239258.0, "step": 8437 }, { "epoch": 1.0734003307467244, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.0887622833251953, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8662358522415161, "num_tokens": 320276872.0, "step": 8438 }, { "epoch": 1.073527541025315, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.087063789367676, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8676207661628723, "num_tokens": 320315340.0, "step": 8439 }, { "epoch": 1.0736547513039054, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.2996387481689453, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8765197992324829, "num_tokens": 320351213.0, "step": 8440 }, { "epoch": 1.073781961582496, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.101929664611816, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8655115962028503, "num_tokens": 320391456.0, "step": 8441 }, { "epoch": 1.0739091718610865, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5053722858428955, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8741989135742188, "num_tokens": 320428193.0, "step": 8442 }, { "epoch": 1.0740363821396768, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.950551748275757, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8732386827468872, "num_tokens": 320466685.0, "step": 8443 }, { "epoch": 1.0741635924182673, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.556858777999878, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8618444800376892, "num_tokens": 320508936.0, "step": 8444 }, { "epoch": 1.0742908026968578, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.775106906890869, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8493149280548096, "num_tokens": 320546644.0, "step": 8445 }, { "epoch": 1.0744180129754484, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.53104567527771, "learning_rate": 1e-05, "loss": 0.3751, "mean_token_accuracy": 0.8905733227729797, "num_tokens": 320571074.0, "step": 8446 }, { "epoch": 1.074545223254039, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000194549560546875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5921385288238525, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8654696941375732, "num_tokens": 320605412.0, "step": 8447 }, { "epoch": 1.0746724335326294, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.453845977783203, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8639395236968994, "num_tokens": 320643077.0, "step": 8448 }, { "epoch": 1.07479964381122, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4573309421539307, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8773242235183716, "num_tokens": 320680630.0, "step": 8449 }, { "epoch": 1.0749268540898105, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3792812824249268, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8794425129890442, "num_tokens": 320720405.0, "step": 8450 }, { "epoch": 1.075054064368401, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.404919385910034, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8737847805023193, "num_tokens": 320753453.0, "step": 8451 }, { "epoch": 1.0751812746469915, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00019550323486328125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3100054264068604, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8739432096481323, "num_tokens": 320795316.0, "step": 8452 }, { "epoch": 1.075308484925582, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3860673904418945, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8740463852882385, "num_tokens": 320834022.0, "step": 8453 }, { "epoch": 1.0754356952041726, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.257031202316284, "learning_rate": 1e-05, "loss": 0.5515, "mean_token_accuracy": 0.8349758386611938, "num_tokens": 320878410.0, "step": 8454 }, { "epoch": 1.0755629054827631, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.386174201965332, "learning_rate": 1e-05, "loss": 0.3847, "mean_token_accuracy": 0.8884125351905823, "num_tokens": 320918767.0, "step": 8455 }, { "epoch": 1.0756901157613534, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1983232498168945, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8713103532791138, "num_tokens": 320963845.0, "step": 8456 }, { "epoch": 1.075817326039944, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4283697605133057, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8762291669845581, "num_tokens": 321004811.0, "step": 8457 }, { "epoch": 1.0759445363185345, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.142975091934204, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8654170632362366, "num_tokens": 321040894.0, "step": 8458 }, { "epoch": 1.076071746597125, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.511314868927002, "learning_rate": 1e-05, "loss": 0.5068, "mean_token_accuracy": 0.8530172109603882, "num_tokens": 321083602.0, "step": 8459 }, { "epoch": 1.0761989568757155, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0825674533843994, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8673556447029114, "num_tokens": 321119444.0, "step": 8460 }, { "epoch": 1.076326167154306, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6416218280792236, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8745689392089844, "num_tokens": 321155125.0, "step": 8461 }, { "epoch": 1.0764533774328966, "ewc_loss": 0.06201171875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 3.0094759464263916, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8788567781448364, "num_tokens": 321193645.0, "step": 8462 }, { "epoch": 1.0765805877114871, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8827321529388428, "learning_rate": 1e-05, "loss": 0.5084, "mean_token_accuracy": 0.855599582195282, "num_tokens": 321227118.0, "step": 8463 }, { "epoch": 1.0767077979900777, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.0554072856903076, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8675891160964966, "num_tokens": 321263245.0, "step": 8464 }, { "epoch": 1.0768350082686682, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.112167835235596, "learning_rate": 1e-05, "loss": 0.527, "mean_token_accuracy": 0.8469090461730957, "num_tokens": 321301476.0, "step": 8465 }, { "epoch": 1.0769622185472587, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.265911102294922, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8720592260360718, "num_tokens": 321333464.0, "step": 8466 }, { "epoch": 1.077089428825849, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.191569805145264, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8688071370124817, "num_tokens": 321367807.0, "step": 8467 }, { "epoch": 1.0772166391044395, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5220165252685547, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8789381384849548, "num_tokens": 321408638.0, "step": 8468 }, { "epoch": 1.07734384938303, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.040178298950195, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8654171228408813, "num_tokens": 321451036.0, "step": 8469 }, { "epoch": 1.0774710596616206, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.611375093460083, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8815176486968994, "num_tokens": 321491423.0, "step": 8470 }, { "epoch": 1.0775982699402111, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.834027051925659, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8597151637077332, "num_tokens": 321527956.0, "step": 8471 }, { "epoch": 1.0777254802188017, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.569019079208374, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8793973922729492, "num_tokens": 321560185.0, "step": 8472 }, { "epoch": 1.0778526904973922, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.646885395050049, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8756200671195984, "num_tokens": 321602073.0, "step": 8473 }, { "epoch": 1.0779799007759827, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.480013132095337, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8676068782806396, "num_tokens": 321636722.0, "step": 8474 }, { "epoch": 1.0781071110545732, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5285632610321045, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8693865537643433, "num_tokens": 321675348.0, "step": 8475 }, { "epoch": 1.0782343213331638, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3868770599365234, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8693960905075073, "num_tokens": 321711983.0, "step": 8476 }, { "epoch": 1.0783615316117543, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.467592239379883, "learning_rate": 1e-05, "loss": 0.4028, "mean_token_accuracy": 0.8836460709571838, "num_tokens": 321748268.0, "step": 8477 }, { "epoch": 1.0784887418903448, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.30117130279541, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8656976222991943, "num_tokens": 321783082.0, "step": 8478 }, { "epoch": 1.0786159521689354, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4559435844421387, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.863763689994812, "num_tokens": 321823634.0, "step": 8479 }, { "epoch": 1.0787431624475257, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2170467376708984, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8769410848617554, "num_tokens": 321859033.0, "step": 8480 }, { "epoch": 1.0788703727261162, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5091562271118164, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.86742103099823, "num_tokens": 321896227.0, "step": 8481 }, { "epoch": 1.0789975830047067, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.140352725982666, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8697895407676697, "num_tokens": 321934329.0, "step": 8482 }, { "epoch": 1.0791247932832972, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6052749156951904, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8647082448005676, "num_tokens": 321977426.0, "step": 8483 }, { "epoch": 1.0792520035618878, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0668320655822754, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8707357048988342, "num_tokens": 322014603.0, "step": 8484 }, { "epoch": 1.0793792138404783, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.782205104827881, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8789116740226746, "num_tokens": 322051809.0, "step": 8485 }, { "epoch": 1.0795064241190688, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0581228733062744, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8685675263404846, "num_tokens": 322091628.0, "step": 8486 }, { "epoch": 1.0796336343976594, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9969725608825684, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8624094724655151, "num_tokens": 322130616.0, "step": 8487 }, { "epoch": 1.0797608446762499, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.192750930786133, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8828520178794861, "num_tokens": 322172525.0, "step": 8488 }, { "epoch": 1.0798880549548404, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.106398105621338, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8749034404754639, "num_tokens": 322213438.0, "step": 8489 }, { "epoch": 1.080015265233431, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.4438223838806152, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8761192560195923, "num_tokens": 322252492.0, "step": 8490 }, { "epoch": 1.0801424755120215, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.980259656906128, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8705193996429443, "num_tokens": 322295320.0, "step": 8491 }, { "epoch": 1.0802696857906118, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.551044225692749, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8840218782424927, "num_tokens": 322335475.0, "step": 8492 }, { "epoch": 1.0803968960692023, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.810380220413208, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8747216463088989, "num_tokens": 322373559.0, "step": 8493 }, { "epoch": 1.0805241063477928, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.560391426086426, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8611871004104614, "num_tokens": 322407978.0, "step": 8494 }, { "epoch": 1.0806513166263834, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5819692611694336, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8695530891418457, "num_tokens": 322443657.0, "step": 8495 }, { "epoch": 1.080778526904974, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4995241165161133, "learning_rate": 1e-05, "loss": 0.3929, "mean_token_accuracy": 0.8880085945129395, "num_tokens": 322479424.0, "step": 8496 }, { "epoch": 1.0809057371835644, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.421086072921753, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8567372560501099, "num_tokens": 322523112.0, "step": 8497 }, { "epoch": 1.081032947462155, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4624478816986084, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8753476142883301, "num_tokens": 322560444.0, "step": 8498 }, { "epoch": 1.0811601577407455, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3271961212158203, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8715858459472656, "num_tokens": 322591211.0, "step": 8499 }, { "epoch": 1.081287368019336, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4438819885253906, "learning_rate": 1e-05, "loss": 0.5123, "mean_token_accuracy": 0.8519896864891052, "num_tokens": 322625129.0, "step": 8500 }, { "epoch": 1.0814145782979265, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.246515989303589, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8733050227165222, "num_tokens": 322663428.0, "step": 8501 }, { "epoch": 1.081541788576517, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4259769916534424, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8778942823410034, "num_tokens": 322707936.0, "step": 8502 }, { "epoch": 1.0816689988551076, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1845970153808594, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8814834356307983, "num_tokens": 322748605.0, "step": 8503 }, { "epoch": 1.0817962091336981, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4776909351348877, "learning_rate": 1e-05, "loss": 0.4943, "mean_token_accuracy": 0.8513762354850769, "num_tokens": 322788009.0, "step": 8504 }, { "epoch": 1.0819234194122884, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1237030029296875, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8661236763000488, "num_tokens": 322825159.0, "step": 8505 }, { "epoch": 1.082050629690879, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.566202163696289, "learning_rate": 1e-05, "loss": 0.5237, "mean_token_accuracy": 0.8452014923095703, "num_tokens": 322866769.0, "step": 8506 }, { "epoch": 1.0821778399694695, "ewc_loss": 0.06298828125, "ewc_loss_diag": 0.0001964569091796875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.058868408203125, "learning_rate": 1e-05, "loss": 0.3957, "mean_token_accuracy": 0.878589928150177, "num_tokens": 322901851.0, "step": 8507 }, { "epoch": 1.08230505024806, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7137434482574463, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8735847473144531, "num_tokens": 322938527.0, "step": 8508 }, { "epoch": 1.0824322605266505, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0288891792297363, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8728044629096985, "num_tokens": 322972135.0, "step": 8509 }, { "epoch": 1.082559470805241, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9246678352355957, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8582752346992493, "num_tokens": 323010451.0, "step": 8510 }, { "epoch": 1.0826866810838316, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.087740421295166, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8710730075836182, "num_tokens": 323047583.0, "step": 8511 }, { "epoch": 1.0828138913624221, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.120222091674805, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.862123429775238, "num_tokens": 323088754.0, "step": 8512 }, { "epoch": 1.0829411016410126, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.315753221511841, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8743857145309448, "num_tokens": 323128510.0, "step": 8513 }, { "epoch": 1.0830683119196032, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.142589092254639, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8648760318756104, "num_tokens": 323172557.0, "step": 8514 }, { "epoch": 1.0831955221981937, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.549825668334961, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.878923773765564, "num_tokens": 323208298.0, "step": 8515 }, { "epoch": 1.083322732476784, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.962975263595581, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8602527379989624, "num_tokens": 323250824.0, "step": 8516 }, { "epoch": 1.0834499427553745, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5928783416748047, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8710851669311523, "num_tokens": 323290792.0, "step": 8517 }, { "epoch": 1.083577153033965, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7543492317199707, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8654505014419556, "num_tokens": 323327806.0, "step": 8518 }, { "epoch": 1.0837043633125556, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.54233980178833, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8692126870155334, "num_tokens": 323364275.0, "step": 8519 }, { "epoch": 1.0838315735911461, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5867652893066406, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8772692680358887, "num_tokens": 323395671.0, "step": 8520 }, { "epoch": 1.0839587838697367, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4579241275787354, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8713488578796387, "num_tokens": 323432719.0, "step": 8521 }, { "epoch": 1.0840859941483272, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4647302627563477, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8623781800270081, "num_tokens": 323474431.0, "step": 8522 }, { "epoch": 1.0842132044269177, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3814892768859863, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8726083636283875, "num_tokens": 323515439.0, "step": 8523 }, { "epoch": 1.0843404147055082, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4080617427825928, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.860394299030304, "num_tokens": 323552613.0, "step": 8524 }, { "epoch": 1.0844676249840988, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3171591758728027, "learning_rate": 1e-05, "loss": 0.3848, "mean_token_accuracy": 0.8888946175575256, "num_tokens": 323590037.0, "step": 8525 }, { "epoch": 1.0845948352626893, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3763949871063232, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8811461925506592, "num_tokens": 323634540.0, "step": 8526 }, { "epoch": 1.0847220455412798, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.271270990371704, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8728480339050293, "num_tokens": 323666772.0, "step": 8527 }, { "epoch": 1.0848492558198704, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3793933391571045, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8698114156723022, "num_tokens": 323707847.0, "step": 8528 }, { "epoch": 1.0849764660984607, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2217977046966553, "learning_rate": 1e-05, "loss": 0.3913, "mean_token_accuracy": 0.8860697746276855, "num_tokens": 323737648.0, "step": 8529 }, { "epoch": 1.0851036763770512, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.397822856903076, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8757296800613403, "num_tokens": 323772621.0, "step": 8530 }, { "epoch": 1.0852308866556417, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1885151863098145, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8594790101051331, "num_tokens": 323806384.0, "step": 8531 }, { "epoch": 1.0853580969342322, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4278082847595215, "learning_rate": 1e-05, "loss": 0.3803, "mean_token_accuracy": 0.8903462886810303, "num_tokens": 323845287.0, "step": 8532 }, { "epoch": 1.0854853072128228, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.113085985183716, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8770233988761902, "num_tokens": 323884182.0, "step": 8533 }, { "epoch": 1.0856125174914133, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.550210475921631, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8523991107940674, "num_tokens": 323919357.0, "step": 8534 }, { "epoch": 1.0857397277700038, "ewc_loss": 0.062255859375, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.034719705581665, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8780509233474731, "num_tokens": 323957993.0, "step": 8535 }, { "epoch": 1.0858669380485944, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7309749126434326, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.880800187587738, "num_tokens": 323991726.0, "step": 8536 }, { "epoch": 1.0859941483271849, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00019741058349609375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.9903781414031982, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8698998093605042, "num_tokens": 324031639.0, "step": 8537 }, { "epoch": 1.0861213586057754, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9971206188201904, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8599660992622375, "num_tokens": 324063359.0, "step": 8538 }, { "epoch": 1.086248568884366, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.091714859008789, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8661715388298035, "num_tokens": 324105058.0, "step": 8539 }, { "epoch": 1.0863757791629565, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.230269432067871, "learning_rate": 1e-05, "loss": 0.5026, "mean_token_accuracy": 0.859994649887085, "num_tokens": 324144209.0, "step": 8540 }, { "epoch": 1.0865029894415468, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4217536449432373, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8743764162063599, "num_tokens": 324179194.0, "step": 8541 }, { "epoch": 1.0866301997201373, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.183279991149902, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8769027590751648, "num_tokens": 324218405.0, "step": 8542 }, { "epoch": 1.0867574099987278, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6375465393066406, "learning_rate": 1e-05, "loss": 0.5369, "mean_token_accuracy": 0.8450762033462524, "num_tokens": 324258298.0, "step": 8543 }, { "epoch": 1.0868846202773184, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9621989727020264, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.878961980342865, "num_tokens": 324296518.0, "step": 8544 }, { "epoch": 1.0870118305559089, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6267974376678467, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8678439855575562, "num_tokens": 324336724.0, "step": 8545 }, { "epoch": 1.0871390408344994, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7787325382232666, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8635903000831604, "num_tokens": 324373464.0, "step": 8546 }, { "epoch": 1.08726625111309, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5308101177215576, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8744693398475647, "num_tokens": 324410893.0, "step": 8547 }, { "epoch": 1.0873934613916805, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.636812210083008, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8758775591850281, "num_tokens": 324446765.0, "step": 8548 }, { "epoch": 1.087520671670271, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.48879337310791, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8715761303901672, "num_tokens": 324482745.0, "step": 8549 }, { "epoch": 1.0876478819488615, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.529510974884033, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8771486282348633, "num_tokens": 324513764.0, "step": 8550 }, { "epoch": 1.087775092227452, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.351323366165161, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8686683177947998, "num_tokens": 324549502.0, "step": 8551 }, { "epoch": 1.0879023025060426, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.481156349182129, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8685247898101807, "num_tokens": 324584456.0, "step": 8552 }, { "epoch": 1.0880295127846331, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2645621299743652, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8628688454627991, "num_tokens": 324624070.0, "step": 8553 }, { "epoch": 1.0881567230632234, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4963245391845703, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8693655133247375, "num_tokens": 324665971.0, "step": 8554 }, { "epoch": 1.088283933341814, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1776747703552246, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.8834960460662842, "num_tokens": 324706705.0, "step": 8555 }, { "epoch": 1.0884111436204045, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5811243057250977, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8609884977340698, "num_tokens": 324744701.0, "step": 8556 }, { "epoch": 1.088538353898995, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1180474758148193, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8680763840675354, "num_tokens": 324781447.0, "step": 8557 }, { "epoch": 1.0886655641775855, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0001983642578125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7135167121887207, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8731710314750671, "num_tokens": 324816753.0, "step": 8558 }, { "epoch": 1.088792774456176, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0829784870147705, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8639345169067383, "num_tokens": 324855093.0, "step": 8559 }, { "epoch": 1.0889199847347666, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.894132375717163, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.871021032333374, "num_tokens": 324893107.0, "step": 8560 }, { "epoch": 1.0890471950133571, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1479969024658203, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8690308332443237, "num_tokens": 324927589.0, "step": 8561 }, { "epoch": 1.0891744052919476, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.084457874298096, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8785126209259033, "num_tokens": 324958801.0, "step": 8562 }, { "epoch": 1.0893016155705382, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.347882032394409, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8744573593139648, "num_tokens": 324995617.0, "step": 8563 }, { "epoch": 1.0894288258491287, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.02976655960083, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8649537563323975, "num_tokens": 325035702.0, "step": 8564 }, { "epoch": 1.089556036127719, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.516923189163208, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8595705628395081, "num_tokens": 325078063.0, "step": 8565 }, { "epoch": 1.0896832464063095, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8474409580230713, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8672329187393188, "num_tokens": 325113127.0, "step": 8566 }, { "epoch": 1.0898104566849, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5538055896759033, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8650611639022827, "num_tokens": 325149451.0, "step": 8567 }, { "epoch": 1.0899376669634906, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6177589893341064, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8667548894882202, "num_tokens": 325193265.0, "step": 8568 }, { "epoch": 1.0900648772420811, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5063834190368652, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8733654022216797, "num_tokens": 325235160.0, "step": 8569 }, { "epoch": 1.0901920875206716, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4483203887939453, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8671561479568481, "num_tokens": 325276299.0, "step": 8570 }, { "epoch": 1.0903192977992622, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.475060224533081, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8554067611694336, "num_tokens": 325313509.0, "step": 8571 }, { "epoch": 1.0904465080778527, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.329615354537964, "learning_rate": 1e-05, "loss": 0.4751, "mean_token_accuracy": 0.8597669005393982, "num_tokens": 325355883.0, "step": 8572 }, { "epoch": 1.0905737183564432, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4551401138305664, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8752819299697876, "num_tokens": 325392937.0, "step": 8573 }, { "epoch": 1.0907009286350338, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2567734718322754, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8709762692451477, "num_tokens": 325422132.0, "step": 8574 }, { "epoch": 1.0908281389136243, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4539637565612793, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8663283586502075, "num_tokens": 325460383.0, "step": 8575 }, { "epoch": 1.0909553491922148, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.188523769378662, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8615573644638062, "num_tokens": 325499242.0, "step": 8576 }, { "epoch": 1.0910825594708053, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.480473518371582, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8757081031799316, "num_tokens": 325535389.0, "step": 8577 }, { "epoch": 1.0912097697493957, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1280529499053955, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8772162199020386, "num_tokens": 325574459.0, "step": 8578 }, { "epoch": 1.0913369800279862, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5736498832702637, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8547281622886658, "num_tokens": 325616296.0, "step": 8579 }, { "epoch": 1.0914641903065767, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.067286252975464, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8595410585403442, "num_tokens": 325652707.0, "step": 8580 }, { "epoch": 1.0915914005851672, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.713857889175415, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8819363117218018, "num_tokens": 325692423.0, "step": 8581 }, { "epoch": 1.0917186108637578, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.034316301345825, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8697572350502014, "num_tokens": 325731638.0, "step": 8582 }, { "epoch": 1.0918458211423483, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9059548377990723, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8765789866447449, "num_tokens": 325771482.0, "step": 8583 }, { "epoch": 1.0919730314209388, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1045846939086914, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8629138469696045, "num_tokens": 325807447.0, "step": 8584 }, { "epoch": 1.0921002416995294, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0782623291015625, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8704867362976074, "num_tokens": 325848070.0, "step": 8585 }, { "epoch": 1.0922274519781199, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.2952964305877686, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8754541277885437, "num_tokens": 325889027.0, "step": 8586 }, { "epoch": 1.0923546622567104, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.119902610778809, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8663554191589355, "num_tokens": 325926547.0, "step": 8587 }, { "epoch": 1.092481872535301, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.507384777069092, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8748157620429993, "num_tokens": 325968968.0, "step": 8588 }, { "epoch": 1.0926090828138915, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.96600604057312, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8698068261146545, "num_tokens": 326005060.0, "step": 8589 }, { "epoch": 1.0927362930924818, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5755093097686768, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.8587638139724731, "num_tokens": 326049895.0, "step": 8590 }, { "epoch": 1.0928635033710723, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.761763572692871, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8605798482894897, "num_tokens": 326087772.0, "step": 8591 }, { "epoch": 1.0929907136496628, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5254268646240234, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8741103410720825, "num_tokens": 326126518.0, "step": 8592 }, { "epoch": 1.0931179239282534, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5801730155944824, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8606416583061218, "num_tokens": 326166672.0, "step": 8593 }, { "epoch": 1.0932451342068439, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.455317258834839, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8657497763633728, "num_tokens": 326209099.0, "step": 8594 }, { "epoch": 1.0933723444854344, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.482318162918091, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8615010976791382, "num_tokens": 326244101.0, "step": 8595 }, { "epoch": 1.093499554764025, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.42655873298645, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.868445873260498, "num_tokens": 326279997.0, "step": 8596 }, { "epoch": 1.0936267650426155, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.388678550720215, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8613979816436768, "num_tokens": 326321241.0, "step": 8597 }, { "epoch": 1.093753975321206, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3363230228424072, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8697466850280762, "num_tokens": 326359662.0, "step": 8598 }, { "epoch": 1.0938811855997965, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3473315238952637, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8778437376022339, "num_tokens": 326397072.0, "step": 8599 }, { "epoch": 1.094008395878387, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2929396629333496, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8652983903884888, "num_tokens": 326440029.0, "step": 8600 }, { "epoch": 1.0941356061569776, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3322911262512207, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8716236352920532, "num_tokens": 326479586.0, "step": 8601 }, { "epoch": 1.094262816435568, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2596068382263184, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8670821189880371, "num_tokens": 326518235.0, "step": 8602 }, { "epoch": 1.0943900267141584, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3379404544830322, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8749866485595703, "num_tokens": 326559420.0, "step": 8603 }, { "epoch": 1.094517236992749, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2286949157714844, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8642833232879639, "num_tokens": 326597020.0, "step": 8604 }, { "epoch": 1.0946444472713395, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3758041858673096, "learning_rate": 1e-05, "loss": 0.5232, "mean_token_accuracy": 0.8444846868515015, "num_tokens": 326632248.0, "step": 8605 }, { "epoch": 1.09477165754993, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.179809331893921, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8730665445327759, "num_tokens": 326666185.0, "step": 8606 }, { "epoch": 1.0948988678285205, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.437567949295044, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8505538702011108, "num_tokens": 326700356.0, "step": 8607 }, { "epoch": 1.095026078107111, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.104487180709839, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8721590042114258, "num_tokens": 326739448.0, "step": 8608 }, { "epoch": 1.0951532883857016, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.547962188720703, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8695772290229797, "num_tokens": 326776290.0, "step": 8609 }, { "epoch": 1.0952804986642921, "ewc_loss": 0.0625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0168168544769287, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8734296560287476, "num_tokens": 326817975.0, "step": 8610 }, { "epoch": 1.0954077089428826, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.754709482192993, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8720430135726929, "num_tokens": 326852404.0, "step": 8611 }, { "epoch": 1.0955349192214732, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 2.9816811084747314, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8595874309539795, "num_tokens": 326888519.0, "step": 8612 }, { "epoch": 1.0956621295000637, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.038808345794678, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8623905181884766, "num_tokens": 326925644.0, "step": 8613 }, { "epoch": 1.095789339778654, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.095163106918335, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8677741289138794, "num_tokens": 326961490.0, "step": 8614 }, { "epoch": 1.0959165500572445, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2936201095581055, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8579649925231934, "num_tokens": 326998726.0, "step": 8615 }, { "epoch": 1.096043760335835, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.452533721923828, "learning_rate": 1e-05, "loss": 0.5102, "mean_token_accuracy": 0.8549939393997192, "num_tokens": 327040717.0, "step": 8616 }, { "epoch": 1.0961709706144256, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2662200927734375, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.866218090057373, "num_tokens": 327070597.0, "step": 8617 }, { "epoch": 1.0962981808930161, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7157955169677734, "learning_rate": 1e-05, "loss": 0.519, "mean_token_accuracy": 0.8529399037361145, "num_tokens": 327108598.0, "step": 8618 }, { "epoch": 1.0964253911716066, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.014525413513184, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8799289464950562, "num_tokens": 327142750.0, "step": 8619 }, { "epoch": 1.0965526014501972, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6864736080169678, "learning_rate": 1e-05, "loss": 0.4891, "mean_token_accuracy": 0.8591327667236328, "num_tokens": 327182536.0, "step": 8620 }, { "epoch": 1.0966798117287877, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8185248374938965, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8676482439041138, "num_tokens": 327218135.0, "step": 8621 }, { "epoch": 1.0968070220073782, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002002716064453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5745668411254883, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8808311223983765, "num_tokens": 327251581.0, "step": 8622 }, { "epoch": 1.0969342322859688, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6797478199005127, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8698330521583557, "num_tokens": 327288466.0, "step": 8623 }, { "epoch": 1.0970614425645593, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4487431049346924, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8740091919898987, "num_tokens": 327325646.0, "step": 8624 }, { "epoch": 1.0971886528431498, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.61271595954895, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8617998361587524, "num_tokens": 327358244.0, "step": 8625 }, { "epoch": 1.0973158631217403, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.338285207748413, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8763360977172852, "num_tokens": 327397315.0, "step": 8626 }, { "epoch": 1.0974430734003306, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6120996475219727, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8690555095672607, "num_tokens": 327432642.0, "step": 8627 }, { "epoch": 1.0975702836789212, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.242058277130127, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8574119210243225, "num_tokens": 327478563.0, "step": 8628 }, { "epoch": 1.0976974939575117, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6633269786834717, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8763821125030518, "num_tokens": 327514295.0, "step": 8629 }, { "epoch": 1.0978247042361022, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1672840118408203, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8758450746536255, "num_tokens": 327550496.0, "step": 8630 }, { "epoch": 1.0979519145146928, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7773821353912354, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8858829140663147, "num_tokens": 327584692.0, "step": 8631 }, { "epoch": 1.0980791247932833, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1474995613098145, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8772488832473755, "num_tokens": 327627710.0, "step": 8632 }, { "epoch": 1.0982063350718738, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.922757625579834, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8620691299438477, "num_tokens": 327666958.0, "step": 8633 }, { "epoch": 1.0983335453504643, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2487103939056396, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8751345872879028, "num_tokens": 327700829.0, "step": 8634 }, { "epoch": 1.0984607556290549, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9957833290100098, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.860909104347229, "num_tokens": 327741410.0, "step": 8635 }, { "epoch": 1.0985879659076454, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4154365062713623, "learning_rate": 1e-05, "loss": 0.4945, "mean_token_accuracy": 0.8548076748847961, "num_tokens": 327778305.0, "step": 8636 }, { "epoch": 1.098715176186236, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.894512414932251, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8603152632713318, "num_tokens": 327813384.0, "step": 8637 }, { "epoch": 1.0988423864648265, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.493849039077759, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8725240230560303, "num_tokens": 327851928.0, "step": 8638 }, { "epoch": 1.0989695967434168, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.69045090675354, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8802765607833862, "num_tokens": 327890053.0, "step": 8639 }, { "epoch": 1.0990968070220073, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5094141960144043, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8669403195381165, "num_tokens": 327928947.0, "step": 8640 }, { "epoch": 1.0992240173005978, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.489658832550049, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8742179870605469, "num_tokens": 327966286.0, "step": 8641 }, { "epoch": 1.0993512275791884, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.487711191177368, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8667291402816772, "num_tokens": 328005480.0, "step": 8642 }, { "epoch": 1.0994784378577789, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3303418159484863, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8726357221603394, "num_tokens": 328040587.0, "step": 8643 }, { "epoch": 1.0996056481363694, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4945297241210938, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8625163435935974, "num_tokens": 328075856.0, "step": 8644 }, { "epoch": 1.09973285841496, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.213484764099121, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8580178618431091, "num_tokens": 328116243.0, "step": 8645 }, { "epoch": 1.0998600686935505, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5304651260375977, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8713126182556152, "num_tokens": 328151854.0, "step": 8646 }, { "epoch": 1.099987278972141, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.134800672531128, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8617389798164368, "num_tokens": 328191455.0, "step": 8647 }, { "epoch": 1.1001144892507315, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6080713272094727, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8750893473625183, "num_tokens": 328232034.0, "step": 8648 }, { "epoch": 1.100241699529322, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0802183151245117, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.870244562625885, "num_tokens": 328265581.0, "step": 8649 }, { "epoch": 1.1003689098079126, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7535479068756104, "learning_rate": 1e-05, "loss": 0.5309, "mean_token_accuracy": 0.8516700863838196, "num_tokens": 328304354.0, "step": 8650 }, { "epoch": 1.100496120086503, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.0632266998291016, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8603116273880005, "num_tokens": 328347924.0, "step": 8651 }, { "epoch": 1.1006233303650934, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.921268939971924, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8844271898269653, "num_tokens": 328385603.0, "step": 8652 }, { "epoch": 1.100750540643684, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.1538259983062744, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8757302165031433, "num_tokens": 328422881.0, "step": 8653 }, { "epoch": 1.1008777509222745, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.06052827835083, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8664936423301697, "num_tokens": 328461191.0, "step": 8654 }, { "epoch": 1.101004961200865, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.345906972885132, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8672541975975037, "num_tokens": 328499824.0, "step": 8655 }, { "epoch": 1.1011321714794555, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.027120590209961, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8749858140945435, "num_tokens": 328538456.0, "step": 8656 }, { "epoch": 1.101259381758046, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5019447803497314, "learning_rate": 1e-05, "loss": 0.4969, "mean_token_accuracy": 0.8545723557472229, "num_tokens": 328574722.0, "step": 8657 }, { "epoch": 1.1013865920366366, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8698394298553467, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8706782460212708, "num_tokens": 328608248.0, "step": 8658 }, { "epoch": 1.101513802315227, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.514476776123047, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.869455099105835, "num_tokens": 328651011.0, "step": 8659 }, { "epoch": 1.1016410125938176, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6867175102233887, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8647866249084473, "num_tokens": 328694438.0, "step": 8660 }, { "epoch": 1.1017682228724082, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.481884002685547, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8698512315750122, "num_tokens": 328731851.0, "step": 8661 }, { "epoch": 1.1018954331509987, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.527332067489624, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8687832355499268, "num_tokens": 328772159.0, "step": 8662 }, { "epoch": 1.102022643429589, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.414440155029297, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8847740888595581, "num_tokens": 328814712.0, "step": 8663 }, { "epoch": 1.1021498537081795, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4317333698272705, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8650408983230591, "num_tokens": 328854243.0, "step": 8664 }, { "epoch": 1.10227706398677, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3711252212524414, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8569349646568298, "num_tokens": 328894109.0, "step": 8665 }, { "epoch": 1.1024042742653606, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.361048698425293, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8798398375511169, "num_tokens": 328937275.0, "step": 8666 }, { "epoch": 1.1025314845439511, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3276851177215576, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8780338168144226, "num_tokens": 328974572.0, "step": 8667 }, { "epoch": 1.1026586948225416, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3434672355651855, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8649061918258667, "num_tokens": 329013871.0, "step": 8668 }, { "epoch": 1.1027859051011322, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.298488140106201, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8543184399604797, "num_tokens": 329050333.0, "step": 8669 }, { "epoch": 1.1029131153797227, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.326176643371582, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8645229935646057, "num_tokens": 329091817.0, "step": 8670 }, { "epoch": 1.1030403256583132, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2609620094299316, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8739521503448486, "num_tokens": 329128523.0, "step": 8671 }, { "epoch": 1.1031675359369038, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3267221450805664, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8855724930763245, "num_tokens": 329166320.0, "step": 8672 }, { "epoch": 1.1032947462154943, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2510268688201904, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8711812496185303, "num_tokens": 329195140.0, "step": 8673 }, { "epoch": 1.1034219564940848, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3264553546905518, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8788720369338989, "num_tokens": 329230814.0, "step": 8674 }, { "epoch": 1.1035491667726753, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2108006477355957, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8760043382644653, "num_tokens": 329270695.0, "step": 8675 }, { "epoch": 1.1036763770512656, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3703272342681885, "learning_rate": 1e-05, "loss": 0.4961, "mean_token_accuracy": 0.85533607006073, "num_tokens": 329308061.0, "step": 8676 }, { "epoch": 1.1038035873298562, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.166628837585449, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8669321537017822, "num_tokens": 329347093.0, "step": 8677 }, { "epoch": 1.1039307976084467, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.428192615509033, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8666955232620239, "num_tokens": 329385009.0, "step": 8678 }, { "epoch": 1.1040580078870372, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.113003730773926, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8671629428863525, "num_tokens": 329425317.0, "step": 8679 }, { "epoch": 1.1041852181656278, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5341274738311768, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8769904375076294, "num_tokens": 329461581.0, "step": 8680 }, { "epoch": 1.1043124284442183, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.035072088241577, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8808262348175049, "num_tokens": 329494064.0, "step": 8681 }, { "epoch": 1.1044396387228088, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.721057891845703, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8746228814125061, "num_tokens": 329530807.0, "step": 8682 }, { "epoch": 1.1045668490013993, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020122528076171875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.962719440460205, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8487279415130615, "num_tokens": 329575689.0, "step": 8683 }, { "epoch": 1.1046940592799899, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.008590221405029, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8623819351196289, "num_tokens": 329616855.0, "step": 8684 }, { "epoch": 1.1048212695585804, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.051246166229248, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.870353102684021, "num_tokens": 329654085.0, "step": 8685 }, { "epoch": 1.104948479837171, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.31358528137207, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8692805767059326, "num_tokens": 329688375.0, "step": 8686 }, { "epoch": 1.1050756901157615, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4057059288024902, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.8580717444419861, "num_tokens": 329734258.0, "step": 8687 }, { "epoch": 1.1052029003943518, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.336684703826904, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.871190071105957, "num_tokens": 329772080.0, "step": 8688 }, { "epoch": 1.1053301106729423, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.731698751449585, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8763854503631592, "num_tokens": 329817809.0, "step": 8689 }, { "epoch": 1.1054573209515328, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.088195323944092, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8759667873382568, "num_tokens": 329856063.0, "step": 8690 }, { "epoch": 1.1055845312301233, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.761887311935425, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8589935898780823, "num_tokens": 329892548.0, "step": 8691 }, { "epoch": 1.1057117415087139, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.857870101928711, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8853819966316223, "num_tokens": 329931897.0, "step": 8692 }, { "epoch": 1.1058389517873044, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6167681217193604, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8635534048080444, "num_tokens": 329967477.0, "step": 8693 }, { "epoch": 1.105966162065895, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.707474946975708, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8678795099258423, "num_tokens": 330004840.0, "step": 8694 }, { "epoch": 1.1060933723444855, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4759373664855957, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8705557584762573, "num_tokens": 330048928.0, "step": 8695 }, { "epoch": 1.106220582623076, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.624140739440918, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.865847647190094, "num_tokens": 330084112.0, "step": 8696 }, { "epoch": 1.1063477929016665, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.355747938156128, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.859811007976532, "num_tokens": 330127533.0, "step": 8697 }, { "epoch": 1.106475003180257, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.608853340148926, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8502583503723145, "num_tokens": 330169379.0, "step": 8698 }, { "epoch": 1.1066022134588476, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2478201389312744, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8778359293937683, "num_tokens": 330205553.0, "step": 8699 }, { "epoch": 1.106729423737438, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.660097599029541, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8728598356246948, "num_tokens": 330241088.0, "step": 8700 }, { "epoch": 1.1068566340160284, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.163760185241699, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8703879117965698, "num_tokens": 330282520.0, "step": 8701 }, { "epoch": 1.106983844294619, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7882158756256104, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8650821447372437, "num_tokens": 330326231.0, "step": 8702 }, { "epoch": 1.1071110545732095, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.138808012008667, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8795552253723145, "num_tokens": 330363674.0, "step": 8703 }, { "epoch": 1.1072382648518, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.958012580871582, "learning_rate": 1e-05, "loss": 0.385, "mean_token_accuracy": 0.8927000164985657, "num_tokens": 330398890.0, "step": 8704 }, { "epoch": 1.1073654751303905, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.248745918273926, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8752095699310303, "num_tokens": 330434564.0, "step": 8705 }, { "epoch": 1.107492685408981, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.034107685089111, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.875643789768219, "num_tokens": 330475580.0, "step": 8706 }, { "epoch": 1.1076198956875716, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.425152540206909, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8816825151443481, "num_tokens": 330513255.0, "step": 8707 }, { "epoch": 1.107747105966162, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9379472732543945, "learning_rate": 1e-05, "loss": 0.4891, "mean_token_accuracy": 0.8607186675071716, "num_tokens": 330554101.0, "step": 8708 }, { "epoch": 1.1078743162447526, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.543647289276123, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8563035726547241, "num_tokens": 330595312.0, "step": 8709 }, { "epoch": 1.1080015265233432, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7114198207855225, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8771697878837585, "num_tokens": 330633510.0, "step": 8710 }, { "epoch": 1.1081287368019337, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.534036874771118, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8732572793960571, "num_tokens": 330673626.0, "step": 8711 }, { "epoch": 1.108255947080524, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5008621215820312, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.866150975227356, "num_tokens": 330715051.0, "step": 8712 }, { "epoch": 1.1083831573591145, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.504593849182129, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8691571950912476, "num_tokens": 330765603.0, "step": 8713 }, { "epoch": 1.108510367637705, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3508872985839844, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8664202690124512, "num_tokens": 330799427.0, "step": 8714 }, { "epoch": 1.1086375779162956, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5035195350646973, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8733686208724976, "num_tokens": 330840810.0, "step": 8715 }, { "epoch": 1.108764788194886, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2421622276306152, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8578819632530212, "num_tokens": 330878325.0, "step": 8716 }, { "epoch": 1.1088919984734766, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5179378986358643, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8816062211990356, "num_tokens": 330914817.0, "step": 8717 }, { "epoch": 1.1090192087520672, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.146848678588867, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8788127303123474, "num_tokens": 330951890.0, "step": 8718 }, { "epoch": 1.1091464190306577, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6085991859436035, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8592836856842041, "num_tokens": 330986807.0, "step": 8719 }, { "epoch": 1.1092736293092482, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.000202178955078125, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.075833797454834, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8622057437896729, "num_tokens": 331027132.0, "step": 8720 }, { "epoch": 1.1094008395878387, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7911789417266846, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.876406192779541, "num_tokens": 331062669.0, "step": 8721 }, { "epoch": 1.1095280498664293, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0649056434631348, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8734561800956726, "num_tokens": 331097376.0, "step": 8722 }, { "epoch": 1.1096552601450198, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9661619663238525, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8796517252922058, "num_tokens": 331130930.0, "step": 8723 }, { "epoch": 1.1097824704236103, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.1644599437713623, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8670538067817688, "num_tokens": 331172012.0, "step": 8724 }, { "epoch": 1.1099096807022006, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.115143775939941, "learning_rate": 1e-05, "loss": 0.5319, "mean_token_accuracy": 0.8476182222366333, "num_tokens": 331210016.0, "step": 8725 }, { "epoch": 1.1100368909807912, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.397994041442871, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8627735376358032, "num_tokens": 331251818.0, "step": 8726 }, { "epoch": 1.1101641012593817, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.067099571228027, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8582019805908203, "num_tokens": 331283034.0, "step": 8727 }, { "epoch": 1.1102913115379722, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5547680854797363, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8648539781570435, "num_tokens": 331324985.0, "step": 8728 }, { "epoch": 1.1104185218165628, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.889761447906494, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8712215423583984, "num_tokens": 331358595.0, "step": 8729 }, { "epoch": 1.1105457320951533, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5804171562194824, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8648442029953003, "num_tokens": 331387388.0, "step": 8730 }, { "epoch": 1.1106729423737438, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6973588466644287, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8637880086898804, "num_tokens": 331423562.0, "step": 8731 }, { "epoch": 1.1108001526523343, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.514082670211792, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8733561038970947, "num_tokens": 331458835.0, "step": 8732 }, { "epoch": 1.1109273629309249, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.542280435562134, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.861452579498291, "num_tokens": 331489889.0, "step": 8733 }, { "epoch": 1.1110545732095154, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.429656505584717, "learning_rate": 1e-05, "loss": 0.394, "mean_token_accuracy": 0.883343517780304, "num_tokens": 331529103.0, "step": 8734 }, { "epoch": 1.111181783488106, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.44739031791687, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8752517700195312, "num_tokens": 331566100.0, "step": 8735 }, { "epoch": 1.1113089937666965, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3571786880493164, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8762203454971313, "num_tokens": 331606704.0, "step": 8736 }, { "epoch": 1.1114362040452868, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4148690700531006, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8787858486175537, "num_tokens": 331642019.0, "step": 8737 }, { "epoch": 1.1115634143238773, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3085086345672607, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.870256245136261, "num_tokens": 331672888.0, "step": 8738 }, { "epoch": 1.1116906246024678, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3994972705841064, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8717994093894958, "num_tokens": 331710383.0, "step": 8739 }, { "epoch": 1.1118178348810583, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2372710704803467, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8737643957138062, "num_tokens": 331754299.0, "step": 8740 }, { "epoch": 1.1119450451596489, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4349350929260254, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8716081380844116, "num_tokens": 331789329.0, "step": 8741 }, { "epoch": 1.1120722554382394, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.181997537612915, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8756150603294373, "num_tokens": 331822727.0, "step": 8742 }, { "epoch": 1.11219946571683, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5013182163238525, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8736231327056885, "num_tokens": 331863091.0, "step": 8743 }, { "epoch": 1.1123266759954205, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1032254695892334, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8728201985359192, "num_tokens": 331900866.0, "step": 8744 }, { "epoch": 1.112453886274011, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.63283109664917, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8835583925247192, "num_tokens": 331936774.0, "step": 8745 }, { "epoch": 1.1125810965526015, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0447309017181396, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8738234043121338, "num_tokens": 331974629.0, "step": 8746 }, { "epoch": 1.112708306831192, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8217649459838867, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.877610981464386, "num_tokens": 332008707.0, "step": 8747 }, { "epoch": 1.1128355171097826, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.0437240600585938, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.882880687713623, "num_tokens": 332046973.0, "step": 8748 }, { "epoch": 1.112962727388373, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.048468589782715, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8620226383209229, "num_tokens": 332092658.0, "step": 8749 }, { "epoch": 1.1130899376669634, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.214144468307495, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8722671270370483, "num_tokens": 332129705.0, "step": 8750 }, { "epoch": 1.113217147945554, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.163550853729248, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8699038624763489, "num_tokens": 332163129.0, "step": 8751 }, { "epoch": 1.1133443582241445, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.482778310775757, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8672198057174683, "num_tokens": 332198148.0, "step": 8752 }, { "epoch": 1.113471568502735, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.015771389007568, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8716001510620117, "num_tokens": 332238068.0, "step": 8753 }, { "epoch": 1.1135987787813255, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.583923101425171, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8594886064529419, "num_tokens": 332277455.0, "step": 8754 }, { "epoch": 1.113725989059916, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.820885419845581, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8660414814949036, "num_tokens": 332314778.0, "step": 8755 }, { "epoch": 1.1138531993385066, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5570051670074463, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.873894989490509, "num_tokens": 332347863.0, "step": 8756 }, { "epoch": 1.113980409617097, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6147124767303467, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8833329677581787, "num_tokens": 332389577.0, "step": 8757 }, { "epoch": 1.1141076198956876, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.475646495819092, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8642113208770752, "num_tokens": 332434140.0, "step": 8758 }, { "epoch": 1.1142348301742782, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4909939765930176, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8734122514724731, "num_tokens": 332468843.0, "step": 8759 }, { "epoch": 1.1143620404528687, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4035470485687256, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.876662015914917, "num_tokens": 332509863.0, "step": 8760 }, { "epoch": 1.114489250731459, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4186813831329346, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8662578463554382, "num_tokens": 332550242.0, "step": 8761 }, { "epoch": 1.1146164610100495, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3429081439971924, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8692511320114136, "num_tokens": 332588606.0, "step": 8762 }, { "epoch": 1.11474367128864, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3906702995300293, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8725313544273376, "num_tokens": 332623787.0, "step": 8763 }, { "epoch": 1.1148708815672306, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.283806562423706, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8615642786026001, "num_tokens": 332665358.0, "step": 8764 }, { "epoch": 1.114998091845821, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3892316818237305, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.8809457421302795, "num_tokens": 332700541.0, "step": 8765 }, { "epoch": 1.1151253021244116, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2454214096069336, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8619706630706787, "num_tokens": 332737555.0, "step": 8766 }, { "epoch": 1.1152525124030022, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4110758304595947, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8613428473472595, "num_tokens": 332775957.0, "step": 8767 }, { "epoch": 1.1153797226815927, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1791844367980957, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8730555772781372, "num_tokens": 332812000.0, "step": 8768 }, { "epoch": 1.1155069329601832, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4656946659088135, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8857956528663635, "num_tokens": 332848974.0, "step": 8769 }, { "epoch": 1.1156341432387737, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1062088012695312, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8742547631263733, "num_tokens": 332890362.0, "step": 8770 }, { "epoch": 1.1157613535173643, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.589275360107422, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8677453398704529, "num_tokens": 332924547.0, "step": 8771 }, { "epoch": 1.1158885637959548, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0422184467315674, "learning_rate": 1e-05, "loss": 0.3975, "mean_token_accuracy": 0.8830264806747437, "num_tokens": 332957697.0, "step": 8772 }, { "epoch": 1.1160157740745453, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7834529876708984, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8717676401138306, "num_tokens": 332992671.0, "step": 8773 }, { "epoch": 1.1161429843531356, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.009587287902832, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8746527433395386, "num_tokens": 333033092.0, "step": 8774 }, { "epoch": 1.1162701946317262, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.032146453857422, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8646075129508972, "num_tokens": 333071631.0, "step": 8775 }, { "epoch": 1.1163974049103167, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.1533684730529785, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8783775568008423, "num_tokens": 333100931.0, "step": 8776 }, { "epoch": 1.1165246151889072, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.208469867706299, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8910064697265625, "num_tokens": 333141063.0, "step": 8777 }, { "epoch": 1.1166518254674977, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.439690351486206, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8700144290924072, "num_tokens": 333180322.0, "step": 8778 }, { "epoch": 1.1167790357460883, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.152673244476318, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8647128939628601, "num_tokens": 333219836.0, "step": 8779 }, { "epoch": 1.1169062460246788, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.630080223083496, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8638197779655457, "num_tokens": 333263748.0, "step": 8780 }, { "epoch": 1.1170334563032693, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9432241916656494, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8702187538146973, "num_tokens": 333296110.0, "step": 8781 }, { "epoch": 1.1171606665818599, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.634068489074707, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8661739230155945, "num_tokens": 333328372.0, "step": 8782 }, { "epoch": 1.1172878768604504, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020313262939453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7255420684814453, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8780537843704224, "num_tokens": 333367680.0, "step": 8783 }, { "epoch": 1.117415087139041, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.532045602798462, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8706516027450562, "num_tokens": 333406761.0, "step": 8784 }, { "epoch": 1.1175422974176314, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.57322359085083, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8739967346191406, "num_tokens": 333443268.0, "step": 8785 }, { "epoch": 1.1176695076962218, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4582159519195557, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8594326972961426, "num_tokens": 333478770.0, "step": 8786 }, { "epoch": 1.1177967179748123, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4729316234588623, "learning_rate": 1e-05, "loss": 0.3971, "mean_token_accuracy": 0.8863596320152283, "num_tokens": 333514971.0, "step": 8787 }, { "epoch": 1.1179239282534028, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.368227243423462, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8693404197692871, "num_tokens": 333553610.0, "step": 8788 }, { "epoch": 1.1180511385319933, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002040863037109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.415266990661621, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8713441491127014, "num_tokens": 333604283.0, "step": 8789 }, { "epoch": 1.1181783488105839, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3010668754577637, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8806461095809937, "num_tokens": 333641621.0, "step": 8790 }, { "epoch": 1.1183055590891744, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4165072441101074, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.879871129989624, "num_tokens": 333676435.0, "step": 8791 }, { "epoch": 1.118432769367765, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.240091562271118, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8819893598556519, "num_tokens": 333715308.0, "step": 8792 }, { "epoch": 1.1185599796463555, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.452965259552002, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8557677865028381, "num_tokens": 333752086.0, "step": 8793 }, { "epoch": 1.118687189924946, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1707963943481445, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8691619634628296, "num_tokens": 333789726.0, "step": 8794 }, { "epoch": 1.1188144002035365, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5344178676605225, "learning_rate": 1e-05, "loss": 0.4951, "mean_token_accuracy": 0.860389232635498, "num_tokens": 333828518.0, "step": 8795 }, { "epoch": 1.118941610482127, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.092816114425659, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8650724291801453, "num_tokens": 333870520.0, "step": 8796 }, { "epoch": 1.1190688207607176, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6697683334350586, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8736295700073242, "num_tokens": 333914676.0, "step": 8797 }, { "epoch": 1.119196031039308, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0288984775543213, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.876803994178772, "num_tokens": 333953381.0, "step": 8798 }, { "epoch": 1.1193232413178984, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.91972017288208, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8577315211296082, "num_tokens": 333983275.0, "step": 8799 }, { "epoch": 1.119450451596489, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.0748584270477295, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8752523064613342, "num_tokens": 334018530.0, "step": 8800 }, { "epoch": 1.1195776618750795, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.147281169891357, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8817135095596313, "num_tokens": 334056606.0, "step": 8801 }, { "epoch": 1.11970487215367, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.312960624694824, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8739118576049805, "num_tokens": 334098541.0, "step": 8802 }, { "epoch": 1.1198320824322605, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.191599369049072, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8611034750938416, "num_tokens": 334140243.0, "step": 8803 }, { "epoch": 1.119959292710851, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.5850770473480225, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8784101009368896, "num_tokens": 334177126.0, "step": 8804 }, { "epoch": 1.1200865029894416, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.001760482788086, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.878848671913147, "num_tokens": 334217058.0, "step": 8805 }, { "epoch": 1.120213713268032, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6389942169189453, "learning_rate": 1e-05, "loss": 0.4011, "mean_token_accuracy": 0.8892157077789307, "num_tokens": 334255060.0, "step": 8806 }, { "epoch": 1.1203409235466226, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.787113666534424, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8671172857284546, "num_tokens": 334290661.0, "step": 8807 }, { "epoch": 1.1204681338252132, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5742037296295166, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8820269703865051, "num_tokens": 334328661.0, "step": 8808 }, { "epoch": 1.1205953441038037, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6029486656188965, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8719188570976257, "num_tokens": 334367535.0, "step": 8809 }, { "epoch": 1.120722554382394, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4892020225524902, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.873999834060669, "num_tokens": 334407366.0, "step": 8810 }, { "epoch": 1.1208497646609845, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.487154483795166, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8728525638580322, "num_tokens": 334444049.0, "step": 8811 }, { "epoch": 1.120976974939575, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4113569259643555, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8645420074462891, "num_tokens": 334485405.0, "step": 8812 }, { "epoch": 1.1211041852181656, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.405316114425659, "learning_rate": 1e-05, "loss": 0.3667, "mean_token_accuracy": 0.8946278095245361, "num_tokens": 334525209.0, "step": 8813 }, { "epoch": 1.121231395496756, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3320095539093018, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.873559296131134, "num_tokens": 334569156.0, "step": 8814 }, { "epoch": 1.1213586057753466, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3955740928649902, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8778551816940308, "num_tokens": 334602692.0, "step": 8815 }, { "epoch": 1.1214858160539372, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.274442434310913, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.866128146648407, "num_tokens": 334644926.0, "step": 8816 }, { "epoch": 1.1216130263325277, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.389533042907715, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8749898076057434, "num_tokens": 334690116.0, "step": 8817 }, { "epoch": 1.1217402366111182, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2138590812683105, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8848037719726562, "num_tokens": 334728403.0, "step": 8818 }, { "epoch": 1.1218674468897087, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4427859783172607, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.87142014503479, "num_tokens": 334763394.0, "step": 8819 }, { "epoch": 1.1219946571682993, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1591954231262207, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.879894495010376, "num_tokens": 334795029.0, "step": 8820 }, { "epoch": 1.1221218674468898, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5190699100494385, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8683030009269714, "num_tokens": 334833298.0, "step": 8821 }, { "epoch": 1.1222490777254803, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.267692565917969e-05, "grad_norm": 3.0822222232818604, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.8837620615959167, "num_tokens": 334865146.0, "step": 8822 }, { "epoch": 1.1223762880040706, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.675776243209839, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.85673588514328, "num_tokens": 334899392.0, "step": 8823 }, { "epoch": 1.1225034982826612, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0257182121276855, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8575884699821472, "num_tokens": 334936501.0, "step": 8824 }, { "epoch": 1.1226307085612517, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.888493061065674, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8626920580863953, "num_tokens": 334975920.0, "step": 8825 }, { "epoch": 1.1227579188398422, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0572917461395264, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8703570365905762, "num_tokens": 335012104.0, "step": 8826 }, { "epoch": 1.1228851291184327, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.122810363769531, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8559890389442444, "num_tokens": 335051302.0, "step": 8827 }, { "epoch": 1.1230123393970233, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2657694816589355, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.870463490486145, "num_tokens": 335091778.0, "step": 8828 }, { "epoch": 1.1231395496756138, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2309417724609375, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.8564884066581726, "num_tokens": 335120687.0, "step": 8829 }, { "epoch": 1.1232667599542043, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5678138732910156, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8679119944572449, "num_tokens": 335158166.0, "step": 8830 }, { "epoch": 1.1233939702327949, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.052241325378418, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8775110244750977, "num_tokens": 335194674.0, "step": 8831 }, { "epoch": 1.1235211805113854, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6432619094848633, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8762096166610718, "num_tokens": 335229768.0, "step": 8832 }, { "epoch": 1.123648390789976, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00020503997802734375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.840329170227051, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8743188381195068, "num_tokens": 335270500.0, "step": 8833 }, { "epoch": 1.1237756010685664, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.57206130027771, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8813244700431824, "num_tokens": 335309672.0, "step": 8834 }, { "epoch": 1.1239028113471567, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.687859535217285, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8809934854507446, "num_tokens": 335341505.0, "step": 8835 }, { "epoch": 1.1240300216257473, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4874789714813232, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8581464290618896, "num_tokens": 335379013.0, "step": 8836 }, { "epoch": 1.1241572319043378, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5671188831329346, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8631294965744019, "num_tokens": 335415170.0, "step": 8837 }, { "epoch": 1.1242844421829283, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.386868953704834, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8792700171470642, "num_tokens": 335455476.0, "step": 8838 }, { "epoch": 1.1244116524615189, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5142226219177246, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8665997385978699, "num_tokens": 335497185.0, "step": 8839 }, { "epoch": 1.1245388627401094, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.309309720993042, "learning_rate": 1e-05, "loss": 0.5058, "mean_token_accuracy": 0.8546385765075684, "num_tokens": 335531754.0, "step": 8840 }, { "epoch": 1.1246660730187, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5024118423461914, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8781970143318176, "num_tokens": 335570486.0, "step": 8841 }, { "epoch": 1.1247932832972904, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020599365234375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.225196123123169, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8612781167030334, "num_tokens": 335610179.0, "step": 8842 }, { "epoch": 1.124920493575881, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5758183002471924, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8563976287841797, "num_tokens": 335639785.0, "step": 8843 }, { "epoch": 1.1250477038544715, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.145085573196411, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8707513213157654, "num_tokens": 335679012.0, "step": 8844 }, { "epoch": 1.125174914133062, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6708269119262695, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8811383247375488, "num_tokens": 335717495.0, "step": 8845 }, { "epoch": 1.1253021244116526, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.099501848220825, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8710085153579712, "num_tokens": 335751617.0, "step": 8846 }, { "epoch": 1.125429334690243, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.848644733428955, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8627521991729736, "num_tokens": 335782698.0, "step": 8847 }, { "epoch": 1.1255565449688334, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.128278970718384, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8653494715690613, "num_tokens": 335823091.0, "step": 8848 }, { "epoch": 1.125683755247424, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.99377179145813, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8701825141906738, "num_tokens": 335863318.0, "step": 8849 }, { "epoch": 1.1258109655260145, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2863006591796875, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8741176128387451, "num_tokens": 335900467.0, "step": 8850 }, { "epoch": 1.125938175804605, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.003362655639648, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8633198142051697, "num_tokens": 335942274.0, "step": 8851 }, { "epoch": 1.1260653860831955, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.4722042083740234, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8678512573242188, "num_tokens": 335975073.0, "step": 8852 }, { "epoch": 1.126192596361786, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.852250099182129, "learning_rate": 1e-05, "loss": 0.5021, "mean_token_accuracy": 0.8560796976089478, "num_tokens": 336011846.0, "step": 8853 }, { "epoch": 1.1263198066403766, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5214359760284424, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8656195998191833, "num_tokens": 336050969.0, "step": 8854 }, { "epoch": 1.126447016918967, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6363255977630615, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8858537673950195, "num_tokens": 336087714.0, "step": 8855 }, { "epoch": 1.1265742271975576, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5092618465423584, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8596965074539185, "num_tokens": 336124937.0, "step": 8856 }, { "epoch": 1.1267014374761481, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4354476928710938, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8856133222579956, "num_tokens": 336167666.0, "step": 8857 }, { "epoch": 1.1268286477547387, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.495905637741089, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8836125135421753, "num_tokens": 336202170.0, "step": 8858 }, { "epoch": 1.126955858033329, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3064773082733154, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8799448013305664, "num_tokens": 336235613.0, "step": 8859 }, { "epoch": 1.1270830683119195, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4995384216308594, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8679844737052917, "num_tokens": 336272027.0, "step": 8860 }, { "epoch": 1.12721027859051, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.21266508102417, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.85588139295578, "num_tokens": 336312554.0, "step": 8861 }, { "epoch": 1.1273374888691006, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5305519104003906, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8711177110671997, "num_tokens": 336356719.0, "step": 8862 }, { "epoch": 1.127464699147691, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.135560989379883, "learning_rate": 1e-05, "loss": 0.4157, "mean_token_accuracy": 0.8796554803848267, "num_tokens": 336395242.0, "step": 8863 }, { "epoch": 1.1275919094262816, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.619016408920288, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8672990798950195, "num_tokens": 336431738.0, "step": 8864 }, { "epoch": 1.1277191197048722, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0820517539978027, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8680250644683838, "num_tokens": 336467799.0, "step": 8865 }, { "epoch": 1.1278463299834627, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.75669002532959, "learning_rate": 1e-05, "loss": 0.5199, "mean_token_accuracy": 0.8507858514785767, "num_tokens": 336506628.0, "step": 8866 }, { "epoch": 1.1279735402620532, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.0625832080841064, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8786727786064148, "num_tokens": 336546034.0, "step": 8867 }, { "epoch": 1.1281007505406437, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9406626224517822, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8696359395980835, "num_tokens": 336578561.0, "step": 8868 }, { "epoch": 1.1282279608192343, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.161403179168701, "learning_rate": 1e-05, "loss": 0.3994, "mean_token_accuracy": 0.8836629390716553, "num_tokens": 336610500.0, "step": 8869 }, { "epoch": 1.1283551710978248, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.054089546203613, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.879012405872345, "num_tokens": 336653889.0, "step": 8870 }, { "epoch": 1.1284823813764153, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3401939868927, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8672596216201782, "num_tokens": 336690605.0, "step": 8871 }, { "epoch": 1.1286095916550056, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.043198108673096, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8657156229019165, "num_tokens": 336729159.0, "step": 8872 }, { "epoch": 1.1287368019335962, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5004525184631348, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.877239465713501, "num_tokens": 336762376.0, "step": 8873 }, { "epoch": 1.1288640122121867, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8828001022338867, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.876337468624115, "num_tokens": 336797060.0, "step": 8874 }, { "epoch": 1.1289912224907772, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5352070331573486, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8522343635559082, "num_tokens": 336838686.0, "step": 8875 }, { "epoch": 1.1291184327693677, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.689018487930298, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.854340672492981, "num_tokens": 336887535.0, "step": 8876 }, { "epoch": 1.1292456430479583, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.483259439468384, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8769713640213013, "num_tokens": 336920167.0, "step": 8877 }, { "epoch": 1.1293728533265488, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.557530403137207, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8727587461471558, "num_tokens": 336953925.0, "step": 8878 }, { "epoch": 1.1295000636051393, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.416323661804199, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8749992847442627, "num_tokens": 336992407.0, "step": 8879 }, { "epoch": 1.1296272738837299, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.47088623046875, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.857858419418335, "num_tokens": 337029495.0, "step": 8880 }, { "epoch": 1.1297544841623204, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.345351457595825, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8684653639793396, "num_tokens": 337067533.0, "step": 8881 }, { "epoch": 1.129881694440911, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4359683990478516, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8576246500015259, "num_tokens": 337105112.0, "step": 8882 }, { "epoch": 1.1300089047195012, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2945356369018555, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.860099196434021, "num_tokens": 337140807.0, "step": 8883 }, { "epoch": 1.1301361149980917, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020694732666015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4152653217315674, "learning_rate": 1e-05, "loss": 0.3878, "mean_token_accuracy": 0.887742280960083, "num_tokens": 337182644.0, "step": 8884 }, { "epoch": 1.1302633252766823, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2254626750946045, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.872709333896637, "num_tokens": 337226868.0, "step": 8885 }, { "epoch": 1.1303905355552728, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4786293506622314, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8593090772628784, "num_tokens": 337262118.0, "step": 8886 }, { "epoch": 1.1305177458338633, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1630730628967285, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8524091243743896, "num_tokens": 337299575.0, "step": 8887 }, { "epoch": 1.1306449561124539, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002079010009765625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.556283473968506, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8740081787109375, "num_tokens": 337335178.0, "step": 8888 }, { "epoch": 1.1307721663910444, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.125410556793213, "learning_rate": 1e-05, "loss": 0.5141, "mean_token_accuracy": 0.8550724387168884, "num_tokens": 337368982.0, "step": 8889 }, { "epoch": 1.130899376669635, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6632654666900635, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8656129240989685, "num_tokens": 337403708.0, "step": 8890 }, { "epoch": 1.1310265869482254, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.055006980895996, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8743705749511719, "num_tokens": 337441265.0, "step": 8891 }, { "epoch": 1.131153797226816, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.845874786376953, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8729000687599182, "num_tokens": 337475600.0, "step": 8892 }, { "epoch": 1.1312810075054065, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.0805044174194336, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8820183277130127, "num_tokens": 337513051.0, "step": 8893 }, { "epoch": 1.131408217783997, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.049622058868408, "learning_rate": 1e-05, "loss": 0.5119, "mean_token_accuracy": 0.8590776324272156, "num_tokens": 337552187.0, "step": 8894 }, { "epoch": 1.1315354280625876, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.248582124710083, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.884365439414978, "num_tokens": 337586877.0, "step": 8895 }, { "epoch": 1.131662638341178, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.094491004943848, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8710033297538757, "num_tokens": 337627979.0, "step": 8896 }, { "epoch": 1.1317898486197684, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.463824510574341, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8718903064727783, "num_tokens": 337670064.0, "step": 8897 }, { "epoch": 1.131917058898359, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9823970794677734, "learning_rate": 1e-05, "loss": 0.5596, "mean_token_accuracy": 0.8430416584014893, "num_tokens": 337709670.0, "step": 8898 }, { "epoch": 1.1320442691769494, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.56539249420166, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8679046630859375, "num_tokens": 337747507.0, "step": 8899 }, { "epoch": 1.13217147945554, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7571003437042236, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8665355443954468, "num_tokens": 337789542.0, "step": 8900 }, { "epoch": 1.1322986897341305, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5210652351379395, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8836346864700317, "num_tokens": 337832100.0, "step": 8901 }, { "epoch": 1.132425900012721, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5852410793304443, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8712759017944336, "num_tokens": 337869395.0, "step": 8902 }, { "epoch": 1.1325531102913116, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4761531352996826, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8546940088272095, "num_tokens": 337909644.0, "step": 8903 }, { "epoch": 1.132680320569902, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4538543224334717, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8564459681510925, "num_tokens": 337947255.0, "step": 8904 }, { "epoch": 1.1328075308484926, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.423168182373047, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.882893443107605, "num_tokens": 337986377.0, "step": 8905 }, { "epoch": 1.1329347411270831, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3579354286193848, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8765585422515869, "num_tokens": 338025234.0, "step": 8906 }, { "epoch": 1.1330619514056737, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.377985715866089, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8590383529663086, "num_tokens": 338069070.0, "step": 8907 }, { "epoch": 1.133189161684264, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2993171215057373, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8747583031654358, "num_tokens": 338111529.0, "step": 8908 }, { "epoch": 1.1333163719628545, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.357384204864502, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8597640991210938, "num_tokens": 338155881.0, "step": 8909 }, { "epoch": 1.133443582241445, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2744672298431396, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8633090853691101, "num_tokens": 338192570.0, "step": 8910 }, { "epoch": 1.1335707925200356, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3557965755462646, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8666386604309082, "num_tokens": 338225088.0, "step": 8911 }, { "epoch": 1.133698002798626, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2455692291259766, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8571010828018188, "num_tokens": 338267301.0, "step": 8912 }, { "epoch": 1.1338252130772166, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.350874423980713, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.876006543636322, "num_tokens": 338305130.0, "step": 8913 }, { "epoch": 1.1339524233558071, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.213876247406006, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8530190587043762, "num_tokens": 338347556.0, "step": 8914 }, { "epoch": 1.1340796336343977, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3837883472442627, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8706912994384766, "num_tokens": 338384994.0, "step": 8915 }, { "epoch": 1.1342068439129882, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1692054271698, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8601235151290894, "num_tokens": 338427311.0, "step": 8916 }, { "epoch": 1.1343340541915787, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.43353271484375, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8779768347740173, "num_tokens": 338464091.0, "step": 8917 }, { "epoch": 1.1344612644701693, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0962040424346924, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8681288957595825, "num_tokens": 338507961.0, "step": 8918 }, { "epoch": 1.1345884747487598, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5617690086364746, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8707002401351929, "num_tokens": 338544028.0, "step": 8919 }, { "epoch": 1.1347156850273503, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.017141580581665, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.855499267578125, "num_tokens": 338587827.0, "step": 8920 }, { "epoch": 1.1348428953059406, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7774226665496826, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8720743656158447, "num_tokens": 338627729.0, "step": 8921 }, { "epoch": 1.1349701055845312, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 2.9656107425689697, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.8824468851089478, "num_tokens": 338663805.0, "step": 8922 }, { "epoch": 1.1350973158631217, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0922770500183105, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8691461682319641, "num_tokens": 338702204.0, "step": 8923 }, { "epoch": 1.1352245261417122, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.0937881469726562, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8737106323242188, "num_tokens": 338748769.0, "step": 8924 }, { "epoch": 1.1353517364203027, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.380658149719238, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8779457807540894, "num_tokens": 338793617.0, "step": 8925 }, { "epoch": 1.1354789466988933, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.5438778400421143, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8743222951889038, "num_tokens": 338827893.0, "step": 8926 }, { "epoch": 1.1356061569774838, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.291074275970459, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8740699291229248, "num_tokens": 338866036.0, "step": 8927 }, { "epoch": 1.1357333672560743, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.7886366844177246, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8627861738204956, "num_tokens": 338900595.0, "step": 8928 }, { "epoch": 1.1358605775346649, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.031229496002197, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8860747814178467, "num_tokens": 338930268.0, "step": 8929 }, { "epoch": 1.1359877878132554, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.709872245788574, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8698903322219849, "num_tokens": 338974633.0, "step": 8930 }, { "epoch": 1.136114998091846, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8425607681274414, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8590582609176636, "num_tokens": 339008884.0, "step": 8931 }, { "epoch": 1.1362422083704362, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5919713973999023, "learning_rate": 1e-05, "loss": 0.5315, "mean_token_accuracy": 0.8444304466247559, "num_tokens": 339043923.0, "step": 8932 }, { "epoch": 1.1363694186490267, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.706491470336914, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8748160004615784, "num_tokens": 339080545.0, "step": 8933 }, { "epoch": 1.1364966289276173, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.453461170196533, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.869509220123291, "num_tokens": 339114243.0, "step": 8934 }, { "epoch": 1.1366238392062078, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6577446460723877, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8769884705543518, "num_tokens": 339147609.0, "step": 8935 }, { "epoch": 1.1367510494847983, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3213913440704346, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8778719902038574, "num_tokens": 339186295.0, "step": 8936 }, { "epoch": 1.1368782597633889, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6934726238250732, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8742012977600098, "num_tokens": 339217572.0, "step": 8937 }, { "epoch": 1.1370054700419794, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.22723388671875, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8740600943565369, "num_tokens": 339257305.0, "step": 8938 }, { "epoch": 1.13713268032057, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.782493829727173, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8644782304763794, "num_tokens": 339295117.0, "step": 8939 }, { "epoch": 1.1372598905991604, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1991214752197266, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8718752861022949, "num_tokens": 339330543.0, "step": 8940 }, { "epoch": 1.137387100877751, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.904249429702759, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8711779713630676, "num_tokens": 339375040.0, "step": 8941 }, { "epoch": 1.1375143111563415, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2480926513671875, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8643823862075806, "num_tokens": 339410862.0, "step": 8942 }, { "epoch": 1.137641521434932, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.008326053619385, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8758295178413391, "num_tokens": 339441929.0, "step": 8943 }, { "epoch": 1.1377687317135226, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3945701122283936, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.871588945388794, "num_tokens": 339486156.0, "step": 8944 }, { "epoch": 1.137895941992113, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.942492723464966, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8576601147651672, "num_tokens": 339531676.0, "step": 8945 }, { "epoch": 1.1380231522707034, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.516655683517456, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.8594449758529663, "num_tokens": 339571003.0, "step": 8946 }, { "epoch": 1.138150362549294, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.758329391479492, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8637336492538452, "num_tokens": 339604389.0, "step": 8947 }, { "epoch": 1.1382775728278844, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5260939598083496, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8572919368743896, "num_tokens": 339649267.0, "step": 8948 }, { "epoch": 1.138404783106475, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.557270050048828, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8603111505508423, "num_tokens": 339685564.0, "step": 8949 }, { "epoch": 1.1385319933850655, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5165960788726807, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8839032053947449, "num_tokens": 339716312.0, "step": 8950 }, { "epoch": 1.138659203663656, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3701844215393066, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8722327351570129, "num_tokens": 339760666.0, "step": 8951 }, { "epoch": 1.1387864139422466, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4970321655273438, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8787516355514526, "num_tokens": 339799432.0, "step": 8952 }, { "epoch": 1.138913624220837, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2520363330841064, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.865539014339447, "num_tokens": 339839787.0, "step": 8953 }, { "epoch": 1.1390408344994276, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5135602951049805, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8709985613822937, "num_tokens": 339876193.0, "step": 8954 }, { "epoch": 1.1391680447780181, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1633365154266357, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8689427971839905, "num_tokens": 339919791.0, "step": 8955 }, { "epoch": 1.1392952550566087, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5839669704437256, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8607996702194214, "num_tokens": 339954432.0, "step": 8956 }, { "epoch": 1.139422465335199, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.090440511703491, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.874186098575592, "num_tokens": 339996327.0, "step": 8957 }, { "epoch": 1.1395496756137895, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7143282890319824, "learning_rate": 1e-05, "loss": 0.3731, "mean_token_accuracy": 0.8988029956817627, "num_tokens": 340031604.0, "step": 8958 }, { "epoch": 1.13967688589238, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0472137928009033, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.87044358253479, "num_tokens": 340072433.0, "step": 8959 }, { "epoch": 1.1398040961709706, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.9094667434692383, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8699865937232971, "num_tokens": 340114306.0, "step": 8960 }, { "epoch": 1.139931306449561, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1258835792541504, "learning_rate": 1e-05, "loss": 0.5109, "mean_token_accuracy": 0.8473225235939026, "num_tokens": 340152826.0, "step": 8961 }, { "epoch": 1.1400585167281516, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.076900005340576, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8624334931373596, "num_tokens": 340191110.0, "step": 8962 }, { "epoch": 1.1401857270067421, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3099238872528076, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.868876576423645, "num_tokens": 340230606.0, "step": 8963 }, { "epoch": 1.1403129372853327, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.095713138580322, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8619692325592041, "num_tokens": 340274831.0, "step": 8964 }, { "epoch": 1.1404401475639232, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.513892412185669, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8712599277496338, "num_tokens": 340308245.0, "step": 8965 }, { "epoch": 1.1405673578425137, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9414188861846924, "learning_rate": 1e-05, "loss": 0.4042, "mean_token_accuracy": 0.890128493309021, "num_tokens": 340343091.0, "step": 8966 }, { "epoch": 1.1406945681211043, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5663082599639893, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8732465505599976, "num_tokens": 340382968.0, "step": 8967 }, { "epoch": 1.1408217783996948, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.737732410430908, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8782073855400085, "num_tokens": 340421740.0, "step": 8968 }, { "epoch": 1.1409489886782853, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5268595218658447, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8679239749908447, "num_tokens": 340460156.0, "step": 8969 }, { "epoch": 1.1410761989568756, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.579838991165161, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8698540925979614, "num_tokens": 340496106.0, "step": 8970 }, { "epoch": 1.1412034092354661, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4673471450805664, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8730466365814209, "num_tokens": 340530338.0, "step": 8971 }, { "epoch": 1.1413306195140567, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.468404769897461, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8718152046203613, "num_tokens": 340565590.0, "step": 8972 }, { "epoch": 1.1414578297926472, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3936328887939453, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8734910488128662, "num_tokens": 340609651.0, "step": 8973 }, { "epoch": 1.1415850400712377, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3985555171966553, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8707467317581177, "num_tokens": 340650119.0, "step": 8974 }, { "epoch": 1.1417122503498283, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3395743370056152, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8708760738372803, "num_tokens": 340686734.0, "step": 8975 }, { "epoch": 1.1418394606284188, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.376603126525879, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8779313564300537, "num_tokens": 340722405.0, "step": 8976 }, { "epoch": 1.1419666709070093, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2785189151763916, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8788439631462097, "num_tokens": 340763402.0, "step": 8977 }, { "epoch": 1.1420938811855998, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.382979154586792, "learning_rate": 1e-05, "loss": 0.366, "mean_token_accuracy": 0.8961614370346069, "num_tokens": 340799410.0, "step": 8978 }, { "epoch": 1.1422210914641904, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2400472164154053, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8610943555831909, "num_tokens": 340832970.0, "step": 8979 }, { "epoch": 1.142348301742781, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4144198894500732, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8770893812179565, "num_tokens": 340872500.0, "step": 8980 }, { "epoch": 1.1424755120213712, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.168769359588623, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8659267425537109, "num_tokens": 340913784.0, "step": 8981 }, { "epoch": 1.1426027222999617, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4888319969177246, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.881615161895752, "num_tokens": 340945660.0, "step": 8982 }, { "epoch": 1.1427299325785523, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1081454753875732, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8687219619750977, "num_tokens": 340979528.0, "step": 8983 }, { "epoch": 1.1428571428571428, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.618983268737793, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8684374690055847, "num_tokens": 341016256.0, "step": 8984 }, { "epoch": 1.1429843531357333, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0357275009155273, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8601090312004089, "num_tokens": 341056339.0, "step": 8985 }, { "epoch": 1.1431115634143239, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8222055435180664, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.876947820186615, "num_tokens": 341093060.0, "step": 8986 }, { "epoch": 1.1432387736929144, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00020885467529296875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.032433271408081, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8635550737380981, "num_tokens": 341128634.0, "step": 8987 }, { "epoch": 1.143365983971505, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.1111040115356445, "learning_rate": 1e-05, "loss": 0.5106, "mean_token_accuracy": 0.8551920056343079, "num_tokens": 341160649.0, "step": 8988 }, { "epoch": 1.1434931942500954, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2124545574188232, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8700230121612549, "num_tokens": 341196660.0, "step": 8989 }, { "epoch": 1.143620404528686, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.235318183898926, "learning_rate": 1e-05, "loss": 0.5002, "mean_token_accuracy": 0.8594244122505188, "num_tokens": 341236704.0, "step": 8990 }, { "epoch": 1.1437476148072765, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.5172784328460693, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.86651611328125, "num_tokens": 341272874.0, "step": 8991 }, { "epoch": 1.143874825085867, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.119711875915527, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8639562129974365, "num_tokens": 341310209.0, "step": 8992 }, { "epoch": 1.1440020353644575, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6585118770599365, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8822242021560669, "num_tokens": 341341717.0, "step": 8993 }, { "epoch": 1.144129245643048, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.903012990951538, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8868796825408936, "num_tokens": 341374298.0, "step": 8994 }, { "epoch": 1.1442564559216384, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6036171913146973, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8691734075546265, "num_tokens": 341416214.0, "step": 8995 }, { "epoch": 1.144383666200229, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7194631099700928, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8714346885681152, "num_tokens": 341457419.0, "step": 8996 }, { "epoch": 1.1445108764788194, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.512593984603882, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8768790364265442, "num_tokens": 341493251.0, "step": 8997 }, { "epoch": 1.14463808675741, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.608783006668091, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8547053337097168, "num_tokens": 341531501.0, "step": 8998 }, { "epoch": 1.1447652970360005, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.3951363563537598, "learning_rate": 1e-05, "loss": 0.3932, "mean_token_accuracy": 0.8853508234024048, "num_tokens": 341572114.0, "step": 8999 }, { "epoch": 1.144892507314591, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5542938709259033, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8649435043334961, "num_tokens": 341617446.0, "step": 9000 }, { "epoch": 1.1450197175931816, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.311089038848877, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8652654886245728, "num_tokens": 341653636.0, "step": 9001 }, { "epoch": 1.145146927871772, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5699729919433594, "learning_rate": 1e-05, "loss": 0.5769, "mean_token_accuracy": 0.8339170217514038, "num_tokens": 341692174.0, "step": 9002 }, { "epoch": 1.1452741381503626, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2198128700256348, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8630604147911072, "num_tokens": 341737215.0, "step": 9003 }, { "epoch": 1.1454013484289531, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.620473623275757, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8658260107040405, "num_tokens": 341778571.0, "step": 9004 }, { "epoch": 1.1455285587075437, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1604490280151367, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8573012948036194, "num_tokens": 341814725.0, "step": 9005 }, { "epoch": 1.145655768986134, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7285048961639404, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.8830876350402832, "num_tokens": 341854982.0, "step": 9006 }, { "epoch": 1.1457829792647245, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1299452781677246, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8783403038978577, "num_tokens": 341888952.0, "step": 9007 }, { "epoch": 1.145910189543315, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8643476963043213, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8773423433303833, "num_tokens": 341927985.0, "step": 9008 }, { "epoch": 1.1460373998219056, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1908936500549316, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8708251118659973, "num_tokens": 341959813.0, "step": 9009 }, { "epoch": 1.146164610100496, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9724061489105225, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8758358955383301, "num_tokens": 341997231.0, "step": 9010 }, { "epoch": 1.1462918203790866, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.322354555130005, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8832166194915771, "num_tokens": 342029277.0, "step": 9011 }, { "epoch": 1.1464190306576771, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9471383094787598, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.879457414150238, "num_tokens": 342069297.0, "step": 9012 }, { "epoch": 1.1465462409362677, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4445323944091797, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8680548667907715, "num_tokens": 342111525.0, "step": 9013 }, { "epoch": 1.1466734512148582, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8055546283721924, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8710760474205017, "num_tokens": 342148445.0, "step": 9014 }, { "epoch": 1.1468006614934487, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5108771324157715, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.869896411895752, "num_tokens": 342186855.0, "step": 9015 }, { "epoch": 1.1469278717720393, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6045093536376953, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8698586225509644, "num_tokens": 342224619.0, "step": 9016 }, { "epoch": 1.1470550820506298, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.487464189529419, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8795004487037659, "num_tokens": 342261775.0, "step": 9017 }, { "epoch": 1.1471822923292203, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.424910545349121, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8758176565170288, "num_tokens": 342296355.0, "step": 9018 }, { "epoch": 1.1473095026078106, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4767906665802, "learning_rate": 1e-05, "loss": 0.5084, "mean_token_accuracy": 0.8517328500747681, "num_tokens": 342337363.0, "step": 9019 }, { "epoch": 1.1474367128864011, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.297227621078491, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8709605932235718, "num_tokens": 342376462.0, "step": 9020 }, { "epoch": 1.1475639231649917, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4797751903533936, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8732824921607971, "num_tokens": 342417341.0, "step": 9021 }, { "epoch": 1.1476911334435822, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.205332040786743, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8791607022285461, "num_tokens": 342455489.0, "step": 9022 }, { "epoch": 1.1478183437221727, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5196402072906494, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8777716755867004, "num_tokens": 342489594.0, "step": 9023 }, { "epoch": 1.1479455540007633, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.149308919906616, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8622329831123352, "num_tokens": 342525473.0, "step": 9024 }, { "epoch": 1.1480727642793538, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.609140396118164, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8822811841964722, "num_tokens": 342557472.0, "step": 9025 }, { "epoch": 1.1481999745579443, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0853025913238525, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8760294318199158, "num_tokens": 342593386.0, "step": 9026 }, { "epoch": 1.1483271848365348, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7557532787323, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8659875392913818, "num_tokens": 342631370.0, "step": 9027 }, { "epoch": 1.1484543951151254, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.055210828781128, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8702501058578491, "num_tokens": 342675570.0, "step": 9028 }, { "epoch": 1.148581605393716, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.952122449874878, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.86708664894104, "num_tokens": 342717519.0, "step": 9029 }, { "epoch": 1.1487088156723062, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1352591514587402, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8836990594863892, "num_tokens": 342754104.0, "step": 9030 }, { "epoch": 1.1488360259508967, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.111861705780029, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8593992590904236, "num_tokens": 342796059.0, "step": 9031 }, { "epoch": 1.1489632362294873, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.36513090133667, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8644474744796753, "num_tokens": 342833731.0, "step": 9032 }, { "epoch": 1.1490904465080778, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.085761070251465, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8791033029556274, "num_tokens": 342871918.0, "step": 9033 }, { "epoch": 1.1492176567866683, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5454463958740234, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8556104302406311, "num_tokens": 342903129.0, "step": 9034 }, { "epoch": 1.1493448670652588, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.911116123199463, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8604260683059692, "num_tokens": 342941503.0, "step": 9035 }, { "epoch": 1.1494720773438494, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5613772869110107, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8649271726608276, "num_tokens": 342977766.0, "step": 9036 }, { "epoch": 1.14959928762244, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7178947925567627, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8732028603553772, "num_tokens": 343019865.0, "step": 9037 }, { "epoch": 1.1497264979010304, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5109026432037354, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8602867126464844, "num_tokens": 343056500.0, "step": 9038 }, { "epoch": 1.149853708179621, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.571464776992798, "learning_rate": 1e-05, "loss": 0.4924, "mean_token_accuracy": 0.8577172756195068, "num_tokens": 343098404.0, "step": 9039 }, { "epoch": 1.1499809184582115, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.428151845932007, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8846220970153809, "num_tokens": 343137360.0, "step": 9040 }, { "epoch": 1.150108128736802, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.479015588760376, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8798618316650391, "num_tokens": 343176038.0, "step": 9041 }, { "epoch": 1.1502353390153925, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3529903888702393, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8729639053344727, "num_tokens": 343217888.0, "step": 9042 }, { "epoch": 1.150362549293983, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4424898624420166, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8741798400878906, "num_tokens": 343257217.0, "step": 9043 }, { "epoch": 1.1504897595725734, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2994558811187744, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8763329982757568, "num_tokens": 343290607.0, "step": 9044 }, { "epoch": 1.150616969851164, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.437978506088257, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8734865188598633, "num_tokens": 343329608.0, "step": 9045 }, { "epoch": 1.1507441801297544, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.23747181892395, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8821887373924255, "num_tokens": 343365459.0, "step": 9046 }, { "epoch": 1.150871390408345, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4650254249572754, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.875934362411499, "num_tokens": 343403909.0, "step": 9047 }, { "epoch": 1.1509986006869355, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1696298122406006, "learning_rate": 1e-05, "loss": 0.3761, "mean_token_accuracy": 0.8921866416931152, "num_tokens": 343438506.0, "step": 9048 }, { "epoch": 1.151125810965526, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5565340518951416, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8668391108512878, "num_tokens": 343475059.0, "step": 9049 }, { "epoch": 1.1512530212441165, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1014156341552734, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8734481334686279, "num_tokens": 343511286.0, "step": 9050 }, { "epoch": 1.151380231522707, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.696441411972046, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8820726871490479, "num_tokens": 343548092.0, "step": 9051 }, { "epoch": 1.1515074418012976, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.050485134124756, "learning_rate": 1e-05, "loss": 0.388, "mean_token_accuracy": 0.8862313032150269, "num_tokens": 343586487.0, "step": 9052 }, { "epoch": 1.1516346520798881, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.899914264678955, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8836869597434998, "num_tokens": 343627747.0, "step": 9053 }, { "epoch": 1.1517618623584787, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1804027557373047, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8588399887084961, "num_tokens": 343672725.0, "step": 9054 }, { "epoch": 1.151889072637069, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.015573501586914, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8681331872940063, "num_tokens": 343714426.0, "step": 9055 }, { "epoch": 1.1520162829156595, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.260226249694824, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8655205965042114, "num_tokens": 343754613.0, "step": 9056 }, { "epoch": 1.15214349319425, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.054430961608887, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8723495006561279, "num_tokens": 343796187.0, "step": 9057 }, { "epoch": 1.1522707034728406, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4451303482055664, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8660260438919067, "num_tokens": 343833323.0, "step": 9058 }, { "epoch": 1.152397913751431, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.942333936691284, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8618530035018921, "num_tokens": 343872642.0, "step": 9059 }, { "epoch": 1.1525251240300216, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.524282932281494, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8690749406814575, "num_tokens": 343909104.0, "step": 9060 }, { "epoch": 1.1526523343086121, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.75734543800354, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8655270338058472, "num_tokens": 343947102.0, "step": 9061 }, { "epoch": 1.1527795445872027, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.4971354007720947, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8782241344451904, "num_tokens": 343985887.0, "step": 9062 }, { "epoch": 1.1529067548657932, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.602686643600464, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.863505482673645, "num_tokens": 344019427.0, "step": 9063 }, { "epoch": 1.1530339651443837, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000209808349609375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.444558620452881, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8629406690597534, "num_tokens": 344061706.0, "step": 9064 }, { "epoch": 1.1531611754229742, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.473428249359131, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8869675397872925, "num_tokens": 344100912.0, "step": 9065 }, { "epoch": 1.1532883857015648, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4079878330230713, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8645398020744324, "num_tokens": 344135645.0, "step": 9066 }, { "epoch": 1.1534155959801553, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3910279273986816, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8718101382255554, "num_tokens": 344176467.0, "step": 9067 }, { "epoch": 1.1535428062587456, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3738291263580322, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8648243546485901, "num_tokens": 344213498.0, "step": 9068 }, { "epoch": 1.1536700165373361, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3365211486816406, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.8874386548995972, "num_tokens": 344250105.0, "step": 9069 }, { "epoch": 1.1537972268159267, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3520402908325195, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8683346509933472, "num_tokens": 344284017.0, "step": 9070 }, { "epoch": 1.1539244370945172, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3060030937194824, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8733773231506348, "num_tokens": 344317729.0, "step": 9071 }, { "epoch": 1.1540516473731077, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.331604480743408, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8535404801368713, "num_tokens": 344356032.0, "step": 9072 }, { "epoch": 1.1541788576516983, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2973198890686035, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8625133037567139, "num_tokens": 344385694.0, "step": 9073 }, { "epoch": 1.1543060679302888, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3170268535614014, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8669285774230957, "num_tokens": 344420435.0, "step": 9074 }, { "epoch": 1.1544332782088793, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2635366916656494, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8800047636032104, "num_tokens": 344461065.0, "step": 9075 }, { "epoch": 1.1545604884874698, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.316863775253296, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8775198459625244, "num_tokens": 344495419.0, "step": 9076 }, { "epoch": 1.1546876987660604, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2489447593688965, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.858323335647583, "num_tokens": 344536304.0, "step": 9077 }, { "epoch": 1.154814909044651, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.328343391418457, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8719675540924072, "num_tokens": 344577118.0, "step": 9078 }, { "epoch": 1.1549421193232412, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.239604949951172, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8617253303527832, "num_tokens": 344609919.0, "step": 9079 }, { "epoch": 1.1550693296018317, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.345576763153076, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8718517422676086, "num_tokens": 344651543.0, "step": 9080 }, { "epoch": 1.1551965398804223, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.193150281906128, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8764647245407104, "num_tokens": 344691477.0, "step": 9081 }, { "epoch": 1.1553237501590128, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.387829542160034, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8698617219924927, "num_tokens": 344731533.0, "step": 9082 }, { "epoch": 1.1554509604376033, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.142296314239502, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.882050096988678, "num_tokens": 344771628.0, "step": 9083 }, { "epoch": 1.1555781707161938, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4798078536987305, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8607009649276733, "num_tokens": 344812079.0, "step": 9084 }, { "epoch": 1.1557053809947844, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0588488578796387, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8760455846786499, "num_tokens": 344856395.0, "step": 9085 }, { "epoch": 1.155832591273375, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.645770788192749, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8787575364112854, "num_tokens": 344894578.0, "step": 9086 }, { "epoch": 1.1559598015519654, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.220008850097656e-05, "grad_norm": 2.9690091609954834, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8566293716430664, "num_tokens": 344937896.0, "step": 9087 }, { "epoch": 1.156087011830556, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.929307460784912, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.884209156036377, "num_tokens": 344972852.0, "step": 9088 }, { "epoch": 1.1562142221091465, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 2.9674227237701416, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8661779165267944, "num_tokens": 345010322.0, "step": 9089 }, { "epoch": 1.156341432387737, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.312371253967285, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8879036903381348, "num_tokens": 345046268.0, "step": 9090 }, { "epoch": 1.1564686426663275, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.295867919921875, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8865841627120972, "num_tokens": 345078601.0, "step": 9091 }, { "epoch": 1.156595852944918, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.474553108215332, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8707427382469177, "num_tokens": 345121863.0, "step": 9092 }, { "epoch": 1.1567230632235084, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.778146982192993, "learning_rate": 1e-05, "loss": 0.3937, "mean_token_accuracy": 0.8930389881134033, "num_tokens": 345153884.0, "step": 9093 }, { "epoch": 1.156850273502099, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.22199821472168, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8776829242706299, "num_tokens": 345188793.0, "step": 9094 }, { "epoch": 1.1569774837806894, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.848970651626587, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8788244724273682, "num_tokens": 345222537.0, "step": 9095 }, { "epoch": 1.15710469405928, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9635562896728516, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8684306144714355, "num_tokens": 345262006.0, "step": 9096 }, { "epoch": 1.1572319043378705, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7127139568328857, "learning_rate": 1e-05, "loss": 0.5056, "mean_token_accuracy": 0.8572102785110474, "num_tokens": 345295388.0, "step": 9097 }, { "epoch": 1.157359114616461, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7958614826202393, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8848031759262085, "num_tokens": 345331144.0, "step": 9098 }, { "epoch": 1.1574863248950515, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5495107173919678, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.853190541267395, "num_tokens": 345373009.0, "step": 9099 }, { "epoch": 1.157613535173642, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.69279146194458, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8830130100250244, "num_tokens": 345411931.0, "step": 9100 }, { "epoch": 1.1577407454522326, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4126315116882324, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8696188926696777, "num_tokens": 345448818.0, "step": 9101 }, { "epoch": 1.1578679557308231, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.675416946411133, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8810582160949707, "num_tokens": 345484867.0, "step": 9102 }, { "epoch": 1.1579951660094137, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2880797386169434, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8726893663406372, "num_tokens": 345525949.0, "step": 9103 }, { "epoch": 1.158122376288004, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7274723052978516, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8852934241294861, "num_tokens": 345563027.0, "step": 9104 }, { "epoch": 1.1582495865665945, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1958823204040527, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8833221197128296, "num_tokens": 345605920.0, "step": 9105 }, { "epoch": 1.158376796845185, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8780782222747803, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8653926849365234, "num_tokens": 345645843.0, "step": 9106 }, { "epoch": 1.1585040071237755, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1986844539642334, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8696095943450928, "num_tokens": 345684207.0, "step": 9107 }, { "epoch": 1.158631217402366, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.039708137512207, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.862636923789978, "num_tokens": 345721628.0, "step": 9108 }, { "epoch": 1.1587584276809566, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3391799926757812, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8774005770683289, "num_tokens": 345761864.0, "step": 9109 }, { "epoch": 1.1588856379595471, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.071254730224609, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8737043142318726, "num_tokens": 345802478.0, "step": 9110 }, { "epoch": 1.1590128482381377, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5163958072662354, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8750348091125488, "num_tokens": 345846598.0, "step": 9111 }, { "epoch": 1.1591400585167282, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9321916103363037, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8591076731681824, "num_tokens": 345885050.0, "step": 9112 }, { "epoch": 1.1592672687953187, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.607924222946167, "learning_rate": 1e-05, "loss": 0.5378, "mean_token_accuracy": 0.8422386050224304, "num_tokens": 345927460.0, "step": 9113 }, { "epoch": 1.1593944790739092, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.704754590988159, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8695914149284363, "num_tokens": 345960855.0, "step": 9114 }, { "epoch": 1.1595216893524998, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.583009719848633, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8778015375137329, "num_tokens": 345997086.0, "step": 9115 }, { "epoch": 1.1596488996310903, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4770491123199463, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8777656555175781, "num_tokens": 346034317.0, "step": 9116 }, { "epoch": 1.1597761099096806, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00021076202392578125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5470964908599854, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8658491373062134, "num_tokens": 346072626.0, "step": 9117 }, { "epoch": 1.1599033201882711, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.342437267303467, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8650636076927185, "num_tokens": 346109792.0, "step": 9118 }, { "epoch": 1.1600305304668617, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.531252384185791, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8797153830528259, "num_tokens": 346147696.0, "step": 9119 }, { "epoch": 1.1601577407454522, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2194643020629883, "learning_rate": 1e-05, "loss": 0.3949, "mean_token_accuracy": 0.883159875869751, "num_tokens": 346183944.0, "step": 9120 }, { "epoch": 1.1602849510240427, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.564030408859253, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8795067071914673, "num_tokens": 346223352.0, "step": 9121 }, { "epoch": 1.1604121613026332, "ewc_loss": 0.06396484375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.1543636322021484, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8688300848007202, "num_tokens": 346255008.0, "step": 9122 }, { "epoch": 1.1605393715812238, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6703479290008545, "learning_rate": 1e-05, "loss": 0.5476, "mean_token_accuracy": 0.8451998233795166, "num_tokens": 346285188.0, "step": 9123 }, { "epoch": 1.1606665818598143, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0819172859191895, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8773761987686157, "num_tokens": 346318691.0, "step": 9124 }, { "epoch": 1.1607937921384048, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.81557035446167, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8847795128822327, "num_tokens": 346353800.0, "step": 9125 }, { "epoch": 1.1609210024169954, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.083024263381958, "learning_rate": 1e-05, "loss": 0.492, "mean_token_accuracy": 0.8542018532752991, "num_tokens": 346396616.0, "step": 9126 }, { "epoch": 1.161048212695586, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.006234169006348, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8716455698013306, "num_tokens": 346438244.0, "step": 9127 }, { "epoch": 1.1611754229741762, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2239246368408203, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8705544471740723, "num_tokens": 346475081.0, "step": 9128 }, { "epoch": 1.1613026332527667, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.103270530700684, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.877458930015564, "num_tokens": 346513435.0, "step": 9129 }, { "epoch": 1.1614298435313573, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4362857341766357, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8653342723846436, "num_tokens": 346553216.0, "step": 9130 }, { "epoch": 1.1615570538099478, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.018490791320801, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8672688007354736, "num_tokens": 346589271.0, "step": 9131 }, { "epoch": 1.1616842640885383, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5498158931732178, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8726222515106201, "num_tokens": 346631806.0, "step": 9132 }, { "epoch": 1.1618114743671288, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8267853260040283, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8842966556549072, "num_tokens": 346670805.0, "step": 9133 }, { "epoch": 1.1619386846457194, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.540975570678711, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8828139305114746, "num_tokens": 346710506.0, "step": 9134 }, { "epoch": 1.16206589492431, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.648838996887207, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8623843193054199, "num_tokens": 346753410.0, "step": 9135 }, { "epoch": 1.1621931052029004, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4873509407043457, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8590652346611023, "num_tokens": 346795104.0, "step": 9136 }, { "epoch": 1.162320315481491, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.511009693145752, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.855860710144043, "num_tokens": 346836888.0, "step": 9137 }, { "epoch": 1.1624475257600815, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4221391677856445, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8684777021408081, "num_tokens": 346874449.0, "step": 9138 }, { "epoch": 1.162574736038672, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4230809211730957, "learning_rate": 1e-05, "loss": 0.3642, "mean_token_accuracy": 0.8940512537956238, "num_tokens": 346906706.0, "step": 9139 }, { "epoch": 1.1627019463172625, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.359332323074341, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8711490631103516, "num_tokens": 346947250.0, "step": 9140 }, { "epoch": 1.162829156595853, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.395580768585205, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8687359690666199, "num_tokens": 346984159.0, "step": 9141 }, { "epoch": 1.1629563668744434, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.325962543487549, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8499711751937866, "num_tokens": 347026449.0, "step": 9142 }, { "epoch": 1.163083577153034, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3695836067199707, "learning_rate": 1e-05, "loss": 0.3978, "mean_token_accuracy": 0.884415864944458, "num_tokens": 347057903.0, "step": 9143 }, { "epoch": 1.1632107874316244, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2717812061309814, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8711399435997009, "num_tokens": 347095805.0, "step": 9144 }, { "epoch": 1.163337997710215, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3784878253936768, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8716292977333069, "num_tokens": 347133357.0, "step": 9145 }, { "epoch": 1.1634652079888055, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.231684446334839, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.860745906829834, "num_tokens": 347169679.0, "step": 9146 }, { "epoch": 1.163592418267396, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4221270084381104, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8642857074737549, "num_tokens": 347207990.0, "step": 9147 }, { "epoch": 1.1637196285459865, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.17445707321167, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8823856115341187, "num_tokens": 347244427.0, "step": 9148 }, { "epoch": 1.163846838824577, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.5054433345794678, "learning_rate": 1e-05, "loss": 0.3799, "mean_token_accuracy": 0.8923957347869873, "num_tokens": 347276303.0, "step": 9149 }, { "epoch": 1.1639740491031676, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1005468368530273, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8758325576782227, "num_tokens": 347318682.0, "step": 9150 }, { "epoch": 1.1641012593817581, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6373159885406494, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8609548807144165, "num_tokens": 347359355.0, "step": 9151 }, { "epoch": 1.1642284696603487, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0255355834960938, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8671858310699463, "num_tokens": 347397617.0, "step": 9152 }, { "epoch": 1.164355679938939, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8661036491394043, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8517997860908508, "num_tokens": 347433067.0, "step": 9153 }, { "epoch": 1.1644828902175295, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.031564712524414, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.8835476636886597, "num_tokens": 347467091.0, "step": 9154 }, { "epoch": 1.16461010049612, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.128438472747803, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8779776096343994, "num_tokens": 347503921.0, "step": 9155 }, { "epoch": 1.1647373107747105, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.226353406906128, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.873528778553009, "num_tokens": 347547061.0, "step": 9156 }, { "epoch": 1.164864521053301, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2765913009643555, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8657923936843872, "num_tokens": 347583845.0, "step": 9157 }, { "epoch": 1.1649917313318916, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.5662100315093994, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8702346682548523, "num_tokens": 347624146.0, "step": 9158 }, { "epoch": 1.1651189416104821, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.116759777069092, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8696352243423462, "num_tokens": 347666171.0, "step": 9159 }, { "epoch": 1.1652461518890727, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.681896448135376, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8692036867141724, "num_tokens": 347706960.0, "step": 9160 }, { "epoch": 1.1653733621676632, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9016833305358887, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8687372207641602, "num_tokens": 347744830.0, "step": 9161 }, { "epoch": 1.1655005724462537, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6065807342529297, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8767900466918945, "num_tokens": 347791033.0, "step": 9162 }, { "epoch": 1.1656277827248442, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7282612323760986, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8677370548248291, "num_tokens": 347832579.0, "step": 9163 }, { "epoch": 1.1657549930034348, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5169594287872314, "learning_rate": 1e-05, "loss": 0.3881, "mean_token_accuracy": 0.8883184790611267, "num_tokens": 347863404.0, "step": 9164 }, { "epoch": 1.1658822032820253, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.613593339920044, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8656328916549683, "num_tokens": 347900882.0, "step": 9165 }, { "epoch": 1.1660094135606156, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4157462120056152, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8636665344238281, "num_tokens": 347943105.0, "step": 9166 }, { "epoch": 1.1661366238392061, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5693793296813965, "learning_rate": 1e-05, "loss": 0.5502, "mean_token_accuracy": 0.8358493447303772, "num_tokens": 347978783.0, "step": 9167 }, { "epoch": 1.1662638341177967, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.315821647644043, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8802019357681274, "num_tokens": 348016169.0, "step": 9168 }, { "epoch": 1.1663910443963872, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5543932914733887, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8639297485351562, "num_tokens": 348052431.0, "step": 9169 }, { "epoch": 1.1665182546749777, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2267282009124756, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8671320080757141, "num_tokens": 348091412.0, "step": 9170 }, { "epoch": 1.1666454649535682, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.612558603286743, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8667595982551575, "num_tokens": 348128370.0, "step": 9171 }, { "epoch": 1.1667726752321588, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.157555103302002, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8638724088668823, "num_tokens": 348167076.0, "step": 9172 }, { "epoch": 1.1668998855107493, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7302229404449463, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8759713172912598, "num_tokens": 348206926.0, "step": 9173 }, { "epoch": 1.1670270957893398, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.13083553314209, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8645342588424683, "num_tokens": 348247209.0, "step": 9174 }, { "epoch": 1.1671543060679304, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.878065586090088, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8676761984825134, "num_tokens": 348286613.0, "step": 9175 }, { "epoch": 1.1672815163465209, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.183985471725464, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8672308921813965, "num_tokens": 348322246.0, "step": 9176 }, { "epoch": 1.1674087266251112, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.983232021331787, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8771811723709106, "num_tokens": 348359435.0, "step": 9177 }, { "epoch": 1.1675359369037017, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3312275409698486, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8883835673332214, "num_tokens": 348397330.0, "step": 9178 }, { "epoch": 1.1676631471822922, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9649951457977295, "learning_rate": 1e-05, "loss": 0.5032, "mean_token_accuracy": 0.8585528135299683, "num_tokens": 348435413.0, "step": 9179 }, { "epoch": 1.1677903574608828, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.4755425453186035, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8580679893493652, "num_tokens": 348474963.0, "step": 9180 }, { "epoch": 1.1679175677394733, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.806004524230957, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8684887886047363, "num_tokens": 348511306.0, "step": 9181 }, { "epoch": 1.1680447780180638, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5043301582336426, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8780466914176941, "num_tokens": 348552257.0, "step": 9182 }, { "epoch": 1.1681719882966544, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6104536056518555, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8676905035972595, "num_tokens": 348588143.0, "step": 9183 }, { "epoch": 1.168299198575245, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4884228706359863, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8698887228965759, "num_tokens": 348633525.0, "step": 9184 }, { "epoch": 1.1684264088538354, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4353208541870117, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8676061630249023, "num_tokens": 348676024.0, "step": 9185 }, { "epoch": 1.168553619132426, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4764444828033447, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8796635866165161, "num_tokens": 348715704.0, "step": 9186 }, { "epoch": 1.1686808294110165, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.315566062927246, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8648315668106079, "num_tokens": 348753160.0, "step": 9187 }, { "epoch": 1.168808039689607, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4642722606658936, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8728867769241333, "num_tokens": 348796423.0, "step": 9188 }, { "epoch": 1.1689352499681975, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2236790657043457, "learning_rate": 1e-05, "loss": 0.3836, "mean_token_accuracy": 0.8899726867675781, "num_tokens": 348832813.0, "step": 9189 }, { "epoch": 1.169062460246788, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.496333599090576, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8706854581832886, "num_tokens": 348868273.0, "step": 9190 }, { "epoch": 1.1691896705253784, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1468989849090576, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8781440258026123, "num_tokens": 348908008.0, "step": 9191 }, { "epoch": 1.169316880803969, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.57781720161438, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8713288307189941, "num_tokens": 348947701.0, "step": 9192 }, { "epoch": 1.1694440910825594, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.0802505016326904, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8692742586135864, "num_tokens": 348992221.0, "step": 9193 }, { "epoch": 1.16957130136115, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7256717681884766, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8806918859481812, "num_tokens": 349031982.0, "step": 9194 }, { "epoch": 1.1696985116397405, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0462138652801514, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8825472593307495, "num_tokens": 349070905.0, "step": 9195 }, { "epoch": 1.169825721918331, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9230616092681885, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8714605569839478, "num_tokens": 349114736.0, "step": 9196 }, { "epoch": 1.1699529321969215, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1221184730529785, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8640971779823303, "num_tokens": 349153760.0, "step": 9197 }, { "epoch": 1.170080142475512, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.097805976867676, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.885610818862915, "num_tokens": 349192464.0, "step": 9198 }, { "epoch": 1.1702073527541026, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.316585063934326, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8696889877319336, "num_tokens": 349234444.0, "step": 9199 }, { "epoch": 1.1703345630326931, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.124707221984863, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8576112985610962, "num_tokens": 349276777.0, "step": 9200 }, { "epoch": 1.1704617733112836, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5204076766967773, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8808366060256958, "num_tokens": 349317158.0, "step": 9201 }, { "epoch": 1.170588983589874, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9732553958892822, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8652093410491943, "num_tokens": 349357043.0, "step": 9202 }, { "epoch": 1.1707161938684645, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5906643867492676, "learning_rate": 1e-05, "loss": 0.5104, "mean_token_accuracy": 0.8530281782150269, "num_tokens": 349395843.0, "step": 9203 }, { "epoch": 1.170843404147055, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.770784854888916, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8645609617233276, "num_tokens": 349429410.0, "step": 9204 }, { "epoch": 1.1709706144256455, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5549161434173584, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8551552295684814, "num_tokens": 349466507.0, "step": 9205 }, { "epoch": 1.171097824704236, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5853168964385986, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8840019702911377, "num_tokens": 349502918.0, "step": 9206 }, { "epoch": 1.1712250349828266, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4756252765655518, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8711532354354858, "num_tokens": 349546740.0, "step": 9207 }, { "epoch": 1.1713522452614171, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.460439443588257, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8750850558280945, "num_tokens": 349590479.0, "step": 9208 }, { "epoch": 1.1714794555400077, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.421975612640381, "learning_rate": 1e-05, "loss": 0.3881, "mean_token_accuracy": 0.887649655342102, "num_tokens": 349627726.0, "step": 9209 }, { "epoch": 1.1716066658185982, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.395873546600342, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8717919588088989, "num_tokens": 349663547.0, "step": 9210 }, { "epoch": 1.1717338760971887, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3806214332580566, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8670262694358826, "num_tokens": 349696927.0, "step": 9211 }, { "epoch": 1.1718610863757792, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3455352783203125, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8782293796539307, "num_tokens": 349736027.0, "step": 9212 }, { "epoch": 1.1719882966543698, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3381905555725098, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.866112470626831, "num_tokens": 349778225.0, "step": 9213 }, { "epoch": 1.1721155069329603, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.32669734954834, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8703371286392212, "num_tokens": 349818093.0, "step": 9214 }, { "epoch": 1.1722427172115506, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3215725421905518, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8681790232658386, "num_tokens": 349856433.0, "step": 9215 }, { "epoch": 1.1723699274901411, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3026561737060547, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8651429414749146, "num_tokens": 349898829.0, "step": 9216 }, { "epoch": 1.1724971377687317, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3010499477386475, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8734258413314819, "num_tokens": 349937059.0, "step": 9217 }, { "epoch": 1.1726243480473222, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.30723237991333, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8620455265045166, "num_tokens": 349973433.0, "step": 9218 }, { "epoch": 1.1727515583259127, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2837352752685547, "learning_rate": 1e-05, "loss": 0.5021, "mean_token_accuracy": 0.8522849678993225, "num_tokens": 350012859.0, "step": 9219 }, { "epoch": 1.1728787686045032, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3039071559906006, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8584965467453003, "num_tokens": 350055128.0, "step": 9220 }, { "epoch": 1.1730059788830938, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.5216495990753174, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8565639853477478, "num_tokens": 350090058.0, "step": 9221 }, { "epoch": 1.1731331891616843, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1560564041137695, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8785153031349182, "num_tokens": 350129441.0, "step": 9222 }, { "epoch": 1.1732603994402748, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.507188320159912, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8770979046821594, "num_tokens": 350168866.0, "step": 9223 }, { "epoch": 1.1733876097188654, "ewc_loss": 0.0634765625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.2438507080078125e-05, "grad_norm": 3.031918525695801, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8669428825378418, "num_tokens": 350207498.0, "step": 9224 }, { "epoch": 1.1735148199974559, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7137742042541504, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8600229620933533, "num_tokens": 350248729.0, "step": 9225 }, { "epoch": 1.1736420302760462, "ewc_loss": 0.064453125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 2.9689605236053467, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.871698260307312, "num_tokens": 350279399.0, "step": 9226 }, { "epoch": 1.1737692405546367, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0172858238220215, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.8521997928619385, "num_tokens": 350321135.0, "step": 9227 }, { "epoch": 1.1738964508332272, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002117156982421875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.0185272693634033, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8668440580368042, "num_tokens": 350359348.0, "step": 9228 }, { "epoch": 1.1740236611118178, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.357693195343018, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8741828203201294, "num_tokens": 350401140.0, "step": 9229 }, { "epoch": 1.1741508713904083, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.403294086456299, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8653926849365234, "num_tokens": 350443936.0, "step": 9230 }, { "epoch": 1.1742780816689988, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.413210868835449, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8800966739654541, "num_tokens": 350486110.0, "step": 9231 }, { "epoch": 1.1744052919475894, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.814087390899658, "learning_rate": 1e-05, "loss": 0.5349, "mean_token_accuracy": 0.8511360287666321, "num_tokens": 350521964.0, "step": 9232 }, { "epoch": 1.1745325022261799, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.123816967010498, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.883370578289032, "num_tokens": 350557578.0, "step": 9233 }, { "epoch": 1.1746597125047704, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.8038392066955566, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8695865869522095, "num_tokens": 350595693.0, "step": 9234 }, { "epoch": 1.174786922783361, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8940017223358154, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8781249523162842, "num_tokens": 350633247.0, "step": 9235 }, { "epoch": 1.1749141330619515, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6723427772521973, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8797316551208496, "num_tokens": 350673048.0, "step": 9236 }, { "epoch": 1.175041343340542, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.726461172103882, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8594571352005005, "num_tokens": 350708840.0, "step": 9237 }, { "epoch": 1.1751685536191325, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5250275135040283, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8696354627609253, "num_tokens": 350748875.0, "step": 9238 }, { "epoch": 1.175295763897723, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6252737045288086, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8686721324920654, "num_tokens": 350783200.0, "step": 9239 }, { "epoch": 1.1754229741763134, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.394779920578003, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8712838292121887, "num_tokens": 350819514.0, "step": 9240 }, { "epoch": 1.175550184454904, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5987234115600586, "learning_rate": 1e-05, "loss": 0.3873, "mean_token_accuracy": 0.8910105228424072, "num_tokens": 350854261.0, "step": 9241 }, { "epoch": 1.1756773947334944, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2758893966674805, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8719557523727417, "num_tokens": 350897847.0, "step": 9242 }, { "epoch": 1.175804605012085, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.665330410003662, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8639942407608032, "num_tokens": 350931926.0, "step": 9243 }, { "epoch": 1.1759318152906755, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1710891723632812, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.8819538950920105, "num_tokens": 350972352.0, "step": 9244 }, { "epoch": 1.176059025569266, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.797603130340576, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8690400123596191, "num_tokens": 351012245.0, "step": 9245 }, { "epoch": 1.1761862358478565, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1424314975738525, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8769726753234863, "num_tokens": 351043005.0, "step": 9246 }, { "epoch": 1.176313446126447, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9840309619903564, "learning_rate": 1e-05, "loss": 0.5066, "mean_token_accuracy": 0.8567392826080322, "num_tokens": 351084810.0, "step": 9247 }, { "epoch": 1.1764406564050376, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2251410484313965, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8748952150344849, "num_tokens": 351124243.0, "step": 9248 }, { "epoch": 1.1765678666836281, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.113178253173828, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8797916769981384, "num_tokens": 351163650.0, "step": 9249 }, { "epoch": 1.1766950769622184, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.4667158126831055, "learning_rate": 1e-05, "loss": 0.5539, "mean_token_accuracy": 0.8456137180328369, "num_tokens": 351205912.0, "step": 9250 }, { "epoch": 1.176822287240809, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.016900539398193, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8701527118682861, "num_tokens": 351250515.0, "step": 9251 }, { "epoch": 1.1769494975193995, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6059811115264893, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.868103563785553, "num_tokens": 351284222.0, "step": 9252 }, { "epoch": 1.17707670779799, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7829251289367676, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8603918552398682, "num_tokens": 351327799.0, "step": 9253 }, { "epoch": 1.1772039180765805, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.604161500930786, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8716634511947632, "num_tokens": 351362194.0, "step": 9254 }, { "epoch": 1.177331128355171, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.55293869972229, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8781565427780151, "num_tokens": 351395186.0, "step": 9255 }, { "epoch": 1.1774583386337616, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.56585693359375, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8719605207443237, "num_tokens": 351435484.0, "step": 9256 }, { "epoch": 1.1775855489123521, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3670690059661865, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8622961044311523, "num_tokens": 351474613.0, "step": 9257 }, { "epoch": 1.1777127591909426, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5593883991241455, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8785562515258789, "num_tokens": 351508612.0, "step": 9258 }, { "epoch": 1.1778399694695332, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.238483190536499, "learning_rate": 1e-05, "loss": 0.5044, "mean_token_accuracy": 0.857052206993103, "num_tokens": 351554811.0, "step": 9259 }, { "epoch": 1.1779671797481237, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5970418453216553, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8731779456138611, "num_tokens": 351591540.0, "step": 9260 }, { "epoch": 1.1780943900267142, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.156045913696289, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8712437152862549, "num_tokens": 351630529.0, "step": 9261 }, { "epoch": 1.1782216003053048, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.703468084335327, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8589884042739868, "num_tokens": 351669412.0, "step": 9262 }, { "epoch": 1.1783488105838953, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.092341423034668, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8653792142868042, "num_tokens": 351711825.0, "step": 9263 }, { "epoch": 1.1784760208624856, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.867392063140869, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8764627575874329, "num_tokens": 351747967.0, "step": 9264 }, { "epoch": 1.1786032311410761, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1211163997650146, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8781713247299194, "num_tokens": 351786208.0, "step": 9265 }, { "epoch": 1.1787304414196667, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.042291164398193, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8676596879959106, "num_tokens": 351825711.0, "step": 9266 }, { "epoch": 1.1788576516982572, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2588202953338623, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.88135826587677, "num_tokens": 351866772.0, "step": 9267 }, { "epoch": 1.1789848619768477, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.12233829498291, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8730231523513794, "num_tokens": 351905812.0, "step": 9268 }, { "epoch": 1.1791120722554382, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.475874900817871, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8778989315032959, "num_tokens": 351950295.0, "step": 9269 }, { "epoch": 1.1792392825340288, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.020374298095703, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.8573940992355347, "num_tokens": 351990980.0, "step": 9270 }, { "epoch": 1.1793664928126193, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5814974308013916, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8575735092163086, "num_tokens": 352034483.0, "step": 9271 }, { "epoch": 1.1794937030912098, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.823023796081543, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8803058862686157, "num_tokens": 352069195.0, "step": 9272 }, { "epoch": 1.1796209133698004, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.560009717941284, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8825361728668213, "num_tokens": 352103393.0, "step": 9273 }, { "epoch": 1.1797481236483909, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.637413740158081, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8784446716308594, "num_tokens": 352138032.0, "step": 9274 }, { "epoch": 1.1798753339269812, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4998762607574463, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8609862327575684, "num_tokens": 352177243.0, "step": 9275 }, { "epoch": 1.1800025442055717, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5137887001037598, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8537463545799255, "num_tokens": 352212461.0, "step": 9276 }, { "epoch": 1.1801297544841622, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4247703552246094, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8598930835723877, "num_tokens": 352250001.0, "step": 9277 }, { "epoch": 1.1802569647627528, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4408466815948486, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8561837673187256, "num_tokens": 352288893.0, "step": 9278 }, { "epoch": 1.1803841750413433, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.346205949783325, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8722454309463501, "num_tokens": 352328255.0, "step": 9279 }, { "epoch": 1.1805113853199338, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4246697425842285, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8687354922294617, "num_tokens": 352364871.0, "step": 9280 }, { "epoch": 1.1806385955985244, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2919859886169434, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8805842399597168, "num_tokens": 352398894.0, "step": 9281 }, { "epoch": 1.1807658058771149, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4306347370147705, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8654218316078186, "num_tokens": 352438327.0, "step": 9282 }, { "epoch": 1.1808930161557054, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2319047451019287, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8599998950958252, "num_tokens": 352475732.0, "step": 9283 }, { "epoch": 1.181020226434296, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4994311332702637, "learning_rate": 1e-05, "loss": 0.5066, "mean_token_accuracy": 0.8547121286392212, "num_tokens": 352509405.0, "step": 9284 }, { "epoch": 1.1811474367128865, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1613266468048096, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8620975017547607, "num_tokens": 352547172.0, "step": 9285 }, { "epoch": 1.181274646991477, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000213623046875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5896074771881104, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8687486052513123, "num_tokens": 352585705.0, "step": 9286 }, { "epoch": 1.1814018572700675, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0836541652679443, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8775807619094849, "num_tokens": 352629246.0, "step": 9287 }, { "epoch": 1.181529067548658, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.756413459777832, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8613177537918091, "num_tokens": 352669504.0, "step": 9288 }, { "epoch": 1.1816562778272484, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0551717281341553, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.865736722946167, "num_tokens": 352708981.0, "step": 9289 }, { "epoch": 1.1817834881058389, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.965954303741455, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8664251565933228, "num_tokens": 352751646.0, "step": 9290 }, { "epoch": 1.1819106983844294, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.1522347927093506, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8698118925094604, "num_tokens": 352789959.0, "step": 9291 }, { "epoch": 1.18203790866302, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.132352352142334, "learning_rate": 1e-05, "loss": 0.5079, "mean_token_accuracy": 0.8559540510177612, "num_tokens": 352828725.0, "step": 9292 }, { "epoch": 1.1821651189416105, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3842544555664062, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.8880715370178223, "num_tokens": 352866373.0, "step": 9293 }, { "epoch": 1.182292329220201, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.083516597747803, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8697727918624878, "num_tokens": 352908964.0, "step": 9294 }, { "epoch": 1.1824195394987915, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.570017099380493, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.87834632396698, "num_tokens": 352946292.0, "step": 9295 }, { "epoch": 1.182546749777382, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8890089988708496, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8740335702896118, "num_tokens": 352985875.0, "step": 9296 }, { "epoch": 1.1826739600559726, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.570469856262207, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8653131723403931, "num_tokens": 353032019.0, "step": 9297 }, { "epoch": 1.1828011703345631, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.703723669052124, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8683360815048218, "num_tokens": 353062624.0, "step": 9298 }, { "epoch": 1.1829283806131534, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5064380168914795, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.874046266078949, "num_tokens": 353105401.0, "step": 9299 }, { "epoch": 1.183055590891744, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5515477657318115, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8541884422302246, "num_tokens": 353143180.0, "step": 9300 }, { "epoch": 1.1831828011703345, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.453801393508911, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8711119890213013, "num_tokens": 353177800.0, "step": 9301 }, { "epoch": 1.183310011448925, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4405059814453125, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8700699210166931, "num_tokens": 353221491.0, "step": 9302 }, { "epoch": 1.1834372217275155, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3862428665161133, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8568814992904663, "num_tokens": 353263851.0, "step": 9303 }, { "epoch": 1.183564432006106, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.41172194480896, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8583168983459473, "num_tokens": 353299197.0, "step": 9304 }, { "epoch": 1.1836916422846966, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.330298662185669, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8476669788360596, "num_tokens": 353339008.0, "step": 9305 }, { "epoch": 1.1838188525632871, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.38472318649292, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8699378967285156, "num_tokens": 353378143.0, "step": 9306 }, { "epoch": 1.1839460628418776, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2713584899902344, "learning_rate": 1e-05, "loss": 0.3853, "mean_token_accuracy": 0.8889858722686768, "num_tokens": 353412659.0, "step": 9307 }, { "epoch": 1.1840732731204682, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4042012691497803, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8562062382698059, "num_tokens": 353454459.0, "step": 9308 }, { "epoch": 1.1842004833990587, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2115066051483154, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8630807399749756, "num_tokens": 353501446.0, "step": 9309 }, { "epoch": 1.1843276936776492, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4600417613983154, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.863909125328064, "num_tokens": 353538129.0, "step": 9310 }, { "epoch": 1.1844549039562398, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1387131214141846, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8798054456710815, "num_tokens": 353581488.0, "step": 9311 }, { "epoch": 1.1845821142348303, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5551702976226807, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8717710971832275, "num_tokens": 353622729.0, "step": 9312 }, { "epoch": 1.1847093245134206, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.063774585723877, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8779709339141846, "num_tokens": 353663696.0, "step": 9313 }, { "epoch": 1.1848365347920111, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7250494956970215, "learning_rate": 1e-05, "loss": 0.49, "mean_token_accuracy": 0.8598012328147888, "num_tokens": 353705934.0, "step": 9314 }, { "epoch": 1.1849637450706016, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0219082832336426, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.8872450590133667, "num_tokens": 353741535.0, "step": 9315 }, { "epoch": 1.1850909553491922, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9648962020874023, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8691293001174927, "num_tokens": 353784446.0, "step": 9316 }, { "epoch": 1.1852181656277827, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.0974106788635254, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8682074546813965, "num_tokens": 353821955.0, "step": 9317 }, { "epoch": 1.1853453759063732, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.205144882202148, "learning_rate": 1e-05, "loss": 0.4794, "mean_token_accuracy": 0.8687451481819153, "num_tokens": 353859994.0, "step": 9318 }, { "epoch": 1.1854725861849638, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.353675127029419, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8749760389328003, "num_tokens": 353903301.0, "step": 9319 }, { "epoch": 1.1855997964635543, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00021457672119140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.237765789031982, "learning_rate": 1e-05, "loss": 0.5405, "mean_token_accuracy": 0.8463815450668335, "num_tokens": 353944527.0, "step": 9320 }, { "epoch": 1.1857270067421448, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6411197185516357, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8791128993034363, "num_tokens": 353977402.0, "step": 9321 }, { "epoch": 1.1858542170207353, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.025099754333496, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8774660229682922, "num_tokens": 354016681.0, "step": 9322 }, { "epoch": 1.1859814272993259, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6667819023132324, "learning_rate": 1e-05, "loss": 0.5017, "mean_token_accuracy": 0.8599127531051636, "num_tokens": 354057923.0, "step": 9323 }, { "epoch": 1.1861086375779162, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8035521507263184, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8691639304161072, "num_tokens": 354097787.0, "step": 9324 }, { "epoch": 1.1862358478565067, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.580617904663086, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8762802481651306, "num_tokens": 354136694.0, "step": 9325 }, { "epoch": 1.1863630581350972, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6388516426086426, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8871892690658569, "num_tokens": 354176326.0, "step": 9326 }, { "epoch": 1.1864902684136878, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4907333850860596, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8647928833961487, "num_tokens": 354214350.0, "step": 9327 }, { "epoch": 1.1866174786922783, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.537602663040161, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8692033290863037, "num_tokens": 354252910.0, "step": 9328 }, { "epoch": 1.1867446889708688, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3932535648345947, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8698329329490662, "num_tokens": 354291105.0, "step": 9329 }, { "epoch": 1.1868718992494594, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4863460063934326, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8741146326065063, "num_tokens": 354329767.0, "step": 9330 }, { "epoch": 1.1869991095280499, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3201286792755127, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8789736032485962, "num_tokens": 354369826.0, "step": 9331 }, { "epoch": 1.1871263198066404, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.481271982192993, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8820366263389587, "num_tokens": 354406088.0, "step": 9332 }, { "epoch": 1.187253530085231, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2438104152679443, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.882598876953125, "num_tokens": 354445675.0, "step": 9333 }, { "epoch": 1.1873807403638215, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5259387493133545, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8522613644599915, "num_tokens": 354482733.0, "step": 9334 }, { "epoch": 1.187507950642412, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.176238536834717, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8645027875900269, "num_tokens": 354515483.0, "step": 9335 }, { "epoch": 1.1876351609210025, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6206719875335693, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8781741857528687, "num_tokens": 354551972.0, "step": 9336 }, { "epoch": 1.187762371199593, "ewc_loss": 0.06494140625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0935757160186768, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8704391717910767, "num_tokens": 354592089.0, "step": 9337 }, { "epoch": 1.1878895814781834, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7800168991088867, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8796195983886719, "num_tokens": 354629582.0, "step": 9338 }, { "epoch": 1.1880167917567739, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.072526454925537, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8469704389572144, "num_tokens": 354677114.0, "step": 9339 }, { "epoch": 1.1881440020353644, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9772300720214844, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8662625551223755, "num_tokens": 354723125.0, "step": 9340 }, { "epoch": 1.188271212313955, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1883556842803955, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8599249124526978, "num_tokens": 354761032.0, "step": 9341 }, { "epoch": 1.1883984225925455, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.106020927429199, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8784367442131042, "num_tokens": 354798522.0, "step": 9342 }, { "epoch": 1.188525632871136, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.4164674282073975, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.871055006980896, "num_tokens": 354833752.0, "step": 9343 }, { "epoch": 1.1886528431497265, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.036794662475586, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8743076324462891, "num_tokens": 354869901.0, "step": 9344 }, { "epoch": 1.188780053428317, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.56485652923584, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8707538843154907, "num_tokens": 354902487.0, "step": 9345 }, { "epoch": 1.1889072637069076, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8243706226348877, "learning_rate": 1e-05, "loss": 0.5218, "mean_token_accuracy": 0.8479934930801392, "num_tokens": 354948778.0, "step": 9346 }, { "epoch": 1.189034473985498, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5559120178222656, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8698101043701172, "num_tokens": 354990242.0, "step": 9347 }, { "epoch": 1.1891616842640884, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6228857040405273, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8826372623443604, "num_tokens": 355025462.0, "step": 9348 }, { "epoch": 1.189288894542679, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.517094850540161, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.860887348651886, "num_tokens": 355067488.0, "step": 9349 }, { "epoch": 1.1894161048212695, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4735970497131348, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8656063079833984, "num_tokens": 355101734.0, "step": 9350 }, { "epoch": 1.18954331509986, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.470470666885376, "learning_rate": 1e-05, "loss": 0.4824, "mean_token_accuracy": 0.8591023683547974, "num_tokens": 355140209.0, "step": 9351 }, { "epoch": 1.1896705253784505, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3660640716552734, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8673231601715088, "num_tokens": 355177424.0, "step": 9352 }, { "epoch": 1.189797735657041, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4353373050689697, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8730848431587219, "num_tokens": 355215112.0, "step": 9353 }, { "epoch": 1.1899249459356316, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.294517755508423, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8814486861228943, "num_tokens": 355250402.0, "step": 9354 }, { "epoch": 1.1900521562142221, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4236528873443604, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.873296856880188, "num_tokens": 355289887.0, "step": 9355 }, { "epoch": 1.1901793664928126, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.235624313354492, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8716601133346558, "num_tokens": 355330946.0, "step": 9356 }, { "epoch": 1.1903065767714032, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4468295574188232, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8712944388389587, "num_tokens": 355369029.0, "step": 9357 }, { "epoch": 1.1904337870499937, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2008512020111084, "learning_rate": 1e-05, "loss": 0.5146, "mean_token_accuracy": 0.8497398495674133, "num_tokens": 355406145.0, "step": 9358 }, { "epoch": 1.1905609973285842, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.491335391998291, "learning_rate": 1e-05, "loss": 0.3837, "mean_token_accuracy": 0.889983594417572, "num_tokens": 355445702.0, "step": 9359 }, { "epoch": 1.1906882076071748, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.128761053085327, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8662768006324768, "num_tokens": 355485592.0, "step": 9360 }, { "epoch": 1.1908154178857653, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.595402240753174, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.88526850938797, "num_tokens": 355523208.0, "step": 9361 }, { "epoch": 1.1909426281643556, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.063735246658325, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8706924319267273, "num_tokens": 355557729.0, "step": 9362 }, { "epoch": 1.1910698384429461, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.753460168838501, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8819310665130615, "num_tokens": 355597222.0, "step": 9363 }, { "epoch": 1.1911970487215366, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0327961444854736, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8862206935882568, "num_tokens": 355630477.0, "step": 9364 }, { "epoch": 1.1913242590001272, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9949588775634766, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8687559962272644, "num_tokens": 355664506.0, "step": 9365 }, { "epoch": 1.1914514692787177, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.1044106483459473, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8705090880393982, "num_tokens": 355710268.0, "step": 9366 }, { "epoch": 1.1915786795573082, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.234429359436035, "learning_rate": 1e-05, "loss": 0.5274, "mean_token_accuracy": 0.8550273180007935, "num_tokens": 355746396.0, "step": 9367 }, { "epoch": 1.1917058898358988, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4071857929229736, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8615155220031738, "num_tokens": 355787483.0, "step": 9368 }, { "epoch": 1.1918331001144893, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.218648433685303, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8625202775001526, "num_tokens": 355824138.0, "step": 9369 }, { "epoch": 1.1919603103930798, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.660008668899536, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8591463565826416, "num_tokens": 355854258.0, "step": 9370 }, { "epoch": 1.1920875206716703, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.001051902770996, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8673221468925476, "num_tokens": 355889750.0, "step": 9371 }, { "epoch": 1.1922147309502609, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.65118145942688, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8799504041671753, "num_tokens": 355923950.0, "step": 9372 }, { "epoch": 1.1923419412288512, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7987372875213623, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8779520392417908, "num_tokens": 355959807.0, "step": 9373 }, { "epoch": 1.1924691515074417, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.568319797515869, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8784344792366028, "num_tokens": 355992574.0, "step": 9374 }, { "epoch": 1.1925963617860322, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6355292797088623, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8620874285697937, "num_tokens": 356033726.0, "step": 9375 }, { "epoch": 1.1927235720646228, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002155303955078125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.48201060295105, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8799999952316284, "num_tokens": 356066959.0, "step": 9376 }, { "epoch": 1.1928507823432133, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5401840209960938, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8710944056510925, "num_tokens": 356102316.0, "step": 9377 }, { "epoch": 1.1929779926218038, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.373870849609375, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8814650177955627, "num_tokens": 356146203.0, "step": 9378 }, { "epoch": 1.1931052029003943, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.501678466796875, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8594434261322021, "num_tokens": 356184673.0, "step": 9379 }, { "epoch": 1.1932324131789849, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.294316053390503, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8661929368972778, "num_tokens": 356224693.0, "step": 9380 }, { "epoch": 1.1933596234575754, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.519094467163086, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8633589744567871, "num_tokens": 356265454.0, "step": 9381 }, { "epoch": 1.193486833736166, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2280468940734863, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8699754476547241, "num_tokens": 356302866.0, "step": 9382 }, { "epoch": 1.1936140440147565, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.557640552520752, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.87325119972229, "num_tokens": 356341897.0, "step": 9383 }, { "epoch": 1.193741254293347, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.147634744644165, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8763971328735352, "num_tokens": 356380637.0, "step": 9384 }, { "epoch": 1.1938684645719375, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.677236318588257, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8596805930137634, "num_tokens": 356422584.0, "step": 9385 }, { "epoch": 1.193995674850528, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.098592519760132, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8773900270462036, "num_tokens": 356457007.0, "step": 9386 }, { "epoch": 1.1941228851291183, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8597943782806396, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8736168146133423, "num_tokens": 356491837.0, "step": 9387 }, { "epoch": 1.1942500954077089, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1305830478668213, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8702582120895386, "num_tokens": 356527892.0, "step": 9388 }, { "epoch": 1.1943773056862994, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.025389194488525, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8643136024475098, "num_tokens": 356564310.0, "step": 9389 }, { "epoch": 1.19450451596489, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2670328617095947, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.878412127494812, "num_tokens": 356606476.0, "step": 9390 }, { "epoch": 1.1946317262434805, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.085259437561035, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8685489296913147, "num_tokens": 356646104.0, "step": 9391 }, { "epoch": 1.194758936522071, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.469583749771118, "learning_rate": 1e-05, "loss": 0.3982, "mean_token_accuracy": 0.8857269883155823, "num_tokens": 356687196.0, "step": 9392 }, { "epoch": 1.1948861468006615, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9680752754211426, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8694588541984558, "num_tokens": 356717634.0, "step": 9393 }, { "epoch": 1.195013357079252, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.580052137374878, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.856589674949646, "num_tokens": 356755494.0, "step": 9394 }, { "epoch": 1.1951405673578426, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7285115718841553, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8785485625267029, "num_tokens": 356800798.0, "step": 9395 }, { "epoch": 1.195267777636433, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5397069454193115, "learning_rate": 1e-05, "loss": 0.3936, "mean_token_accuracy": 0.8908467292785645, "num_tokens": 356837819.0, "step": 9396 }, { "epoch": 1.1953949879150234, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.53402042388916, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8771024942398071, "num_tokens": 356876371.0, "step": 9397 }, { "epoch": 1.195522198193614, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5077459812164307, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8704894185066223, "num_tokens": 356911427.0, "step": 9398 }, { "epoch": 1.1956494084722045, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3899855613708496, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8786509037017822, "num_tokens": 356951229.0, "step": 9399 }, { "epoch": 1.195776618750795, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.469696283340454, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8780789375305176, "num_tokens": 356988954.0, "step": 9400 }, { "epoch": 1.1959038290293855, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.298912763595581, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8562440276145935, "num_tokens": 357027982.0, "step": 9401 }, { "epoch": 1.196031039307976, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.475055694580078, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8685739040374756, "num_tokens": 357061434.0, "step": 9402 }, { "epoch": 1.1961582495865666, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.225175142288208, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8673740029335022, "num_tokens": 357099255.0, "step": 9403 }, { "epoch": 1.196285459865157, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.505053758621216, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8778663277626038, "num_tokens": 357131917.0, "step": 9404 }, { "epoch": 1.1964126701437476, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.152952194213867, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8791623115539551, "num_tokens": 357171520.0, "step": 9405 }, { "epoch": 1.1965398804223382, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5871143341064453, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8716081976890564, "num_tokens": 357208517.0, "step": 9406 }, { "epoch": 1.1966670907009287, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0944762229919434, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8802770376205444, "num_tokens": 357245300.0, "step": 9407 }, { "epoch": 1.1967943009795192, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7145235538482666, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8854974508285522, "num_tokens": 357284765.0, "step": 9408 }, { "epoch": 1.1969215112581097, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.062180995941162, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8613799810409546, "num_tokens": 357322968.0, "step": 9409 }, { "epoch": 1.1970487215367003, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9079928398132324, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8745150566101074, "num_tokens": 357360546.0, "step": 9410 }, { "epoch": 1.1971759318152906, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.106912136077881, "learning_rate": 1e-05, "loss": 0.3934, "mean_token_accuracy": 0.8849955797195435, "num_tokens": 357394433.0, "step": 9411 }, { "epoch": 1.1973031420938811, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.102475643157959, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8682825565338135, "num_tokens": 357431498.0, "step": 9412 }, { "epoch": 1.1974303523724716, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3103151321411133, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8662526607513428, "num_tokens": 357466930.0, "step": 9413 }, { "epoch": 1.1975575626510622, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.126885414123535, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8607196807861328, "num_tokens": 357509343.0, "step": 9414 }, { "epoch": 1.1976847729296527, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5152809619903564, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8716793656349182, "num_tokens": 357541980.0, "step": 9415 }, { "epoch": 1.1978119832082432, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.981920003890991, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8782921433448792, "num_tokens": 357579380.0, "step": 9416 }, { "epoch": 1.1979391934868338, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.572141170501709, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.880164384841919, "num_tokens": 357614666.0, "step": 9417 }, { "epoch": 1.1980664037654243, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7986602783203125, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.878538966178894, "num_tokens": 357648532.0, "step": 9418 }, { "epoch": 1.1981936140440148, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5387961864471436, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8761829137802124, "num_tokens": 357683585.0, "step": 9419 }, { "epoch": 1.1983208243226053, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00021648406982421875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6243982315063477, "learning_rate": 1e-05, "loss": 0.5126, "mean_token_accuracy": 0.8483995199203491, "num_tokens": 357725580.0, "step": 9420 }, { "epoch": 1.1984480346011959, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4870874881744385, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.868137538433075, "num_tokens": 357763017.0, "step": 9421 }, { "epoch": 1.1985752448797862, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4997808933258057, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8815054297447205, "num_tokens": 357797433.0, "step": 9422 }, { "epoch": 1.1987024551583767, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4094018936157227, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8685788512229919, "num_tokens": 357835348.0, "step": 9423 }, { "epoch": 1.1988296654369672, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4279541969299316, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8644956350326538, "num_tokens": 357877376.0, "step": 9424 }, { "epoch": 1.1989568757155578, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3528761863708496, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8693819046020508, "num_tokens": 357912983.0, "step": 9425 }, { "epoch": 1.1990840859941483, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.409717321395874, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8782501220703125, "num_tokens": 357948068.0, "step": 9426 }, { "epoch": 1.1992112962727388, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3037893772125244, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8761026859283447, "num_tokens": 357984019.0, "step": 9427 }, { "epoch": 1.1993385065513293, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.392918109893799, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.878221869468689, "num_tokens": 358022440.0, "step": 9428 }, { "epoch": 1.1994657168299199, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.252354860305786, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8694437742233276, "num_tokens": 358061615.0, "step": 9429 }, { "epoch": 1.1995929271085104, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.428745746612549, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8664929866790771, "num_tokens": 358097042.0, "step": 9430 }, { "epoch": 1.199720137387101, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.194706916809082, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8656021356582642, "num_tokens": 358134067.0, "step": 9431 }, { "epoch": 1.1998473476656915, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.484989881515503, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8692284822463989, "num_tokens": 358178975.0, "step": 9432 }, { "epoch": 1.199974557944282, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1316769123077393, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8729672431945801, "num_tokens": 358212879.0, "step": 9433 }, { "epoch": 1.2001017682228725, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.591510057449341, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8836768865585327, "num_tokens": 358245711.0, "step": 9434 }, { "epoch": 1.200228978501463, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0597352981567383, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.8835415244102478, "num_tokens": 358283073.0, "step": 9435 }, { "epoch": 1.2003561887800533, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7708728313446045, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8796635866165161, "num_tokens": 358320338.0, "step": 9436 }, { "epoch": 1.2004833990586439, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.039635181427002, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8705248832702637, "num_tokens": 358358704.0, "step": 9437 }, { "epoch": 1.2006106093372344, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.001485824584961, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8775238990783691, "num_tokens": 358393512.0, "step": 9438 }, { "epoch": 1.200737819615825, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.131150484085083, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8622947931289673, "num_tokens": 358435843.0, "step": 9439 }, { "epoch": 1.2008650298944155, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.203646183013916, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8880955576896667, "num_tokens": 358468918.0, "step": 9440 }, { "epoch": 1.200992240173006, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.409980297088623, "learning_rate": 1e-05, "loss": 0.524, "mean_token_accuracy": 0.8507315516471863, "num_tokens": 358511275.0, "step": 9441 }, { "epoch": 1.2011194504515965, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.175961494445801, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8633968830108643, "num_tokens": 358552665.0, "step": 9442 }, { "epoch": 1.201246660730187, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.623704671859741, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8787993788719177, "num_tokens": 358584725.0, "step": 9443 }, { "epoch": 1.2013738710087776, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9729881286621094, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.875645101070404, "num_tokens": 358620026.0, "step": 9444 }, { "epoch": 1.201501081287368, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6248395442962646, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8686719536781311, "num_tokens": 358657386.0, "step": 9445 }, { "epoch": 1.2016282915659584, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.775059461593628, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8739838600158691, "num_tokens": 358695069.0, "step": 9446 }, { "epoch": 1.201755501844549, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5482373237609863, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8780585527420044, "num_tokens": 358728091.0, "step": 9447 }, { "epoch": 1.2018827121231395, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.625946521759033, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8585203886032104, "num_tokens": 358767930.0, "step": 9448 }, { "epoch": 1.20200992240173, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.461005926132202, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8668907880783081, "num_tokens": 358803705.0, "step": 9449 }, { "epoch": 1.2021371326803205, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5266740322113037, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8707343339920044, "num_tokens": 358840730.0, "step": 9450 }, { "epoch": 1.202264342958911, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.375710964202881, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8649979829788208, "num_tokens": 358881457.0, "step": 9451 }, { "epoch": 1.2023915532375016, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4948205947875977, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8639379143714905, "num_tokens": 358917034.0, "step": 9452 }, { "epoch": 1.202518763516092, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2958269119262695, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8636358976364136, "num_tokens": 358957851.0, "step": 9453 }, { "epoch": 1.2026459737946826, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5064501762390137, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8617714643478394, "num_tokens": 358999816.0, "step": 9454 }, { "epoch": 1.2027731840732732, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.230649948120117, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8500322103500366, "num_tokens": 359041067.0, "step": 9455 }, { "epoch": 1.2029003943518637, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5563039779663086, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8687443733215332, "num_tokens": 359080495.0, "step": 9456 }, { "epoch": 1.2030276046304542, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1459617614746094, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8827114105224609, "num_tokens": 359118245.0, "step": 9457 }, { "epoch": 1.2031548149090447, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6787517070770264, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8593941330909729, "num_tokens": 359159840.0, "step": 9458 }, { "epoch": 1.2032820251876353, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0970962047576904, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8670403957366943, "num_tokens": 359198917.0, "step": 9459 }, { "epoch": 1.2034092354662256, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.854228973388672, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8680831789970398, "num_tokens": 359231904.0, "step": 9460 }, { "epoch": 1.203536445744816, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.117563247680664, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.857403039932251, "num_tokens": 359271713.0, "step": 9461 }, { "epoch": 1.2036636560234066, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.020669460296631, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8692156076431274, "num_tokens": 359312548.0, "step": 9462 }, { "epoch": 1.2037908663019972, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2634191513061523, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8796820640563965, "num_tokens": 359348171.0, "step": 9463 }, { "epoch": 1.2039180765805877, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.095579624176025, "learning_rate": 1e-05, "loss": 0.5145, "mean_token_accuracy": 0.8602498173713684, "num_tokens": 359383863.0, "step": 9464 }, { "epoch": 1.2040452868591782, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.47983455657959, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8660193085670471, "num_tokens": 359421595.0, "step": 9465 }, { "epoch": 1.2041724971377687, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9440813064575195, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.8827217221260071, "num_tokens": 359457889.0, "step": 9466 }, { "epoch": 1.2042997074163593, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.559783935546875, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8648227453231812, "num_tokens": 359496938.0, "step": 9467 }, { "epoch": 1.2044269176949498, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7414090633392334, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8632912039756775, "num_tokens": 359536124.0, "step": 9468 }, { "epoch": 1.2045541279735403, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5419602394104004, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8613430261611938, "num_tokens": 359573525.0, "step": 9469 }, { "epoch": 1.2046813382521309, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.541409969329834, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.873421847820282, "num_tokens": 359614960.0, "step": 9470 }, { "epoch": 1.2048085485307212, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5131332874298096, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.864121675491333, "num_tokens": 359653367.0, "step": 9471 }, { "epoch": 1.2049357588093117, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.39890193939209, "learning_rate": 1e-05, "loss": 0.5085, "mean_token_accuracy": 0.8504073023796082, "num_tokens": 359692983.0, "step": 9472 }, { "epoch": 1.2050629690879022, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000217437744140625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4863407611846924, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8608661890029907, "num_tokens": 359729487.0, "step": 9473 }, { "epoch": 1.2051901793664928, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.290383815765381, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.877314567565918, "num_tokens": 359764805.0, "step": 9474 }, { "epoch": 1.2053173896450833, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.488002300262451, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.858481764793396, "num_tokens": 359803014.0, "step": 9475 }, { "epoch": 1.2054445999236738, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.229398012161255, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8708714842796326, "num_tokens": 359837866.0, "step": 9476 }, { "epoch": 1.2055718102022643, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5111351013183594, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8834011554718018, "num_tokens": 359874052.0, "step": 9477 }, { "epoch": 1.2056990204808549, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1482250690460205, "learning_rate": 1e-05, "loss": 0.3892, "mean_token_accuracy": 0.8879395723342896, "num_tokens": 359915746.0, "step": 9478 }, { "epoch": 1.2058262307594454, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6100399494171143, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8832025527954102, "num_tokens": 359945857.0, "step": 9479 }, { "epoch": 1.205953441038036, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.082753896713257, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8761392831802368, "num_tokens": 359988455.0, "step": 9480 }, { "epoch": 1.2060806513166265, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7657155990600586, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8762161135673523, "num_tokens": 360030610.0, "step": 9481 }, { "epoch": 1.206207861595217, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.064718008041382, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8617822527885437, "num_tokens": 360068919.0, "step": 9482 }, { "epoch": 1.2063350718738075, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9712562561035156, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8680820465087891, "num_tokens": 360099073.0, "step": 9483 }, { "epoch": 1.206462282152398, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.146803140640259, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8715283274650574, "num_tokens": 360141947.0, "step": 9484 }, { "epoch": 1.2065894924309883, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.146307468414307, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.8671940565109253, "num_tokens": 360179435.0, "step": 9485 }, { "epoch": 1.2067167027095789, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.3739469051361084, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8765918612480164, "num_tokens": 360221254.0, "step": 9486 }, { "epoch": 1.2068439129881694, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.12231969833374, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.879163384437561, "num_tokens": 360261657.0, "step": 9487 }, { "epoch": 1.20697112326676, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5656726360321045, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8742225170135498, "num_tokens": 360294408.0, "step": 9488 }, { "epoch": 1.2070983335453505, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9534943103790283, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.879127025604248, "num_tokens": 360332275.0, "step": 9489 }, { "epoch": 1.207225543823941, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5835604667663574, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.876936137676239, "num_tokens": 360372381.0, "step": 9490 }, { "epoch": 1.2073527541025315, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.784024477005005, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8654733896255493, "num_tokens": 360404066.0, "step": 9491 }, { "epoch": 1.207479964381122, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.550318717956543, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8791948556900024, "num_tokens": 360434457.0, "step": 9492 }, { "epoch": 1.2076071746597126, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.610612154006958, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8711602687835693, "num_tokens": 360467905.0, "step": 9493 }, { "epoch": 1.207734384938303, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4642391204833984, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8735877275466919, "num_tokens": 360505190.0, "step": 9494 }, { "epoch": 1.2078615952168934, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.507070541381836, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8740226030349731, "num_tokens": 360539191.0, "step": 9495 }, { "epoch": 1.207988805495484, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.413090944290161, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.858544111251831, "num_tokens": 360569508.0, "step": 9496 }, { "epoch": 1.2081160157740745, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.442521333694458, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8795797228813171, "num_tokens": 360604669.0, "step": 9497 }, { "epoch": 1.208243226052665, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.33090877532959, "learning_rate": 1e-05, "loss": 0.3692, "mean_token_accuracy": 0.8933788537979126, "num_tokens": 360639009.0, "step": 9498 }, { "epoch": 1.2083704363312555, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4261722564697266, "learning_rate": 1e-05, "loss": 0.3842, "mean_token_accuracy": 0.8865577578544617, "num_tokens": 360673857.0, "step": 9499 }, { "epoch": 1.208497646609846, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.26641583442688, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.861952543258667, "num_tokens": 360717553.0, "step": 9500 }, { "epoch": 1.2086248568884366, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4486658573150635, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8728493452072144, "num_tokens": 360761013.0, "step": 9501 }, { "epoch": 1.208752067167027, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1942923069000244, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8706401586532593, "num_tokens": 360804302.0, "step": 9502 }, { "epoch": 1.2088792774456176, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.521040916442871, "learning_rate": 1e-05, "loss": 0.3665, "mean_token_accuracy": 0.8922812342643738, "num_tokens": 360837045.0, "step": 9503 }, { "epoch": 1.2090064877242082, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1326377391815186, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8714867830276489, "num_tokens": 360876243.0, "step": 9504 }, { "epoch": 1.2091336980027987, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6421170234680176, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8666968941688538, "num_tokens": 360915112.0, "step": 9505 }, { "epoch": 1.2092609082813892, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.069946050643921, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8731122016906738, "num_tokens": 360951856.0, "step": 9506 }, { "epoch": 1.2093881185599797, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.823465585708618, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8780773878097534, "num_tokens": 360987243.0, "step": 9507 }, { "epoch": 1.2095153288385703, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.063542366027832, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8758994340896606, "num_tokens": 361025099.0, "step": 9508 }, { "epoch": 1.2096425391171606, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.080466270446777, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8537153601646423, "num_tokens": 361055232.0, "step": 9509 }, { "epoch": 1.209769749395751, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.2368781566619873, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.870914101600647, "num_tokens": 361092500.0, "step": 9510 }, { "epoch": 1.2098969596743416, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1959547996521, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8772882223129272, "num_tokens": 361127044.0, "step": 9511 }, { "epoch": 1.2100241699529322, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5009734630584717, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.8912819623947144, "num_tokens": 361163065.0, "step": 9512 }, { "epoch": 1.2101513802315227, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.069708824157715, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8625375032424927, "num_tokens": 361204465.0, "step": 9513 }, { "epoch": 1.2102785905101132, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6241602897644043, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8498275876045227, "num_tokens": 361243788.0, "step": 9514 }, { "epoch": 1.2104058007887037, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8623459339141846, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8760189414024353, "num_tokens": 361281427.0, "step": 9515 }, { "epoch": 1.2105330110672943, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5941481590270996, "learning_rate": 1e-05, "loss": 0.5022, "mean_token_accuracy": 0.8568665981292725, "num_tokens": 361324695.0, "step": 9516 }, { "epoch": 1.2106602213458848, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6584300994873047, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.869701623916626, "num_tokens": 361370446.0, "step": 9517 }, { "epoch": 1.2107874316244753, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5129969120025635, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8699861764907837, "num_tokens": 361414622.0, "step": 9518 }, { "epoch": 1.2109146419030659, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5260167121887207, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8818225860595703, "num_tokens": 361452363.0, "step": 9519 }, { "epoch": 1.2110418521816562, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4581375122070312, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8547991514205933, "num_tokens": 361484195.0, "step": 9520 }, { "epoch": 1.2111690624602467, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4376277923583984, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8706716895103455, "num_tokens": 361523604.0, "step": 9521 }, { "epoch": 1.2112962727388372, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3921256065368652, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8552293181419373, "num_tokens": 361563723.0, "step": 9522 }, { "epoch": 1.2114234830174277, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.38165545463562, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8727065324783325, "num_tokens": 361602578.0, "step": 9523 }, { "epoch": 1.2115506932960183, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.350083351135254, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8696737885475159, "num_tokens": 361644255.0, "step": 9524 }, { "epoch": 1.2116779035746088, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3560218811035156, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8711118102073669, "num_tokens": 361683424.0, "step": 9525 }, { "epoch": 1.2118051138531993, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3022069931030273, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8814513683319092, "num_tokens": 361720965.0, "step": 9526 }, { "epoch": 1.2119323241317899, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3819432258605957, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8579107522964478, "num_tokens": 361753900.0, "step": 9527 }, { "epoch": 1.2120595344103804, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.255047559738159, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8710016012191772, "num_tokens": 361793072.0, "step": 9528 }, { "epoch": 1.212186744688971, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3968703746795654, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8543283939361572, "num_tokens": 361831082.0, "step": 9529 }, { "epoch": 1.2123139549675614, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.196770668029785, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8737211227416992, "num_tokens": 361864834.0, "step": 9530 }, { "epoch": 1.212441165246152, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.455484390258789, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8725265264511108, "num_tokens": 361905567.0, "step": 9531 }, { "epoch": 1.2125683755247425, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.13131046295166, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8806135058403015, "num_tokens": 361945718.0, "step": 9532 }, { "epoch": 1.212695585803333, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.562551498413086, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8640788197517395, "num_tokens": 361985317.0, "step": 9533 }, { "epoch": 1.2128227960819233, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.315376281738281e-05, "grad_norm": 3.057281017303467, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8726442456245422, "num_tokens": 362020725.0, "step": 9534 }, { "epoch": 1.2129500063605139, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.763394832611084, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8600667715072632, "num_tokens": 362054398.0, "step": 9535 }, { "epoch": 1.2130772166391044, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 2.997485637664795, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8559272289276123, "num_tokens": 362092075.0, "step": 9536 }, { "epoch": 1.213204426917695, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0381903648376465, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8915544748306274, "num_tokens": 362128286.0, "step": 9537 }, { "epoch": 1.2133316371962855, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.0858211517333984, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8623595237731934, "num_tokens": 362169965.0, "step": 9538 }, { "epoch": 1.213458847474876, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.325385093688965, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8689169883728027, "num_tokens": 362211638.0, "step": 9539 }, { "epoch": 1.2135860577534665, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.45822811126709, "learning_rate": 1e-05, "loss": 0.4009, "mean_token_accuracy": 0.8880159854888916, "num_tokens": 362241597.0, "step": 9540 }, { "epoch": 1.213713268032057, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.3037309646606445, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.868493914604187, "num_tokens": 362280028.0, "step": 9541 }, { "epoch": 1.2138404783106476, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.7271664142608643, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8697940707206726, "num_tokens": 362323611.0, "step": 9542 }, { "epoch": 1.213967688589238, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.067225456237793, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8715028762817383, "num_tokens": 362359097.0, "step": 9543 }, { "epoch": 1.2140948988678284, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.724119186401367, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.868749737739563, "num_tokens": 362402175.0, "step": 9544 }, { "epoch": 1.214222109146419, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8569960594177246, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8717352151870728, "num_tokens": 362442581.0, "step": 9545 }, { "epoch": 1.2143493194250095, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6078221797943115, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8795610666275024, "num_tokens": 362480691.0, "step": 9546 }, { "epoch": 1.2144765297036, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.706678628921509, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8725182414054871, "num_tokens": 362522864.0, "step": 9547 }, { "epoch": 1.2146037399821905, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4847846031188965, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8727458119392395, "num_tokens": 362564431.0, "step": 9548 }, { "epoch": 1.214730950260781, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6291310787200928, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8827847242355347, "num_tokens": 362599670.0, "step": 9549 }, { "epoch": 1.2148581605393716, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.381330966949463, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8662094473838806, "num_tokens": 362637588.0, "step": 9550 }, { "epoch": 1.214985370817962, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.605255365371704, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8728254437446594, "num_tokens": 362674679.0, "step": 9551 }, { "epoch": 1.2151125810965526, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.290485143661499, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.866542398929596, "num_tokens": 362705840.0, "step": 9552 }, { "epoch": 1.2152397913751432, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.640014886856079, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8684722185134888, "num_tokens": 362739880.0, "step": 9553 }, { "epoch": 1.2153670016537337, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1968324184417725, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8790675401687622, "num_tokens": 362779101.0, "step": 9554 }, { "epoch": 1.2154942119323242, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7466588020324707, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8750327229499817, "num_tokens": 362821034.0, "step": 9555 }, { "epoch": 1.2156214222109147, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.154421329498291, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8797376155853271, "num_tokens": 362856895.0, "step": 9556 }, { "epoch": 1.2157486324895053, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9221057891845703, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8694117069244385, "num_tokens": 362894955.0, "step": 9557 }, { "epoch": 1.2158758427680956, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2151126861572266, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8659623861312866, "num_tokens": 362930397.0, "step": 9558 }, { "epoch": 1.216003053046686, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.027265548706055, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8882098197937012, "num_tokens": 362970728.0, "step": 9559 }, { "epoch": 1.2161302633252766, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3752262592315674, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8851625323295593, "num_tokens": 363007803.0, "step": 9560 }, { "epoch": 1.2162574736038672, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.997157573699951, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8686118721961975, "num_tokens": 363047815.0, "step": 9561 }, { "epoch": 1.2163846838824577, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5203452110290527, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8788287043571472, "num_tokens": 363084156.0, "step": 9562 }, { "epoch": 1.2165118941610482, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8090133666992188, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8686907291412354, "num_tokens": 363125390.0, "step": 9563 }, { "epoch": 1.2166391044396387, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.561990976333618, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8722223043441772, "num_tokens": 363164210.0, "step": 9564 }, { "epoch": 1.2167663147182293, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.589064121246338, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8545601963996887, "num_tokens": 363200356.0, "step": 9565 }, { "epoch": 1.2168935249968198, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5421881675720215, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8522372841835022, "num_tokens": 363239622.0, "step": 9566 }, { "epoch": 1.2170207352754103, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.401095390319824, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8751604557037354, "num_tokens": 363277348.0, "step": 9567 }, { "epoch": 1.2171479455540009, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.54404616355896, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8712238073348999, "num_tokens": 363310077.0, "step": 9568 }, { "epoch": 1.2172751558325912, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.278193712234497, "learning_rate": 1e-05, "loss": 0.5443, "mean_token_accuracy": 0.8375110626220703, "num_tokens": 363347919.0, "step": 9569 }, { "epoch": 1.2174023661111817, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5458130836486816, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8737962245941162, "num_tokens": 363390818.0, "step": 9570 }, { "epoch": 1.2175295763897722, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.2010622024536133, "learning_rate": 1e-05, "loss": 0.5317, "mean_token_accuracy": 0.8456737995147705, "num_tokens": 363424028.0, "step": 9571 }, { "epoch": 1.2176567866683627, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6132607460021973, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8685282468795776, "num_tokens": 363456862.0, "step": 9572 }, { "epoch": 1.2177839969469533, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1056582927703857, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.873779833316803, "num_tokens": 363498538.0, "step": 9573 }, { "epoch": 1.2179112072255438, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7386763095855713, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8752185106277466, "num_tokens": 363533050.0, "step": 9574 }, { "epoch": 1.2180384175041343, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.094940662384033, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8700002431869507, "num_tokens": 363574151.0, "step": 9575 }, { "epoch": 1.2181656277827249, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8947107791900635, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8659784197807312, "num_tokens": 363615829.0, "step": 9576 }, { "epoch": 1.2182928380613154, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1463282108306885, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8686597943305969, "num_tokens": 363649974.0, "step": 9577 }, { "epoch": 1.218420048339906, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.047098159790039, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8729101419448853, "num_tokens": 363691196.0, "step": 9578 }, { "epoch": 1.2185472586184964, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.296678066253662, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.8825916051864624, "num_tokens": 363726496.0, "step": 9579 }, { "epoch": 1.218674468897087, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.074952125549316, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8605566620826721, "num_tokens": 363769119.0, "step": 9580 }, { "epoch": 1.2188016791756775, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.489607810974121, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.8546477556228638, "num_tokens": 363803363.0, "step": 9581 }, { "epoch": 1.218928889454268, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9422078132629395, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8747816681861877, "num_tokens": 363842602.0, "step": 9582 }, { "epoch": 1.2190560997328583, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5305981636047363, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8695731163024902, "num_tokens": 363882758.0, "step": 9583 }, { "epoch": 1.2191833100114489, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.759054660797119, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.884831428527832, "num_tokens": 363923738.0, "step": 9584 }, { "epoch": 1.2193105202900394, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5108091831207275, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8788701295852661, "num_tokens": 363957999.0, "step": 9585 }, { "epoch": 1.21943773056863, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6082563400268555, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.873023509979248, "num_tokens": 363992653.0, "step": 9586 }, { "epoch": 1.2195649408472204, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.45333194732666, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8847247362136841, "num_tokens": 364027376.0, "step": 9587 }, { "epoch": 1.219692151125811, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.49796462059021, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8690629601478577, "num_tokens": 364063299.0, "step": 9588 }, { "epoch": 1.2198193614044015, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.39652419090271, "learning_rate": 1e-05, "loss": 0.527, "mean_token_accuracy": 0.8436024188995361, "num_tokens": 364105724.0, "step": 9589 }, { "epoch": 1.219946571682992, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4547109603881836, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8609751462936401, "num_tokens": 364140427.0, "step": 9590 }, { "epoch": 1.2200737819615826, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3253626823425293, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8592493534088135, "num_tokens": 364185262.0, "step": 9591 }, { "epoch": 1.220200992240173, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.440220594406128, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8721346855163574, "num_tokens": 364219006.0, "step": 9592 }, { "epoch": 1.2203282025187634, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2597053050994873, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8762186169624329, "num_tokens": 364259699.0, "step": 9593 }, { "epoch": 1.220455412797354, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4683587551116943, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8766252994537354, "num_tokens": 364297759.0, "step": 9594 }, { "epoch": 1.2205826230759445, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.20491099357605, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8671804666519165, "num_tokens": 364338683.0, "step": 9595 }, { "epoch": 1.220709833354535, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.540484666824341, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8712615966796875, "num_tokens": 364377587.0, "step": 9596 }, { "epoch": 1.2208370436331255, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1369025707244873, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8639582991600037, "num_tokens": 364416740.0, "step": 9597 }, { "epoch": 1.220964253911716, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.655545711517334, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8606588840484619, "num_tokens": 364458695.0, "step": 9598 }, { "epoch": 1.2210914641903066, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0834248065948486, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8618277311325073, "num_tokens": 364501395.0, "step": 9599 }, { "epoch": 1.221218674468897, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8273332118988037, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8772433996200562, "num_tokens": 364534126.0, "step": 9600 }, { "epoch": 1.2213458847474876, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.092148542404175, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.884302020072937, "num_tokens": 364573776.0, "step": 9601 }, { "epoch": 1.2214730950260781, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.014840602874756, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8671801686286926, "num_tokens": 364617900.0, "step": 9602 }, { "epoch": 1.2216003053046687, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2285783290863037, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.878255307674408, "num_tokens": 364658992.0, "step": 9603 }, { "epoch": 1.2217275155832592, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.103764057159424, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8851795792579651, "num_tokens": 364695224.0, "step": 9604 }, { "epoch": 1.2218547258618497, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.4566752910614014, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8733310103416443, "num_tokens": 364731476.0, "step": 9605 }, { "epoch": 1.2219819361404403, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.987208127975464, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8715890645980835, "num_tokens": 364774129.0, "step": 9606 }, { "epoch": 1.2221091464190306, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5622220039367676, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8717976808547974, "num_tokens": 364808097.0, "step": 9607 }, { "epoch": 1.222236356697621, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7830519676208496, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8791792392730713, "num_tokens": 364847592.0, "step": 9608 }, { "epoch": 1.2223635669762116, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5412042140960693, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8739240765571594, "num_tokens": 364885846.0, "step": 9609 }, { "epoch": 1.2224907772548022, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002193450927734375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.592688798904419, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.856624960899353, "num_tokens": 364927536.0, "step": 9610 }, { "epoch": 1.2226179875333927, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5076098442077637, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8709012269973755, "num_tokens": 364961050.0, "step": 9611 }, { "epoch": 1.2227451978119832, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.449918746948242, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8654760718345642, "num_tokens": 365001058.0, "step": 9612 }, { "epoch": 1.2228724080905737, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.447721004486084, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.878265917301178, "num_tokens": 365041345.0, "step": 9613 }, { "epoch": 1.2229996183691643, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.351557731628418, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8684821128845215, "num_tokens": 365085669.0, "step": 9614 }, { "epoch": 1.2231268286477548, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4156010150909424, "learning_rate": 1e-05, "loss": 0.4023, "mean_token_accuracy": 0.8861343860626221, "num_tokens": 365125630.0, "step": 9615 }, { "epoch": 1.2232540389263453, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.297755002975464, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8737159967422485, "num_tokens": 365162904.0, "step": 9616 }, { "epoch": 1.2233812492049359, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.420383930206299, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8740127086639404, "num_tokens": 365196509.0, "step": 9617 }, { "epoch": 1.2235084594835262, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2576193809509277, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8576363325119019, "num_tokens": 365232947.0, "step": 9618 }, { "epoch": 1.2236356697621167, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.421463966369629, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8727365732192993, "num_tokens": 365273434.0, "step": 9619 }, { "epoch": 1.2237628800407072, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.215043306350708, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.865520715713501, "num_tokens": 365316638.0, "step": 9620 }, { "epoch": 1.2238900903192977, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.474489212036133, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8662481904029846, "num_tokens": 365348519.0, "step": 9621 }, { "epoch": 1.2240173005978883, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.15356707572937, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8847812414169312, "num_tokens": 365383782.0, "step": 9622 }, { "epoch": 1.2241445108764788, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5664851665496826, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8613158464431763, "num_tokens": 365421363.0, "step": 9623 }, { "epoch": 1.2242717211550693, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0655808448791504, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8708420991897583, "num_tokens": 365468207.0, "step": 9624 }, { "epoch": 1.2243989314336599, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7442281246185303, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.876802921295166, "num_tokens": 365498318.0, "step": 9625 }, { "epoch": 1.2245261417122504, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0151665210723877, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8727786540985107, "num_tokens": 365536140.0, "step": 9626 }, { "epoch": 1.224653351990841, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9998042583465576, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8565064668655396, "num_tokens": 365576490.0, "step": 9627 }, { "epoch": 1.2247805622694314, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.0887181758880615, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8673195242881775, "num_tokens": 365618043.0, "step": 9628 }, { "epoch": 1.224907772548022, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.271825313568115, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8633267283439636, "num_tokens": 365648303.0, "step": 9629 }, { "epoch": 1.2250349828266125, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4195950031280518, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8706282377243042, "num_tokens": 365684739.0, "step": 9630 }, { "epoch": 1.225162193105203, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.259011745452881, "learning_rate": 1e-05, "loss": 0.506, "mean_token_accuracy": 0.8566718101501465, "num_tokens": 365726542.0, "step": 9631 }, { "epoch": 1.2252894033837933, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.64093017578125, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8803995251655579, "num_tokens": 365769949.0, "step": 9632 }, { "epoch": 1.2254166136623839, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.077076435089111, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8737788200378418, "num_tokens": 365805891.0, "step": 9633 }, { "epoch": 1.2255438239409744, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6896915435791016, "learning_rate": 1e-05, "loss": 0.5263, "mean_token_accuracy": 0.8524337410926819, "num_tokens": 365847011.0, "step": 9634 }, { "epoch": 1.225671034219565, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8725168704986572, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8729066252708435, "num_tokens": 365884546.0, "step": 9635 }, { "epoch": 1.2257982444981554, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.605227470397949, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8794317245483398, "num_tokens": 365919619.0, "step": 9636 }, { "epoch": 1.225925454776746, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7151668071746826, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8775449991226196, "num_tokens": 365956020.0, "step": 9637 }, { "epoch": 1.2260526650553365, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.500143051147461, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.8490707874298096, "num_tokens": 365994945.0, "step": 9638 }, { "epoch": 1.226179875333927, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6144418716430664, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8624527454376221, "num_tokens": 366036476.0, "step": 9639 }, { "epoch": 1.2263070856125176, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.39164400100708, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8779796957969666, "num_tokens": 366072966.0, "step": 9640 }, { "epoch": 1.226434295891108, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.578894853591919, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8718686103820801, "num_tokens": 366111535.0, "step": 9641 }, { "epoch": 1.2265615061696984, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.292699098587036, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.866214394569397, "num_tokens": 366151104.0, "step": 9642 }, { "epoch": 1.226688716448289, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022029876708984375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6313858032226562, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8616647720336914, "num_tokens": 366184062.0, "step": 9643 }, { "epoch": 1.2268159267268794, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.203061103820801, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8738659620285034, "num_tokens": 366217774.0, "step": 9644 }, { "epoch": 1.22694313700547, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.724879264831543, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8790445327758789, "num_tokens": 366252776.0, "step": 9645 }, { "epoch": 1.2270703472840605, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1532936096191406, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8686580657958984, "num_tokens": 366289031.0, "step": 9646 }, { "epoch": 1.227197557562651, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8712072372436523, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8816414475440979, "num_tokens": 366327919.0, "step": 9647 }, { "epoch": 1.2273247678412416, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.180351972579956, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8561792969703674, "num_tokens": 366368655.0, "step": 9648 }, { "epoch": 1.227451978119832, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.025215148925781, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8704338073730469, "num_tokens": 366406987.0, "step": 9649 }, { "epoch": 1.2275791883984226, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.333556890487671, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8571484088897705, "num_tokens": 366447360.0, "step": 9650 }, { "epoch": 1.2277063986770131, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.023129940032959, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8777961730957031, "num_tokens": 366487640.0, "step": 9651 }, { "epoch": 1.2278336089556037, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5058951377868652, "learning_rate": 1e-05, "loss": 0.4931, "mean_token_accuracy": 0.8614844679832458, "num_tokens": 366526569.0, "step": 9652 }, { "epoch": 1.2279608192341942, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.845247745513916, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8659336566925049, "num_tokens": 366567503.0, "step": 9653 }, { "epoch": 1.2280880295127847, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5528945922851562, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8695335388183594, "num_tokens": 366609230.0, "step": 9654 }, { "epoch": 1.2282152397913753, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6292667388916016, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8733563423156738, "num_tokens": 366646982.0, "step": 9655 }, { "epoch": 1.2283424500699656, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5538527965545654, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8484058380126953, "num_tokens": 366684051.0, "step": 9656 }, { "epoch": 1.228469660348556, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.421027183532715, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8544520735740662, "num_tokens": 366724617.0, "step": 9657 }, { "epoch": 1.2285968706271466, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5468316078186035, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8855550289154053, "num_tokens": 366756642.0, "step": 9658 }, { "epoch": 1.2287240809057371, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2833316326141357, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8701069355010986, "num_tokens": 366789197.0, "step": 9659 }, { "epoch": 1.2288512911843277, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5690908432006836, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8674367666244507, "num_tokens": 366825684.0, "step": 9660 }, { "epoch": 1.2289785014629182, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1751444339752197, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8589383363723755, "num_tokens": 366871597.0, "step": 9661 }, { "epoch": 1.2291057117415087, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.643859624862671, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8857349157333374, "num_tokens": 366910961.0, "step": 9662 }, { "epoch": 1.2292329220200993, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0991902351379395, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8840616345405579, "num_tokens": 366951701.0, "step": 9663 }, { "epoch": 1.2293601322986898, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7973687648773193, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8602643609046936, "num_tokens": 366992472.0, "step": 9664 }, { "epoch": 1.2294873425772803, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1103620529174805, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.861734926700592, "num_tokens": 367031220.0, "step": 9665 }, { "epoch": 1.2296145528558708, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9956958293914795, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8627187013626099, "num_tokens": 367067275.0, "step": 9666 }, { "epoch": 1.2297417631344612, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.200666904449463, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.8539782166481018, "num_tokens": 367111851.0, "step": 9667 }, { "epoch": 1.2298689734130517, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.082908630371094, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8640961647033691, "num_tokens": 367150895.0, "step": 9668 }, { "epoch": 1.2299961836916422, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.395357370376587, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8577564358711243, "num_tokens": 367187874.0, "step": 9669 }, { "epoch": 1.2301233939702327, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0323896408081055, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8750388622283936, "num_tokens": 367228859.0, "step": 9670 }, { "epoch": 1.2302506042488233, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.521723985671997, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8850811719894409, "num_tokens": 367264686.0, "step": 9671 }, { "epoch": 1.2303778145274138, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8647780418395996, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8746956586837769, "num_tokens": 367309909.0, "step": 9672 }, { "epoch": 1.2305050248060043, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.538228750228882, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8782753944396973, "num_tokens": 367343996.0, "step": 9673 }, { "epoch": 1.2306322350845948, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.696911096572876, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8667702674865723, "num_tokens": 367386267.0, "step": 9674 }, { "epoch": 1.2307594453631854, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4824178218841553, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8624663352966309, "num_tokens": 367430256.0, "step": 9675 }, { "epoch": 1.230886655641776, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5622406005859375, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8563461303710938, "num_tokens": 367469189.0, "step": 9676 }, { "epoch": 1.2310138659203664, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.453094244003296, "learning_rate": 1e-05, "loss": 0.5002, "mean_token_accuracy": 0.8555874824523926, "num_tokens": 367501323.0, "step": 9677 }, { "epoch": 1.231141076198957, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.458159923553467, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8651862144470215, "num_tokens": 367542361.0, "step": 9678 }, { "epoch": 1.2312682864775475, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3997321128845215, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8539021611213684, "num_tokens": 367575859.0, "step": 9679 }, { "epoch": 1.231395496756138, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4105188846588135, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8705700635910034, "num_tokens": 367611330.0, "step": 9680 }, { "epoch": 1.2315227070347283, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.317671537399292, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.8809896111488342, "num_tokens": 367653110.0, "step": 9681 }, { "epoch": 1.2316499173133189, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4136059284210205, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8769688010215759, "num_tokens": 367687799.0, "step": 9682 }, { "epoch": 1.2317771275919094, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.261802911758423, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8714505434036255, "num_tokens": 367728253.0, "step": 9683 }, { "epoch": 1.2319043378705, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4255917072296143, "learning_rate": 1e-05, "loss": 0.3631, "mean_token_accuracy": 0.8981608152389526, "num_tokens": 367770397.0, "step": 9684 }, { "epoch": 1.2320315481490904, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2197675704956055, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8696116209030151, "num_tokens": 367810630.0, "step": 9685 }, { "epoch": 1.232158758427681, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4931042194366455, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8630238175392151, "num_tokens": 367853563.0, "step": 9686 }, { "epoch": 1.2322859687062715, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1709110736846924, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8569425344467163, "num_tokens": 367888113.0, "step": 9687 }, { "epoch": 1.232413178984862, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.581284999847412, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8649858236312866, "num_tokens": 367925839.0, "step": 9688 }, { "epoch": 1.2325403892634526, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0869874954223633, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.860062837600708, "num_tokens": 367966944.0, "step": 9689 }, { "epoch": 1.232667599542043, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.745744466781616, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8715355396270752, "num_tokens": 367999470.0, "step": 9690 }, { "epoch": 1.2327948098206334, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0260138511657715, "learning_rate": 1e-05, "loss": 0.3861, "mean_token_accuracy": 0.8876942992210388, "num_tokens": 368042780.0, "step": 9691 }, { "epoch": 1.232922020099224, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9933085441589355, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.874111533164978, "num_tokens": 368078062.0, "step": 9692 }, { "epoch": 1.2330492303778144, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.122222900390625, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.871094822883606, "num_tokens": 368121570.0, "step": 9693 }, { "epoch": 1.233176440656405, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2170209884643555, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8766178488731384, "num_tokens": 368158576.0, "step": 9694 }, { "epoch": 1.2333036509349955, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.3928885459899902, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8880410194396973, "num_tokens": 368199939.0, "step": 9695 }, { "epoch": 1.233430861213586, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.214233875274658, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8637778162956238, "num_tokens": 368240156.0, "step": 9696 }, { "epoch": 1.2335580714921766, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6600513458251953, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8515889644622803, "num_tokens": 368270852.0, "step": 9697 }, { "epoch": 1.233685281770767, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9931399822235107, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8642169237136841, "num_tokens": 368311986.0, "step": 9698 }, { "epoch": 1.2338124920493576, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6447787284851074, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8633934259414673, "num_tokens": 368354330.0, "step": 9699 }, { "epoch": 1.2339397023279481, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.792337656021118, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8781847953796387, "num_tokens": 368392712.0, "step": 9700 }, { "epoch": 1.2340669126065387, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5668647289276123, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8632635474205017, "num_tokens": 368430913.0, "step": 9701 }, { "epoch": 1.2341941228851292, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6445109844207764, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8847000598907471, "num_tokens": 368462353.0, "step": 9702 }, { "epoch": 1.2343213331637197, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022125244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.46647310256958, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8622039556503296, "num_tokens": 368506061.0, "step": 9703 }, { "epoch": 1.2344485434423103, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5603792667388916, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8771381378173828, "num_tokens": 368539435.0, "step": 9704 }, { "epoch": 1.2345757537209006, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3704652786254883, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8674336075782776, "num_tokens": 368578249.0, "step": 9705 }, { "epoch": 1.234702963999491, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.521825075149536, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8840818405151367, "num_tokens": 368617766.0, "step": 9706 }, { "epoch": 1.2348301742780816, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2861952781677246, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8780080080032349, "num_tokens": 368651458.0, "step": 9707 }, { "epoch": 1.2349573845566721, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5551273822784424, "learning_rate": 1e-05, "loss": 0.5102, "mean_token_accuracy": 0.8529629707336426, "num_tokens": 368688689.0, "step": 9708 }, { "epoch": 1.2350845948352627, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.194380760192871, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8504345417022705, "num_tokens": 368731996.0, "step": 9709 }, { "epoch": 1.2352118051138532, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6471774578094482, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8799600005149841, "num_tokens": 368772712.0, "step": 9710 }, { "epoch": 1.2353390153924437, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1244161128997803, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8754844069480896, "num_tokens": 368816582.0, "step": 9711 }, { "epoch": 1.2354662256710343, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.802016019821167, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8742407560348511, "num_tokens": 368854284.0, "step": 9712 }, { "epoch": 1.2355934359496248, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1235544681549072, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8723396062850952, "num_tokens": 368891564.0, "step": 9713 }, { "epoch": 1.2357206462282153, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.980254650115967, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8546870946884155, "num_tokens": 368931176.0, "step": 9714 }, { "epoch": 1.2358478565068058, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2448461055755615, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8759442567825317, "num_tokens": 368967730.0, "step": 9715 }, { "epoch": 1.2359750667853961, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.047023773193359, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8737128973007202, "num_tokens": 369009378.0, "step": 9716 }, { "epoch": 1.2361022770639867, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.416098117828369, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8643521070480347, "num_tokens": 369045716.0, "step": 9717 }, { "epoch": 1.2362294873425772, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.967468738555908, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8798767328262329, "num_tokens": 369082022.0, "step": 9718 }, { "epoch": 1.2363566976211677, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5268187522888184, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8626297116279602, "num_tokens": 369122456.0, "step": 9719 }, { "epoch": 1.2364839078997583, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7886126041412354, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8584553003311157, "num_tokens": 369156977.0, "step": 9720 }, { "epoch": 1.2366111181783488, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5442662239074707, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8594998121261597, "num_tokens": 369195113.0, "step": 9721 }, { "epoch": 1.2367383284569393, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.598039388656616, "learning_rate": 1e-05, "loss": 0.4167, "mean_token_accuracy": 0.8812155723571777, "num_tokens": 369238650.0, "step": 9722 }, { "epoch": 1.2368655387355298, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.529205560684204, "learning_rate": 1e-05, "loss": 0.5108, "mean_token_accuracy": 0.8515821695327759, "num_tokens": 369275753.0, "step": 9723 }, { "epoch": 1.2369927490141204, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.418224811553955, "learning_rate": 1e-05, "loss": 0.5127, "mean_token_accuracy": 0.8460071086883545, "num_tokens": 369317364.0, "step": 9724 }, { "epoch": 1.237119959292711, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.499199867248535, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8733109831809998, "num_tokens": 369355889.0, "step": 9725 }, { "epoch": 1.2372471695713014, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2871286869049072, "learning_rate": 1e-05, "loss": 0.3926, "mean_token_accuracy": 0.88430255651474, "num_tokens": 369393749.0, "step": 9726 }, { "epoch": 1.237374379849892, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.51584529876709, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8694632053375244, "num_tokens": 369432442.0, "step": 9727 }, { "epoch": 1.2375015901284825, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2137484550476074, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.867544412612915, "num_tokens": 369469513.0, "step": 9728 }, { "epoch": 1.237628800407073, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5679187774658203, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8707218170166016, "num_tokens": 369505720.0, "step": 9729 }, { "epoch": 1.2377560106856633, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.154247760772705, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8767091035842896, "num_tokens": 369535225.0, "step": 9730 }, { "epoch": 1.2378832209642538, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6654624938964844, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.874024510383606, "num_tokens": 369572580.0, "step": 9731 }, { "epoch": 1.2380104312428444, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.082597494125366, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8765614628791809, "num_tokens": 369611566.0, "step": 9732 }, { "epoch": 1.238137641521435, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8264176845550537, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.871596097946167, "num_tokens": 369652541.0, "step": 9733 }, { "epoch": 1.2382648518000254, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1058173179626465, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.856950044631958, "num_tokens": 369686744.0, "step": 9734 }, { "epoch": 1.238392062078616, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9978630542755127, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8795157074928284, "num_tokens": 369725090.0, "step": 9735 }, { "epoch": 1.2385192723572065, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.196556568145752, "learning_rate": 1e-05, "loss": 0.3915, "mean_token_accuracy": 0.8880727291107178, "num_tokens": 369764255.0, "step": 9736 }, { "epoch": 1.238646482635797, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1293158531188965, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.87531977891922, "num_tokens": 369803589.0, "step": 9737 }, { "epoch": 1.2387736929143875, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4170126914978027, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8827478885650635, "num_tokens": 369837984.0, "step": 9738 }, { "epoch": 1.238900903192978, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.073453426361084, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8705942630767822, "num_tokens": 369881398.0, "step": 9739 }, { "epoch": 1.2390281134715684, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5641767978668213, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8693639039993286, "num_tokens": 369915230.0, "step": 9740 }, { "epoch": 1.239155323750159, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.895667791366577, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8701271414756775, "num_tokens": 369955237.0, "step": 9741 }, { "epoch": 1.2392825340287494, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.559439182281494, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8658161163330078, "num_tokens": 369995938.0, "step": 9742 }, { "epoch": 1.23940974430734, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7506821155548096, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8616349101066589, "num_tokens": 370037346.0, "step": 9743 }, { "epoch": 1.2395369545859305, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.525904893875122, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8691704869270325, "num_tokens": 370067222.0, "step": 9744 }, { "epoch": 1.239664164864521, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5699410438537598, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8818671703338623, "num_tokens": 370101438.0, "step": 9745 }, { "epoch": 1.2397913751431116, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.462038040161133, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.86748206615448, "num_tokens": 370136887.0, "step": 9746 }, { "epoch": 1.239918585421702, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4643285274505615, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8789141774177551, "num_tokens": 370176797.0, "step": 9747 }, { "epoch": 1.2400457957002926, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.427086353302002, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8649856448173523, "num_tokens": 370206246.0, "step": 9748 }, { "epoch": 1.2401730059788831, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.396838665008545, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8698331117630005, "num_tokens": 370246132.0, "step": 9749 }, { "epoch": 1.2403002162574737, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3635241985321045, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8759295344352722, "num_tokens": 370280039.0, "step": 9750 }, { "epoch": 1.2404274265360642, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3686516284942627, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8758955001831055, "num_tokens": 370317522.0, "step": 9751 }, { "epoch": 1.2405546368146547, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3180227279663086, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8725681304931641, "num_tokens": 370352350.0, "step": 9752 }, { "epoch": 1.2406818470932452, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3772032260894775, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8809459805488586, "num_tokens": 370382645.0, "step": 9753 }, { "epoch": 1.2408090573718356, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2792105674743652, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8744277954101562, "num_tokens": 370422679.0, "step": 9754 }, { "epoch": 1.240936267650426, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3917062282562256, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8677549362182617, "num_tokens": 370458475.0, "step": 9755 }, { "epoch": 1.2410634779290166, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022220611572265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2463815212249756, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8653610944747925, "num_tokens": 370489641.0, "step": 9756 }, { "epoch": 1.2411906882076071, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022411346435546875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4250433444976807, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8706798553466797, "num_tokens": 370526649.0, "step": 9757 }, { "epoch": 1.2413178984861977, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022411346435546875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.188915252685547, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.865487813949585, "num_tokens": 370564664.0, "step": 9758 }, { "epoch": 1.2414451087647882, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022411346435546875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4925496578216553, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.880286693572998, "num_tokens": 370598000.0, "step": 9759 }, { "epoch": 1.2415723190433787, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022411346435546875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.122885227203369, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8601935505867004, "num_tokens": 370639318.0, "step": 9760 }, { "epoch": 1.2416995293219693, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022411346435546875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6231138706207275, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.860980212688446, "num_tokens": 370676064.0, "step": 9761 }, { "epoch": 1.2418267396005598, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022411346435546875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0387768745422363, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.873732328414917, "num_tokens": 370713190.0, "step": 9762 }, { "epoch": 1.2419539498791503, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00022411346435546875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8397204875946045, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.884209156036377, "num_tokens": 370748433.0, "step": 9763 }, { "epoch": 1.2420811601577408, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022411346435546875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0167295932769775, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8776484727859497, "num_tokens": 370781522.0, "step": 9764 }, { "epoch": 1.2422083704363311, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.121008396148682, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8730295896530151, "num_tokens": 370821878.0, "step": 9765 }, { "epoch": 1.2423355807149217, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.1853437423706055, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8665001392364502, "num_tokens": 370857754.0, "step": 9766 }, { "epoch": 1.2424627909935122, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.331151485443115, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8642915487289429, "num_tokens": 370893672.0, "step": 9767 }, { "epoch": 1.2425900012721027, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5538036823272705, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8657140731811523, "num_tokens": 370928118.0, "step": 9768 }, { "epoch": 1.2427172115506933, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.241208553314209, "learning_rate": 1e-05, "loss": 0.5205, "mean_token_accuracy": 0.8529009819030762, "num_tokens": 370960748.0, "step": 9769 }, { "epoch": 1.2428444218292838, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7439377307891846, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8826471567153931, "num_tokens": 370992408.0, "step": 9770 }, { "epoch": 1.2429716321078743, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9827449321746826, "learning_rate": 1e-05, "loss": 0.5161, "mean_token_accuracy": 0.8555815815925598, "num_tokens": 371035020.0, "step": 9771 }, { "epoch": 1.2430988423864648, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6809568405151367, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8765885829925537, "num_tokens": 371072738.0, "step": 9772 }, { "epoch": 1.2432260526650554, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7949392795562744, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8777652978897095, "num_tokens": 371109246.0, "step": 9773 }, { "epoch": 1.243353262943646, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5601322650909424, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8802566528320312, "num_tokens": 371148322.0, "step": 9774 }, { "epoch": 1.2434804732222364, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.67553448677063, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.8599467277526855, "num_tokens": 371186534.0, "step": 9775 }, { "epoch": 1.243607683500827, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.451683282852173, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8707836866378784, "num_tokens": 371224973.0, "step": 9776 }, { "epoch": 1.2437348937794175, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.616750955581665, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.87190181016922, "num_tokens": 371262629.0, "step": 9777 }, { "epoch": 1.243862104058008, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.339247226715088, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8685730695724487, "num_tokens": 371298421.0, "step": 9778 }, { "epoch": 1.2439893143365983, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6537723541259766, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8473281264305115, "num_tokens": 371337393.0, "step": 9779 }, { "epoch": 1.2441165246151888, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2450921535491943, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8739413022994995, "num_tokens": 371372436.0, "step": 9780 }, { "epoch": 1.2442437348937794, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7184340953826904, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8729796409606934, "num_tokens": 371404269.0, "step": 9781 }, { "epoch": 1.24437094517237, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.168163299560547, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.872064471244812, "num_tokens": 371446494.0, "step": 9782 }, { "epoch": 1.2444981554509604, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8655662536621094, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.8631782531738281, "num_tokens": 371482524.0, "step": 9783 }, { "epoch": 1.244625365729551, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2068212032318115, "learning_rate": 1e-05, "loss": 0.5272, "mean_token_accuracy": 0.8500292897224426, "num_tokens": 371519745.0, "step": 9784 }, { "epoch": 1.2447525760081415, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9788177013397217, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8734212517738342, "num_tokens": 371563593.0, "step": 9785 }, { "epoch": 1.244879786286732, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3073065280914307, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8765746355056763, "num_tokens": 371600646.0, "step": 9786 }, { "epoch": 1.2450069965653225, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.010852336883545, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8783354759216309, "num_tokens": 371635071.0, "step": 9787 }, { "epoch": 1.245134206843913, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.4609789848327637, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8700413703918457, "num_tokens": 371672441.0, "step": 9788 }, { "epoch": 1.2452614171225034, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.878803253173828, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8759266138076782, "num_tokens": 371712751.0, "step": 9789 }, { "epoch": 1.245388627401094, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.542276620864868, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8721328377723694, "num_tokens": 371748863.0, "step": 9790 }, { "epoch": 1.2455158376796844, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6783313751220703, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8603623509407043, "num_tokens": 371784763.0, "step": 9791 }, { "epoch": 1.245643047958275, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.542638063430786, "learning_rate": 1e-05, "loss": 0.5123, "mean_token_accuracy": 0.8492993116378784, "num_tokens": 371822722.0, "step": 9792 }, { "epoch": 1.2457702582368655, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.472834587097168, "learning_rate": 1e-05, "loss": 0.384, "mean_token_accuracy": 0.8902254104614258, "num_tokens": 371855423.0, "step": 9793 }, { "epoch": 1.245897468515456, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5228288173675537, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8651334643363953, "num_tokens": 371897900.0, "step": 9794 }, { "epoch": 1.2460246787940465, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3172240257263184, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8721206188201904, "num_tokens": 371942394.0, "step": 9795 }, { "epoch": 1.246151889072637, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.53800106048584, "learning_rate": 1e-05, "loss": 0.5104, "mean_token_accuracy": 0.850310206413269, "num_tokens": 371982391.0, "step": 9796 }, { "epoch": 1.2462790993512276, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.210294246673584, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8716397285461426, "num_tokens": 372018599.0, "step": 9797 }, { "epoch": 1.2464063096298181, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.607365131378174, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8661818504333496, "num_tokens": 372058650.0, "step": 9798 }, { "epoch": 1.2465335199084087, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1257059574127197, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8652461171150208, "num_tokens": 372102272.0, "step": 9799 }, { "epoch": 1.2466607301869992, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7135813236236572, "learning_rate": 1e-05, "loss": 0.3941, "mean_token_accuracy": 0.8893766403198242, "num_tokens": 372139086.0, "step": 9800 }, { "epoch": 1.2467879404655897, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0822746753692627, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8722008466720581, "num_tokens": 372178418.0, "step": 9801 }, { "epoch": 1.2469151507441802, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.899625539779663, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8822363615036011, "num_tokens": 372218837.0, "step": 9802 }, { "epoch": 1.2470423610227706, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1328978538513184, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8552837371826172, "num_tokens": 372256620.0, "step": 9803 }, { "epoch": 1.247169571301361, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.083555221557617, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8644585609436035, "num_tokens": 372292379.0, "step": 9804 }, { "epoch": 1.2472967815799516, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.288051128387451, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8797739744186401, "num_tokens": 372336319.0, "step": 9805 }, { "epoch": 1.2474239918585421, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.157049179077148, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.8642275333404541, "num_tokens": 372372965.0, "step": 9806 }, { "epoch": 1.2475512021371327, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.521888494491577, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8726320862770081, "num_tokens": 372415453.0, "step": 9807 }, { "epoch": 1.2476784124157232, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0056304931640625, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8652648329734802, "num_tokens": 372452822.0, "step": 9808 }, { "epoch": 1.2478056226943137, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.592987537384033, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8679013252258301, "num_tokens": 372494316.0, "step": 9809 }, { "epoch": 1.2479328329729042, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8067941665649414, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8809178471565247, "num_tokens": 372531334.0, "step": 9810 }, { "epoch": 1.2480600432514948, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.566956043243408, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8782724738121033, "num_tokens": 372568303.0, "step": 9811 }, { "epoch": 1.2481872535300853, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6299705505371094, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8814864158630371, "num_tokens": 372602783.0, "step": 9812 }, { "epoch": 1.2483144638086758, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4947872161865234, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8736585974693298, "num_tokens": 372643883.0, "step": 9813 }, { "epoch": 1.2484416740872661, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.509139060974121, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8677653670310974, "num_tokens": 372681864.0, "step": 9814 }, { "epoch": 1.2485688843658567, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4260902404785156, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8669235706329346, "num_tokens": 372717657.0, "step": 9815 }, { "epoch": 1.2486960946444472, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.426253318786621, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8864734172821045, "num_tokens": 372756702.0, "step": 9816 }, { "epoch": 1.2488233049230377, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.373136281967163, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8669620752334595, "num_tokens": 372795620.0, "step": 9817 }, { "epoch": 1.2489505152016283, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.391340970993042, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8652530908584595, "num_tokens": 372833803.0, "step": 9818 }, { "epoch": 1.2490777254802188, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3293538093566895, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.87569260597229, "num_tokens": 372864229.0, "step": 9819 }, { "epoch": 1.2492049357588093, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.390490770339966, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8733274340629578, "num_tokens": 372900321.0, "step": 9820 }, { "epoch": 1.2493321460373998, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.267378091812134, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8694307208061218, "num_tokens": 372941237.0, "step": 9821 }, { "epoch": 1.2494593563159904, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4137330055236816, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8669228553771973, "num_tokens": 372980057.0, "step": 9822 }, { "epoch": 1.249586566594581, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2297630310058594, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.871013879776001, "num_tokens": 373014746.0, "step": 9823 }, { "epoch": 1.2497137768731714, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.475621223449707, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8613765835762024, "num_tokens": 373053298.0, "step": 9824 }, { "epoch": 1.249840987151762, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1452274322509766, "learning_rate": 1e-05, "loss": 0.3918, "mean_token_accuracy": 0.8857250213623047, "num_tokens": 373100155.0, "step": 9825 }, { "epoch": 1.2499681974303525, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.598374366760254, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8631942272186279, "num_tokens": 373138554.0, "step": 9826 }, { "epoch": 1.250095407708943, "ewc_loss": 0.0654296875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.291534423828125e-05, "grad_norm": 3.0900797843933105, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8581916689872742, "num_tokens": 373173023.0, "step": 9827 }, { "epoch": 1.2502226179875333, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.776416778564453, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8657674789428711, "num_tokens": 373210630.0, "step": 9828 }, { "epoch": 1.2503498282661238, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.051823854446411, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8630009293556213, "num_tokens": 373247859.0, "step": 9829 }, { "epoch": 1.2504770385447144, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 4.0071587562561035, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8715581297874451, "num_tokens": 373285663.0, "step": 9830 }, { "epoch": 1.250604248823305, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.1173949241638184, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8626558780670166, "num_tokens": 373321743.0, "step": 9831 }, { "epoch": 1.2507314591018954, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.24584436416626, "learning_rate": 1e-05, "loss": 0.5378, "mean_token_accuracy": 0.84726881980896, "num_tokens": 373355768.0, "step": 9832 }, { "epoch": 1.250858669380486, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4070656299591064, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8657345771789551, "num_tokens": 373399335.0, "step": 9833 }, { "epoch": 1.2509858796590765, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.236867904663086, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8630963563919067, "num_tokens": 373433708.0, "step": 9834 }, { "epoch": 1.251113089937667, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6627249717712402, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8745542764663696, "num_tokens": 373472145.0, "step": 9835 }, { "epoch": 1.2512403002162575, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.018637180328369, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8762083053588867, "num_tokens": 373506578.0, "step": 9836 }, { "epoch": 1.2513675104948478, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6677210330963135, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8785855770111084, "num_tokens": 373545988.0, "step": 9837 }, { "epoch": 1.2514947207734384, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8045709133148193, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8755224943161011, "num_tokens": 373581001.0, "step": 9838 }, { "epoch": 1.251621931052029, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5953166484832764, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.864600419998169, "num_tokens": 373614961.0, "step": 9839 }, { "epoch": 1.2517491413306194, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.640862464904785, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8837366700172424, "num_tokens": 373652385.0, "step": 9840 }, { "epoch": 1.25187635160921, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4840011596679688, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8651145100593567, "num_tokens": 373692468.0, "step": 9841 }, { "epoch": 1.2520035618878005, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.540130853652954, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8748482465744019, "num_tokens": 373736885.0, "step": 9842 }, { "epoch": 1.252130772166391, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.398049831390381, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8643080592155457, "num_tokens": 373777354.0, "step": 9843 }, { "epoch": 1.2522579824449815, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.511925458908081, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8616366386413574, "num_tokens": 373812400.0, "step": 9844 }, { "epoch": 1.252385192723572, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3074748516082764, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.884211003780365, "num_tokens": 373846033.0, "step": 9845 }, { "epoch": 1.2525124030021626, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5166828632354736, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8776863813400269, "num_tokens": 373887888.0, "step": 9846 }, { "epoch": 1.2526396132807531, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.220792293548584, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8837705850601196, "num_tokens": 373928082.0, "step": 9847 }, { "epoch": 1.2527668235593437, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5883402824401855, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8600103855133057, "num_tokens": 373964078.0, "step": 9848 }, { "epoch": 1.2528940338379342, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.000225067138671875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1675331592559814, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8525997996330261, "num_tokens": 373998163.0, "step": 9849 }, { "epoch": 1.2530212441165247, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.70292067527771, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8691537380218506, "num_tokens": 374036322.0, "step": 9850 }, { "epoch": 1.2531484543951152, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0927436351776123, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8656761646270752, "num_tokens": 374079144.0, "step": 9851 }, { "epoch": 1.2532756646737058, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.905324935913086, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8684467077255249, "num_tokens": 374116200.0, "step": 9852 }, { "epoch": 1.253402874952296, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.142465829849243, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8646941781044006, "num_tokens": 374151009.0, "step": 9853 }, { "epoch": 1.2535300852308866, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.092695713043213, "learning_rate": 1e-05, "loss": 0.536, "mean_token_accuracy": 0.8465322852134705, "num_tokens": 374187974.0, "step": 9854 }, { "epoch": 1.2536572955094771, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.32212233543396, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8591673970222473, "num_tokens": 374229203.0, "step": 9855 }, { "epoch": 1.2537845057880677, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.120810508728027, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8806136250495911, "num_tokens": 374261606.0, "step": 9856 }, { "epoch": 1.2539117160666582, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5349740982055664, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8626288771629333, "num_tokens": 374300274.0, "step": 9857 }, { "epoch": 1.2540389263452487, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9725499153137207, "learning_rate": 1e-05, "loss": 0.5293, "mean_token_accuracy": 0.8489712476730347, "num_tokens": 374336942.0, "step": 9858 }, { "epoch": 1.2541661366238392, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.613415241241455, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8604767322540283, "num_tokens": 374370213.0, "step": 9859 }, { "epoch": 1.2542933469024298, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7321457862854004, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8611313104629517, "num_tokens": 374410653.0, "step": 9860 }, { "epoch": 1.2544205571810203, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5634684562683105, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8566254377365112, "num_tokens": 374454449.0, "step": 9861 }, { "epoch": 1.2545477674596106, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5309059619903564, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8743274211883545, "num_tokens": 374497494.0, "step": 9862 }, { "epoch": 1.2546749777382011, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.521270513534546, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8744336366653442, "num_tokens": 374534739.0, "step": 9863 }, { "epoch": 1.2548021880167917, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.406987428665161, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8639475703239441, "num_tokens": 374573609.0, "step": 9864 }, { "epoch": 1.2549293982953822, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4892492294311523, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8761500120162964, "num_tokens": 374610269.0, "step": 9865 }, { "epoch": 1.2550566085739727, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3277552127838135, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8656426668167114, "num_tokens": 374644888.0, "step": 9866 }, { "epoch": 1.2551838188525632, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4552061557769775, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8711063861846924, "num_tokens": 374685845.0, "step": 9867 }, { "epoch": 1.2553110291311538, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2647769451141357, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8728278875350952, "num_tokens": 374723501.0, "step": 9868 }, { "epoch": 1.2554382394097443, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.470839023590088, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8629215955734253, "num_tokens": 374763082.0, "step": 9869 }, { "epoch": 1.2555654496883348, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.198204278945923, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.876153826713562, "num_tokens": 374802829.0, "step": 9870 }, { "epoch": 1.2556926599669254, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.523191213607788, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8773466944694519, "num_tokens": 374838802.0, "step": 9871 }, { "epoch": 1.255819870245516, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.125211715698242, "learning_rate": 1e-05, "loss": 0.3881, "mean_token_accuracy": 0.888360857963562, "num_tokens": 374876831.0, "step": 9872 }, { "epoch": 1.2559470805241064, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6407859325408936, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.8868008852005005, "num_tokens": 374918120.0, "step": 9873 }, { "epoch": 1.256074290802697, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.060619592666626, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8634586334228516, "num_tokens": 374958247.0, "step": 9874 }, { "epoch": 1.2562015010812875, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.83012318611145, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8639074563980103, "num_tokens": 374998461.0, "step": 9875 }, { "epoch": 1.256328711359878, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.055614709854126, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.878157377243042, "num_tokens": 375029813.0, "step": 9876 }, { "epoch": 1.2564559216384683, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0744123458862305, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8736900687217712, "num_tokens": 375070724.0, "step": 9877 }, { "epoch": 1.2565831319170588, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.1832878589630127, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8682816624641418, "num_tokens": 375110634.0, "step": 9878 }, { "epoch": 1.2567103421956494, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2521514892578125, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8778808116912842, "num_tokens": 375150054.0, "step": 9879 }, { "epoch": 1.25683755247424, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.509766101837158, "learning_rate": 1e-05, "loss": 0.5043, "mean_token_accuracy": 0.859174370765686, "num_tokens": 375185406.0, "step": 9880 }, { "epoch": 1.2569647627528304, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.147709369659424, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8763514161109924, "num_tokens": 375227544.0, "step": 9881 }, { "epoch": 1.257091973031421, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6568684577941895, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8746905326843262, "num_tokens": 375263523.0, "step": 9882 }, { "epoch": 1.2572191833100115, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9368503093719482, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8770259618759155, "num_tokens": 375304764.0, "step": 9883 }, { "epoch": 1.257346393588602, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6354267597198486, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8743948936462402, "num_tokens": 375340989.0, "step": 9884 }, { "epoch": 1.2574736038671925, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7523512840270996, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8688259124755859, "num_tokens": 375374987.0, "step": 9885 }, { "epoch": 1.2576008141457828, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5477426052093506, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8632029294967651, "num_tokens": 375415847.0, "step": 9886 }, { "epoch": 1.2577280244243734, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.598586082458496, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8827072381973267, "num_tokens": 375455761.0, "step": 9887 }, { "epoch": 1.257855234702964, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.452880620956421, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8867459297180176, "num_tokens": 375493774.0, "step": 9888 }, { "epoch": 1.2579824449815544, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.523369073867798, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8639867901802063, "num_tokens": 375531947.0, "step": 9889 }, { "epoch": 1.258109655260145, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3648436069488525, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8805643320083618, "num_tokens": 375573546.0, "step": 9890 }, { "epoch": 1.2582368655387355, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5000579357147217, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8483683466911316, "num_tokens": 375617351.0, "step": 9891 }, { "epoch": 1.258364075817326, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.295842170715332, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8685609102249146, "num_tokens": 375653939.0, "step": 9892 }, { "epoch": 1.2584912860959165, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5157899856567383, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8628283739089966, "num_tokens": 375695345.0, "step": 9893 }, { "epoch": 1.258618496374507, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.251671314239502, "learning_rate": 1e-05, "loss": 0.3874, "mean_token_accuracy": 0.8891972303390503, "num_tokens": 375731786.0, "step": 9894 }, { "epoch": 1.2587457066530976, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5440711975097656, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8703104257583618, "num_tokens": 375773258.0, "step": 9895 }, { "epoch": 1.2588729169316881, "ewc_loss": 0.06591796875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.1735615730285645, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.8835538029670715, "num_tokens": 375805712.0, "step": 9896 }, { "epoch": 1.2590001272102787, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6622369289398193, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8676955699920654, "num_tokens": 375835927.0, "step": 9897 }, { "epoch": 1.2591273374888692, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.111224412918091, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8796287178993225, "num_tokens": 375870568.0, "step": 9898 }, { "epoch": 1.2592545477674597, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8216660022735596, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8684802651405334, "num_tokens": 375904941.0, "step": 9899 }, { "epoch": 1.2593817580460502, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.112863302230835, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.8536670207977295, "num_tokens": 375941193.0, "step": 9900 }, { "epoch": 1.2595089683246408, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.994112491607666, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8760868310928345, "num_tokens": 375977952.0, "step": 9901 }, { "epoch": 1.259636178603231, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2323756217956543, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8593606948852539, "num_tokens": 376016711.0, "step": 9902 }, { "epoch": 1.2597633888818216, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.076933860778809, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8594707250595093, "num_tokens": 376061753.0, "step": 9903 }, { "epoch": 1.2598905991604121, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4308085441589355, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8613258004188538, "num_tokens": 376100148.0, "step": 9904 }, { "epoch": 1.2600178094390027, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9823055267333984, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8743141293525696, "num_tokens": 376140663.0, "step": 9905 }, { "epoch": 1.2601450197175932, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5451221466064453, "learning_rate": 1e-05, "loss": 0.5234, "mean_token_accuracy": 0.8516758680343628, "num_tokens": 376186702.0, "step": 9906 }, { "epoch": 1.2602722299961837, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7905654907226562, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8713897466659546, "num_tokens": 376222736.0, "step": 9907 }, { "epoch": 1.2603994402747742, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5524351596832275, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8565518856048584, "num_tokens": 376258871.0, "step": 9908 }, { "epoch": 1.2605266505533648, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.59289813041687, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8678905367851257, "num_tokens": 376295486.0, "step": 9909 }, { "epoch": 1.2606538608319553, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5022530555725098, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8744304180145264, "num_tokens": 376333138.0, "step": 9910 }, { "epoch": 1.2607810711105456, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4506874084472656, "learning_rate": 1e-05, "loss": 0.4125, "mean_token_accuracy": 0.8799988627433777, "num_tokens": 376367284.0, "step": 9911 }, { "epoch": 1.2609082813891361, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.495293617248535, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8583681583404541, "num_tokens": 376398318.0, "step": 9912 }, { "epoch": 1.2610354916677267, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3507094383239746, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8782435655593872, "num_tokens": 376430599.0, "step": 9913 }, { "epoch": 1.2611627019463172, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.458278179168701, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8724843263626099, "num_tokens": 376467238.0, "step": 9914 }, { "epoch": 1.2612899122249077, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.288947105407715, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8642641305923462, "num_tokens": 376505728.0, "step": 9915 }, { "epoch": 1.2614171225034982, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4551448822021484, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.851845920085907, "num_tokens": 376550387.0, "step": 9916 }, { "epoch": 1.2615443327820888, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2404325008392334, "learning_rate": 1e-05, "loss": 0.3932, "mean_token_accuracy": 0.8864386081695557, "num_tokens": 376583365.0, "step": 9917 }, { "epoch": 1.2616715430606793, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4765138626098633, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8783602714538574, "num_tokens": 376620625.0, "step": 9918 }, { "epoch": 1.2617987533392698, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1870272159576416, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8726028203964233, "num_tokens": 376657048.0, "step": 9919 }, { "epoch": 1.2619259636178604, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5400943756103516, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8649693727493286, "num_tokens": 376697000.0, "step": 9920 }, { "epoch": 1.2620531738964509, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.114223003387451, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8701560497283936, "num_tokens": 376739306.0, "step": 9921 }, { "epoch": 1.2621803841750414, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6684679985046387, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8765280842781067, "num_tokens": 376777136.0, "step": 9922 }, { "epoch": 1.262307594453632, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0530149936676025, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8788073062896729, "num_tokens": 376823875.0, "step": 9923 }, { "epoch": 1.2624348047322225, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.88096284866333, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8676640391349792, "num_tokens": 376861335.0, "step": 9924 }, { "epoch": 1.262562015010813, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.078792095184326, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8567744493484497, "num_tokens": 376896076.0, "step": 9925 }, { "epoch": 1.2626892252894033, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.097382068634033, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.861214280128479, "num_tokens": 376943761.0, "step": 9926 }, { "epoch": 1.2628164355679938, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.224113702774048, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8653510212898254, "num_tokens": 376983526.0, "step": 9927 }, { "epoch": 1.2629436458465844, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.224947452545166, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.883060097694397, "num_tokens": 377025293.0, "step": 9928 }, { "epoch": 1.263070856125175, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.486907720565796, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8695681691169739, "num_tokens": 377071380.0, "step": 9929 }, { "epoch": 1.2631980664037654, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.151303768157959, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8598910570144653, "num_tokens": 377110215.0, "step": 9930 }, { "epoch": 1.263325276682356, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.655364751815796, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8815611600875854, "num_tokens": 377150191.0, "step": 9931 }, { "epoch": 1.2634524869609465, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9507179260253906, "learning_rate": 1e-05, "loss": 0.5277, "mean_token_accuracy": 0.8490409255027771, "num_tokens": 377190636.0, "step": 9932 }, { "epoch": 1.263579697239537, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6278982162475586, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.869870126247406, "num_tokens": 377230314.0, "step": 9933 }, { "epoch": 1.2637069075181275, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7601029872894287, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8637204170227051, "num_tokens": 377270830.0, "step": 9934 }, { "epoch": 1.2638341177967178, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5504302978515625, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8519644737243652, "num_tokens": 377313862.0, "step": 9935 }, { "epoch": 1.2639613280753084, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6088767051696777, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8675096035003662, "num_tokens": 377355451.0, "step": 9936 }, { "epoch": 1.264088538353899, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4620394706726074, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8652652502059937, "num_tokens": 377397608.0, "step": 9937 }, { "epoch": 1.2642157486324894, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5263724327087402, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8702630400657654, "num_tokens": 377436268.0, "step": 9938 }, { "epoch": 1.26434295891108, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.387728452682495, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8682894110679626, "num_tokens": 377470896.0, "step": 9939 }, { "epoch": 1.2644701691896705, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4917540550231934, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8758419752120972, "num_tokens": 377508828.0, "step": 9940 }, { "epoch": 1.264597379468261, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.320704936981201, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8552836775779724, "num_tokens": 377542007.0, "step": 9941 }, { "epoch": 1.2647245897468515, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4945342540740967, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8661115169525146, "num_tokens": 377582671.0, "step": 9942 }, { "epoch": 1.264851800025442, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.230308771133423, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8729685544967651, "num_tokens": 377624840.0, "step": 9943 }, { "epoch": 1.2649790103040326, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.55501389503479, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8673391342163086, "num_tokens": 377661955.0, "step": 9944 }, { "epoch": 1.2651062205826231, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.161761999130249, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8762261867523193, "num_tokens": 377698574.0, "step": 9945 }, { "epoch": 1.2652334308612136, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.668133497238159, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8602485656738281, "num_tokens": 377739874.0, "step": 9946 }, { "epoch": 1.2653606411398042, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1087162494659424, "learning_rate": 1e-05, "loss": 0.5026, "mean_token_accuracy": 0.8526690006256104, "num_tokens": 377775644.0, "step": 9947 }, { "epoch": 1.2654878514183947, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8382575511932373, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8633509874343872, "num_tokens": 377812648.0, "step": 9948 }, { "epoch": 1.2656150616969852, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.109656810760498, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8582403659820557, "num_tokens": 377849928.0, "step": 9949 }, { "epoch": 1.2657422719755758, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.018918514251709, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8754681348800659, "num_tokens": 377891154.0, "step": 9950 }, { "epoch": 1.265869482254166, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.250077247619629, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.866722583770752, "num_tokens": 377926858.0, "step": 9951 }, { "epoch": 1.2659966925327566, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.09212589263916, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8773277997970581, "num_tokens": 377976895.0, "step": 9952 }, { "epoch": 1.2661239028113471, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.464695692062378, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8702302575111389, "num_tokens": 378010300.0, "step": 9953 }, { "epoch": 1.2662511130899377, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9786198139190674, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8689258098602295, "num_tokens": 378051764.0, "step": 9954 }, { "epoch": 1.2663783233685282, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5576281547546387, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8639193177223206, "num_tokens": 378087976.0, "step": 9955 }, { "epoch": 1.2665055336471187, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.78558611869812, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8713172674179077, "num_tokens": 378124062.0, "step": 9956 }, { "epoch": 1.2666327439257092, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5572497844696045, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8679181337356567, "num_tokens": 378161248.0, "step": 9957 }, { "epoch": 1.2667599542042998, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.599717617034912, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8615485429763794, "num_tokens": 378197959.0, "step": 9958 }, { "epoch": 1.2668871644828903, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.515259265899658, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8631868362426758, "num_tokens": 378233030.0, "step": 9959 }, { "epoch": 1.2670143747614806, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4733433723449707, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8632189035415649, "num_tokens": 378265147.0, "step": 9960 }, { "epoch": 1.2671415850400711, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.468981981277466, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.865154504776001, "num_tokens": 378298759.0, "step": 9961 }, { "epoch": 1.2672687953186617, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.373567819595337, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8530007004737854, "num_tokens": 378338456.0, "step": 9962 }, { "epoch": 1.2673960055972522, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4316906929016113, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8714570999145508, "num_tokens": 378374755.0, "step": 9963 }, { "epoch": 1.2675232158758427, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2943663597106934, "learning_rate": 1e-05, "loss": 0.3975, "mean_token_accuracy": 0.8841726779937744, "num_tokens": 378415302.0, "step": 9964 }, { "epoch": 1.2676504261544332, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4256911277770996, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8659652471542358, "num_tokens": 378454400.0, "step": 9965 }, { "epoch": 1.2677776364330238, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.256035566329956, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8715450763702393, "num_tokens": 378496986.0, "step": 9966 }, { "epoch": 1.2679048467116143, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4395785331726074, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8777616024017334, "num_tokens": 378533937.0, "step": 9967 }, { "epoch": 1.2680320569902048, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.217346429824829, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8773780465126038, "num_tokens": 378567420.0, "step": 9968 }, { "epoch": 1.2681592672687954, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.495448589324951, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8752830028533936, "num_tokens": 378600939.0, "step": 9969 }, { "epoch": 1.2682864775473859, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.139005184173584, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8620424270629883, "num_tokens": 378647858.0, "step": 9970 }, { "epoch": 1.2684136878259764, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.604301929473877, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8682942390441895, "num_tokens": 378686695.0, "step": 9971 }, { "epoch": 1.268540898104567, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.0562350749969482, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8740349411964417, "num_tokens": 378732583.0, "step": 9972 }, { "epoch": 1.2686681083831575, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7934060096740723, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8796207904815674, "num_tokens": 378768134.0, "step": 9973 }, { "epoch": 1.268795318661748, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0313172340393066, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8841615915298462, "num_tokens": 378799684.0, "step": 9974 }, { "epoch": 1.2689225289403383, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.058091640472412, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8633028864860535, "num_tokens": 378837831.0, "step": 9975 }, { "epoch": 1.2690497392189288, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.1392879486083984, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8687294721603394, "num_tokens": 378876697.0, "step": 9976 }, { "epoch": 1.2691769494975194, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.276296615600586, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8828056454658508, "num_tokens": 378916301.0, "step": 9977 }, { "epoch": 1.2693041597761099, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.4573893547058105, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8640494346618652, "num_tokens": 378956294.0, "step": 9978 }, { "epoch": 1.2694313700547004, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.251387119293213, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8626805543899536, "num_tokens": 378992704.0, "step": 9979 }, { "epoch": 1.269558580333291, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.68648099899292, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8774544596672058, "num_tokens": 379034955.0, "step": 9980 }, { "epoch": 1.2696857906118815, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.023307800292969, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8799546360969543, "num_tokens": 379075026.0, "step": 9981 }, { "epoch": 1.269813000890472, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6909825801849365, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8678979873657227, "num_tokens": 379113452.0, "step": 9982 }, { "epoch": 1.2699402111690625, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.822333574295044, "learning_rate": 1e-05, "loss": 0.5317, "mean_token_accuracy": 0.8449763059616089, "num_tokens": 379155632.0, "step": 9983 }, { "epoch": 1.2700674214476528, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022602081298828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6067941188812256, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8768518567085266, "num_tokens": 379188987.0, "step": 9984 }, { "epoch": 1.2701946317262434, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.652327060699463, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8749749660491943, "num_tokens": 379229099.0, "step": 9985 }, { "epoch": 1.270321842004834, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5032927989959717, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8813371658325195, "num_tokens": 379265365.0, "step": 9986 }, { "epoch": 1.2704490522834244, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5524561405181885, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8769223093986511, "num_tokens": 379301170.0, "step": 9987 }, { "epoch": 1.270576262562015, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.410787582397461, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8664982318878174, "num_tokens": 379334983.0, "step": 9988 }, { "epoch": 1.2707034728406055, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.512993097305298, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8592849373817444, "num_tokens": 379371584.0, "step": 9989 }, { "epoch": 1.270830683119196, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3101117610931396, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8715759515762329, "num_tokens": 379415363.0, "step": 9990 }, { "epoch": 1.2709578933977865, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5232746601104736, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8820750117301941, "num_tokens": 379447019.0, "step": 9991 }, { "epoch": 1.271085103676377, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.223674774169922, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8767938017845154, "num_tokens": 379490229.0, "step": 9992 }, { "epoch": 1.2712123139549676, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.603181838989258, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8694150447845459, "num_tokens": 379528608.0, "step": 9993 }, { "epoch": 1.2713395242335581, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1570684909820557, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8625937700271606, "num_tokens": 379566916.0, "step": 9994 }, { "epoch": 1.2714667345121486, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7325503826141357, "learning_rate": 1e-05, "loss": 0.407, "mean_token_accuracy": 0.8840363621711731, "num_tokens": 379600648.0, "step": 9995 }, { "epoch": 1.2715939447907392, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1142547130584717, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8636441826820374, "num_tokens": 379639519.0, "step": 9996 }, { "epoch": 1.2717211550693297, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.9186360836029053, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8762128949165344, "num_tokens": 379674426.0, "step": 9997 }, { "epoch": 1.2718483653479202, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.158160448074341, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8778036832809448, "num_tokens": 379713121.0, "step": 9998 }, { "epoch": 1.2719755756265108, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.081007957458496, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8772459030151367, "num_tokens": 379749952.0, "step": 9999 }, { "epoch": 1.272102785905101, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3568389415740967, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8595507144927979, "num_tokens": 379786057.0, "step": 10000 }, { "epoch": 1.2722299961836916, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.086112022399902, "learning_rate": 1e-05, "loss": 0.488, "mean_token_accuracy": 0.8621866703033447, "num_tokens": 379820034.0, "step": 10001 }, { "epoch": 1.2723572064622821, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5241339206695557, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8737785816192627, "num_tokens": 379863969.0, "step": 10002 }, { "epoch": 1.2724844167408726, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9147145748138428, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8774973154067993, "num_tokens": 379903360.0, "step": 10003 }, { "epoch": 1.2726116270194632, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.591662883758545, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8719147443771362, "num_tokens": 379935874.0, "step": 10004 }, { "epoch": 1.2727388372980537, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.694608688354492, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.865258514881134, "num_tokens": 379972525.0, "step": 10005 }, { "epoch": 1.2728660475766442, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5394885540008545, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8723592162132263, "num_tokens": 380012925.0, "step": 10006 }, { "epoch": 1.2729932578552348, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.522932767868042, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8599337339401245, "num_tokens": 380048235.0, "step": 10007 }, { "epoch": 1.2731204681338253, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5114643573760986, "learning_rate": 1e-05, "loss": 0.5231, "mean_token_accuracy": 0.8492550849914551, "num_tokens": 380084607.0, "step": 10008 }, { "epoch": 1.2732476784124156, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3941380977630615, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8592684864997864, "num_tokens": 380121514.0, "step": 10009 }, { "epoch": 1.2733748886910061, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.472003221511841, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8635393381118774, "num_tokens": 380158640.0, "step": 10010 }, { "epoch": 1.2735020989695967, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3090908527374268, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8686497211456299, "num_tokens": 380196195.0, "step": 10011 }, { "epoch": 1.2736293092481872, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.468712091445923, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8670821785926819, "num_tokens": 380234777.0, "step": 10012 }, { "epoch": 1.2737565195267777, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.249723434448242, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.863865852355957, "num_tokens": 380276494.0, "step": 10013 }, { "epoch": 1.2738837298053682, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4898288249969482, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8669936060905457, "num_tokens": 380314791.0, "step": 10014 }, { "epoch": 1.2740109400839588, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1988883018493652, "learning_rate": 1e-05, "loss": 0.4751, "mean_token_accuracy": 0.8663846254348755, "num_tokens": 380354372.0, "step": 10015 }, { "epoch": 1.2741381503625493, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.549302101135254, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8802080750465393, "num_tokens": 380388022.0, "step": 10016 }, { "epoch": 1.2742653606411398, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1326868534088135, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8628033399581909, "num_tokens": 380426665.0, "step": 10017 }, { "epoch": 1.2743925709197303, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6616618633270264, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8775254487991333, "num_tokens": 380466068.0, "step": 10018 }, { "epoch": 1.2745197811983209, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0736796855926514, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.8832300901412964, "num_tokens": 380500370.0, "step": 10019 }, { "epoch": 1.2746469914769114, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8403453826904297, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8716486692428589, "num_tokens": 380538886.0, "step": 10020 }, { "epoch": 1.274774201755502, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.0671029090881348, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8711820244789124, "num_tokens": 380577425.0, "step": 10021 }, { "epoch": 1.2749014120340925, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.079378128051758, "learning_rate": 1e-05, "loss": 0.4997, "mean_token_accuracy": 0.8559232950210571, "num_tokens": 380616999.0, "step": 10022 }, { "epoch": 1.275028622312683, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2155184745788574, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8671365976333618, "num_tokens": 380654065.0, "step": 10023 }, { "epoch": 1.2751558325912733, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.228453159332275, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8642377853393555, "num_tokens": 380687947.0, "step": 10024 }, { "epoch": 1.2752830428698638, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.4851956367492676, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8776738047599792, "num_tokens": 380728387.0, "step": 10025 }, { "epoch": 1.2754102531484544, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.157058238983154, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8629003763198853, "num_tokens": 380763906.0, "step": 10026 }, { "epoch": 1.2755374634270449, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6534290313720703, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8588349223136902, "num_tokens": 380806904.0, "step": 10027 }, { "epoch": 1.2756646737056354, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9315829277038574, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8785355091094971, "num_tokens": 380842694.0, "step": 10028 }, { "epoch": 1.275791883984226, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6275014877319336, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.879345178604126, "num_tokens": 380879624.0, "step": 10029 }, { "epoch": 1.2759190942628165, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.737520217895508, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8722556829452515, "num_tokens": 380916358.0, "step": 10030 }, { "epoch": 1.276046304541407, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5437192916870117, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8805497884750366, "num_tokens": 380957858.0, "step": 10031 }, { "epoch": 1.2761735148199975, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5930869579315186, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8766500353813171, "num_tokens": 380997301.0, "step": 10032 }, { "epoch": 1.2763007250985878, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4584217071533203, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8764181137084961, "num_tokens": 381034023.0, "step": 10033 }, { "epoch": 1.2764279353771784, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5229568481445312, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8553550243377686, "num_tokens": 381073402.0, "step": 10034 }, { "epoch": 1.2765551456557689, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3714520931243896, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8768760561943054, "num_tokens": 381111182.0, "step": 10035 }, { "epoch": 1.2766823559343594, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4836084842681885, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8709732890129089, "num_tokens": 381153527.0, "step": 10036 }, { "epoch": 1.27680956621295, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2965078353881836, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8683974146842957, "num_tokens": 381196337.0, "step": 10037 }, { "epoch": 1.2769367764915405, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5026934146881104, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8668506145477295, "num_tokens": 381239065.0, "step": 10038 }, { "epoch": 1.277063986770131, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.227001428604126, "learning_rate": 1e-05, "loss": 0.3908, "mean_token_accuracy": 0.8859309554100037, "num_tokens": 381274826.0, "step": 10039 }, { "epoch": 1.2771911970487215, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5646865367889404, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8693946003913879, "num_tokens": 381318043.0, "step": 10040 }, { "epoch": 1.277318407327312, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.159292221069336, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8686537146568298, "num_tokens": 381352583.0, "step": 10041 }, { "epoch": 1.2774456176059026, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.686189889907837, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.861587643623352, "num_tokens": 381387271.0, "step": 10042 }, { "epoch": 1.2775728278844931, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0885300636291504, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.862932562828064, "num_tokens": 381429749.0, "step": 10043 }, { "epoch": 1.2777000381630836, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.886312246322632, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8804329633712769, "num_tokens": 381468763.0, "step": 10044 }, { "epoch": 1.2778272484416742, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1097373962402344, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8730612993240356, "num_tokens": 381505789.0, "step": 10045 }, { "epoch": 1.2779544587202647, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.082435607910156, "learning_rate": 1e-05, "loss": 0.3872, "mean_token_accuracy": 0.8930374383926392, "num_tokens": 381545861.0, "step": 10046 }, { "epoch": 1.2780816689988552, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.303281784057617, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.8539934158325195, "num_tokens": 381585373.0, "step": 10047 }, { "epoch": 1.2782088792774458, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.133829116821289, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8745991587638855, "num_tokens": 381623029.0, "step": 10048 }, { "epoch": 1.278336089556036, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.521819591522217, "learning_rate": 1e-05, "loss": 0.4956, "mean_token_accuracy": 0.8621750473976135, "num_tokens": 381663481.0, "step": 10049 }, { "epoch": 1.2784632998346266, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9795265197753906, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.8879613876342773, "num_tokens": 381703977.0, "step": 10050 }, { "epoch": 1.2785905101132171, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.592817544937134, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8655246496200562, "num_tokens": 381741409.0, "step": 10051 }, { "epoch": 1.2787177203918076, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.783665418624878, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8669935464859009, "num_tokens": 381779013.0, "step": 10052 }, { "epoch": 1.2788449306703982, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5680155754089355, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8671907186508179, "num_tokens": 381813460.0, "step": 10053 }, { "epoch": 1.2789721409489887, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.588456392288208, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8639288544654846, "num_tokens": 381853821.0, "step": 10054 }, { "epoch": 1.2790993512275792, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.504676342010498, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8629806637763977, "num_tokens": 381895411.0, "step": 10055 }, { "epoch": 1.2792265615061698, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.470486879348755, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8640036582946777, "num_tokens": 381929067.0, "step": 10056 }, { "epoch": 1.2793537717847603, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002269744873046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4529049396514893, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8724049925804138, "num_tokens": 381964002.0, "step": 10057 }, { "epoch": 1.2794809820633506, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3816518783569336, "learning_rate": 1e-05, "loss": 0.3911, "mean_token_accuracy": 0.8862201571464539, "num_tokens": 382002913.0, "step": 10058 }, { "epoch": 1.2796081923419411, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.414691925048828, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8801866769790649, "num_tokens": 382039026.0, "step": 10059 }, { "epoch": 1.2797354026205316, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3309109210968018, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8715143203735352, "num_tokens": 382077413.0, "step": 10060 }, { "epoch": 1.2798626128991222, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.403229236602783, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8626802563667297, "num_tokens": 382112686.0, "step": 10061 }, { "epoch": 1.2799898231777127, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2854225635528564, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8665937185287476, "num_tokens": 382152184.0, "step": 10062 }, { "epoch": 1.2801170334563032, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4178571701049805, "learning_rate": 1e-05, "loss": 0.5055, "mean_token_accuracy": 0.8613266944885254, "num_tokens": 382190916.0, "step": 10063 }, { "epoch": 1.2802442437348938, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2530932426452637, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8577088117599487, "num_tokens": 382222986.0, "step": 10064 }, { "epoch": 1.2803714540134843, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.421433210372925, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8801150918006897, "num_tokens": 382266039.0, "step": 10065 }, { "epoch": 1.2804986642920748, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1990575790405273, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8732322454452515, "num_tokens": 382302071.0, "step": 10066 }, { "epoch": 1.2806258745706653, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4805831909179688, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8687088489532471, "num_tokens": 382340240.0, "step": 10067 }, { "epoch": 1.2807530848492559, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.154719829559326, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8730612397193909, "num_tokens": 382368308.0, "step": 10068 }, { "epoch": 1.2808802951278464, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.580474853515625, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8746281266212463, "num_tokens": 382406365.0, "step": 10069 }, { "epoch": 1.281007505406437, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0659356117248535, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.859333872795105, "num_tokens": 382449005.0, "step": 10070 }, { "epoch": 1.2811347156850275, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.769577980041504, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8655147552490234, "num_tokens": 382487744.0, "step": 10071 }, { "epoch": 1.281261925963618, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.010371685028076, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8686059713363647, "num_tokens": 382526954.0, "step": 10072 }, { "epoch": 1.2813891362422083, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.032352924346924, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8799117803573608, "num_tokens": 382565257.0, "step": 10073 }, { "epoch": 1.2815163465207988, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.100731611251831, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8653278350830078, "num_tokens": 382601992.0, "step": 10074 }, { "epoch": 1.2816435567993893, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.306698799133301, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8679624199867249, "num_tokens": 382637181.0, "step": 10075 }, { "epoch": 1.2817707670779799, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.431276321411133, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8576247096061707, "num_tokens": 382679886.0, "step": 10076 }, { "epoch": 1.2818979773565704, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.304605007171631, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.864133358001709, "num_tokens": 382717062.0, "step": 10077 }, { "epoch": 1.282025187635161, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.7187423706054688, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8661594986915588, "num_tokens": 382752188.0, "step": 10078 }, { "epoch": 1.2821523979137515, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.075615882873535, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8673214912414551, "num_tokens": 382785229.0, "step": 10079 }, { "epoch": 1.282279608192342, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.721400737762451, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8666115403175354, "num_tokens": 382823432.0, "step": 10080 }, { "epoch": 1.2824068184709325, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8534815311431885, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8605965375900269, "num_tokens": 382862763.0, "step": 10081 }, { "epoch": 1.2825340287495228, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6236984729766846, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8707141280174255, "num_tokens": 382898312.0, "step": 10082 }, { "epoch": 1.2826612390281134, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.705742120742798, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8694832921028137, "num_tokens": 382933494.0, "step": 10083 }, { "epoch": 1.2827884493067039, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.500014305114746, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8828219175338745, "num_tokens": 382971029.0, "step": 10084 }, { "epoch": 1.2829156595852944, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.632526159286499, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8612065315246582, "num_tokens": 383003911.0, "step": 10085 }, { "epoch": 1.283042869863885, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.390453577041626, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8636114597320557, "num_tokens": 383042945.0, "step": 10086 }, { "epoch": 1.2831700801424755, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5939042568206787, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8724355697631836, "num_tokens": 383085944.0, "step": 10087 }, { "epoch": 1.283297290421066, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2953343391418457, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8643532395362854, "num_tokens": 383124854.0, "step": 10088 }, { "epoch": 1.2834245006996565, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6328728199005127, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8748847246170044, "num_tokens": 383159427.0, "step": 10089 }, { "epoch": 1.283551710978247, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1996257305145264, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8727982044219971, "num_tokens": 383199320.0, "step": 10090 }, { "epoch": 1.2836789212568376, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.753892183303833, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8825709819793701, "num_tokens": 383231367.0, "step": 10091 }, { "epoch": 1.283806131535428, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.155758857727051, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8726860284805298, "num_tokens": 383268238.0, "step": 10092 }, { "epoch": 1.2839333418140186, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9212234020233154, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8772211074829102, "num_tokens": 383303445.0, "step": 10093 }, { "epoch": 1.2840605520926092, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2207961082458496, "learning_rate": 1e-05, "loss": 0.5049, "mean_token_accuracy": 0.8515737056732178, "num_tokens": 383341106.0, "step": 10094 }, { "epoch": 1.2841877623711997, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.034134864807129, "learning_rate": 1e-05, "loss": 0.501, "mean_token_accuracy": 0.8617923259735107, "num_tokens": 383379225.0, "step": 10095 }, { "epoch": 1.2843149726497902, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.353825092315674, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8744440674781799, "num_tokens": 383420798.0, "step": 10096 }, { "epoch": 1.2844421829283807, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.022603988647461, "learning_rate": 1e-05, "loss": 0.5056, "mean_token_accuracy": 0.8562588095664978, "num_tokens": 383467145.0, "step": 10097 }, { "epoch": 1.284569393206971, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5239503383636475, "learning_rate": 1e-05, "loss": 0.4945, "mean_token_accuracy": 0.8565431833267212, "num_tokens": 383505143.0, "step": 10098 }, { "epoch": 1.2846966034855616, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8469083309173584, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8618191480636597, "num_tokens": 383545987.0, "step": 10099 }, { "epoch": 1.2848238137641521, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.569912910461426, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8738629221916199, "num_tokens": 383578546.0, "step": 10100 }, { "epoch": 1.2849510240427426, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6191914081573486, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8728905916213989, "num_tokens": 383619603.0, "step": 10101 }, { "epoch": 1.2850782343213332, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5446219444274902, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8569890260696411, "num_tokens": 383660716.0, "step": 10102 }, { "epoch": 1.2852054445999237, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.427288055419922, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8738677501678467, "num_tokens": 383704344.0, "step": 10103 }, { "epoch": 1.2853326548785142, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.555649757385254, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8599228858947754, "num_tokens": 383738141.0, "step": 10104 }, { "epoch": 1.2854598651571048, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2944819927215576, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8596375584602356, "num_tokens": 383774682.0, "step": 10105 }, { "epoch": 1.2855870754356953, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5659031867980957, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8781511783599854, "num_tokens": 383814056.0, "step": 10106 }, { "epoch": 1.2857142857142856, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1911048889160156, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8588242530822754, "num_tokens": 383858151.0, "step": 10107 }, { "epoch": 1.2858414959928761, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.643069267272949, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8681045770645142, "num_tokens": 383895121.0, "step": 10108 }, { "epoch": 1.2859687062714666, "ewc_loss": 0.06640625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.133638620376587, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8562275171279907, "num_tokens": 383929273.0, "step": 10109 }, { "epoch": 1.2860959165500572, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7660915851593018, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.870474636554718, "num_tokens": 383970374.0, "step": 10110 }, { "epoch": 1.2862231268286477, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.096022605895996, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8686696290969849, "num_tokens": 384010375.0, "step": 10111 }, { "epoch": 1.2863503371072382, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.964906930923462, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.861868143081665, "num_tokens": 384040547.0, "step": 10112 }, { "epoch": 1.2864775473858288, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.18129563331604, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8635410070419312, "num_tokens": 384076005.0, "step": 10113 }, { "epoch": 1.2866047576644193, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.093816757202148, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8784240484237671, "num_tokens": 384112568.0, "step": 10114 }, { "epoch": 1.2867319679430098, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.344719648361206, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8778790831565857, "num_tokens": 384151942.0, "step": 10115 }, { "epoch": 1.2868591782216003, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.097311019897461, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8680121898651123, "num_tokens": 384192534.0, "step": 10116 }, { "epoch": 1.2869863885001909, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.519428253173828, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.875481367111206, "num_tokens": 384230600.0, "step": 10117 }, { "epoch": 1.2871135987787814, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9521372318267822, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8877023458480835, "num_tokens": 384266566.0, "step": 10118 }, { "epoch": 1.287240809057372, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5744755268096924, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.862108588218689, "num_tokens": 384304536.0, "step": 10119 }, { "epoch": 1.2873680193359625, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00022792816162109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7573084831237793, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.867637038230896, "num_tokens": 384347865.0, "step": 10120 }, { "epoch": 1.287495229614553, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.545983076095581, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8653519153594971, "num_tokens": 384380060.0, "step": 10121 }, { "epoch": 1.2876224398931433, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6090211868286133, "learning_rate": 1e-05, "loss": 0.5072, "mean_token_accuracy": 0.8541339039802551, "num_tokens": 384417574.0, "step": 10122 }, { "epoch": 1.2877496501717338, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.466306686401367, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8741282224655151, "num_tokens": 384454591.0, "step": 10123 }, { "epoch": 1.2878768604503243, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.504500389099121, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8699564933776855, "num_tokens": 384489642.0, "step": 10124 }, { "epoch": 1.2880040707289149, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.398576498031616, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8663533329963684, "num_tokens": 384528993.0, "step": 10125 }, { "epoch": 1.2881312810075054, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.459796905517578, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.866950511932373, "num_tokens": 384569321.0, "step": 10126 }, { "epoch": 1.288258491286096, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.328765630722046, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8634940385818481, "num_tokens": 384614821.0, "step": 10127 }, { "epoch": 1.2883857015646865, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4503862857818604, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8794056177139282, "num_tokens": 384658557.0, "step": 10128 }, { "epoch": 1.288512911843277, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.277216911315918, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8570573329925537, "num_tokens": 384697121.0, "step": 10129 }, { "epoch": 1.2886401221218675, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4855005741119385, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.8519541025161743, "num_tokens": 384735410.0, "step": 10130 }, { "epoch": 1.2887673324004578, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.224217414855957, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8680135011672974, "num_tokens": 384769394.0, "step": 10131 }, { "epoch": 1.2888945426790483, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5420594215393066, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8708862662315369, "num_tokens": 384808845.0, "step": 10132 }, { "epoch": 1.2890217529576389, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1492998600006104, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8686507940292358, "num_tokens": 384848129.0, "step": 10133 }, { "epoch": 1.2891489632362294, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6511428356170654, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8836209774017334, "num_tokens": 384885181.0, "step": 10134 }, { "epoch": 1.28927617351482, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0956645011901855, "learning_rate": 1e-05, "loss": 0.4121, "mean_token_accuracy": 0.8773025870323181, "num_tokens": 384923499.0, "step": 10135 }, { "epoch": 1.2894033837934105, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8293395042419434, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8790479898452759, "num_tokens": 384957367.0, "step": 10136 }, { "epoch": 1.289530594072001, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.0831120014190674, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8737702965736389, "num_tokens": 385000409.0, "step": 10137 }, { "epoch": 1.2896578043505915, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0413594245910645, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8644311428070068, "num_tokens": 385038439.0, "step": 10138 }, { "epoch": 1.289785014629182, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.234131097793579, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8711762428283691, "num_tokens": 385073950.0, "step": 10139 }, { "epoch": 1.2899122249077726, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.149909019470215, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8615759015083313, "num_tokens": 385112394.0, "step": 10140 }, { "epoch": 1.290039435186363, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.458549976348877, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8662899136543274, "num_tokens": 385149096.0, "step": 10141 }, { "epoch": 1.2901666454649536, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0609893798828125, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8637747764587402, "num_tokens": 385181986.0, "step": 10142 }, { "epoch": 1.2902938557435442, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5933542251586914, "learning_rate": 1e-05, "loss": 0.5512, "mean_token_accuracy": 0.8424564599990845, "num_tokens": 385220730.0, "step": 10143 }, { "epoch": 1.2904210660221347, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8560690879821777, "learning_rate": 1e-05, "loss": 0.5177, "mean_token_accuracy": 0.849908173084259, "num_tokens": 385262250.0, "step": 10144 }, { "epoch": 1.2905482763007252, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5656933784484863, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8629966378211975, "num_tokens": 385303518.0, "step": 10145 }, { "epoch": 1.2906754865793157, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6722238063812256, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8680049777030945, "num_tokens": 385342303.0, "step": 10146 }, { "epoch": 1.290802696857906, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5079760551452637, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.873685359954834, "num_tokens": 385381581.0, "step": 10147 }, { "epoch": 1.2909299071364966, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.538740396499634, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8780493140220642, "num_tokens": 385414835.0, "step": 10148 }, { "epoch": 1.291057117415087, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4439969062805176, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8683756589889526, "num_tokens": 385454022.0, "step": 10149 }, { "epoch": 1.2911843276936776, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4725594520568848, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8583770990371704, "num_tokens": 385486479.0, "step": 10150 }, { "epoch": 1.2913115379722682, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.373021364212036, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8600059151649475, "num_tokens": 385527942.0, "step": 10151 }, { "epoch": 1.2914387482508587, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4197545051574707, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8662993907928467, "num_tokens": 385567994.0, "step": 10152 }, { "epoch": 1.2915659585294492, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3192129135131836, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8749485611915588, "num_tokens": 385603908.0, "step": 10153 }, { "epoch": 1.2916931688080397, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.425349235534668, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8676654100418091, "num_tokens": 385641461.0, "step": 10154 }, { "epoch": 1.2918203790866303, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002288818359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2641406059265137, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8771357536315918, "num_tokens": 385676677.0, "step": 10155 }, { "epoch": 1.2919475893652206, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.467677354812622, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8682891130447388, "num_tokens": 385715674.0, "step": 10156 }, { "epoch": 1.2920747996438111, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2052605152130127, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8648316860198975, "num_tokens": 385747947.0, "step": 10157 }, { "epoch": 1.2922020099224016, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5294694900512695, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8610213994979858, "num_tokens": 385787395.0, "step": 10158 }, { "epoch": 1.2923292202009922, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1313414573669434, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8797053098678589, "num_tokens": 385821724.0, "step": 10159 }, { "epoch": 1.2924564304795827, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.648373603820801, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8898433446884155, "num_tokens": 385856281.0, "step": 10160 }, { "epoch": 1.2925836407581732, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0621087551116943, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8634271621704102, "num_tokens": 385895979.0, "step": 10161 }, { "epoch": 1.2927108510367638, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8473334312438965, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8774825930595398, "num_tokens": 385933134.0, "step": 10162 }, { "epoch": 1.2928380613153543, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0518014430999756, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8650510311126709, "num_tokens": 385974487.0, "step": 10163 }, { "epoch": 1.2929652715939448, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.105401515960693, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8594510555267334, "num_tokens": 386014594.0, "step": 10164 }, { "epoch": 1.2930924818725353, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2513391971588135, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.861105740070343, "num_tokens": 386044982.0, "step": 10165 }, { "epoch": 1.2932196921511259, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.208434581756592, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8774158954620361, "num_tokens": 386084517.0, "step": 10166 }, { "epoch": 1.2933469024297164, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.4920413494110107, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8689761757850647, "num_tokens": 386121841.0, "step": 10167 }, { "epoch": 1.293474112708307, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.127154350280762, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8709529638290405, "num_tokens": 386157426.0, "step": 10168 }, { "epoch": 1.2936013229868975, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6309103965759277, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8642619252204895, "num_tokens": 386196739.0, "step": 10169 }, { "epoch": 1.293728533265488, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9247798919677734, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8848130702972412, "num_tokens": 386234335.0, "step": 10170 }, { "epoch": 1.2938557435440783, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6124963760375977, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8678107261657715, "num_tokens": 386274166.0, "step": 10171 }, { "epoch": 1.2939829538226688, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.745710611343384, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8597269654273987, "num_tokens": 386313709.0, "step": 10172 }, { "epoch": 1.2941101641012593, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5525801181793213, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.851193368434906, "num_tokens": 386345516.0, "step": 10173 }, { "epoch": 1.2942373743798499, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.597113847732544, "learning_rate": 1e-05, "loss": 0.4123, "mean_token_accuracy": 0.8817343711853027, "num_tokens": 386383284.0, "step": 10174 }, { "epoch": 1.2943645846584404, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.463176965713501, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8623377084732056, "num_tokens": 386423141.0, "step": 10175 }, { "epoch": 1.294491794937031, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5157454013824463, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8755875825881958, "num_tokens": 386461888.0, "step": 10176 }, { "epoch": 1.2946190052156215, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.376241445541382, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8667594790458679, "num_tokens": 386500061.0, "step": 10177 }, { "epoch": 1.294746215494212, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4839038848876953, "learning_rate": 1e-05, "loss": 0.3943, "mean_token_accuracy": 0.8879885077476501, "num_tokens": 386538843.0, "step": 10178 }, { "epoch": 1.2948734257728025, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2903852462768555, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8698458671569824, "num_tokens": 386584346.0, "step": 10179 }, { "epoch": 1.2950006360513928, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.516312599182129, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8821234107017517, "num_tokens": 386619574.0, "step": 10180 }, { "epoch": 1.2951278463299833, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.226830244064331, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8704043030738831, "num_tokens": 386655111.0, "step": 10181 }, { "epoch": 1.2952550566085739, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.568624258041382, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8711051940917969, "num_tokens": 386701102.0, "step": 10182 }, { "epoch": 1.2953822668871644, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.154071092605591, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8699382543563843, "num_tokens": 386740259.0, "step": 10183 }, { "epoch": 1.295509477165755, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6939423084259033, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8769025802612305, "num_tokens": 386780345.0, "step": 10184 }, { "epoch": 1.2956366874443455, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.098971366882324, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8730728626251221, "num_tokens": 386819401.0, "step": 10185 }, { "epoch": 1.295763897722936, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.891470432281494, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8766454458236694, "num_tokens": 386856617.0, "step": 10186 }, { "epoch": 1.2958911080015265, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1240150928497314, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8578253984451294, "num_tokens": 386895426.0, "step": 10187 }, { "epoch": 1.296018318280117, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.086567401885986, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8644286394119263, "num_tokens": 386931577.0, "step": 10188 }, { "epoch": 1.2961455285587076, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2846076488494873, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8762166500091553, "num_tokens": 386968658.0, "step": 10189 }, { "epoch": 1.296272738837298, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.159289836883545, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8732764720916748, "num_tokens": 387005364.0, "step": 10190 }, { "epoch": 1.2963999491158886, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.512653350830078, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8750501275062561, "num_tokens": 387048767.0, "step": 10191 }, { "epoch": 1.2965271593944792, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.018920421600342, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8815109729766846, "num_tokens": 387083456.0, "step": 10192 }, { "epoch": 1.2966543696730697, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6142327785491943, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8750249743461609, "num_tokens": 387124220.0, "step": 10193 }, { "epoch": 1.2967815799516602, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.808321237564087, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8595247268676758, "num_tokens": 387163825.0, "step": 10194 }, { "epoch": 1.2969087902302507, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5869710445404053, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.871026337146759, "num_tokens": 387201230.0, "step": 10195 }, { "epoch": 1.297036000508841, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6134214401245117, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8665128946304321, "num_tokens": 387238426.0, "step": 10196 }, { "epoch": 1.2971632107874316, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.534090518951416, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8738224506378174, "num_tokens": 387275883.0, "step": 10197 }, { "epoch": 1.297290421066022, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4570693969726562, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8791534900665283, "num_tokens": 387319053.0, "step": 10198 }, { "epoch": 1.2974176313446126, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4841396808624268, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.878593385219574, "num_tokens": 387355013.0, "step": 10199 }, { "epoch": 1.2975448416232032, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3690779209136963, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8806328773498535, "num_tokens": 387392736.0, "step": 10200 }, { "epoch": 1.2976720519017937, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4571895599365234, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8579301238059998, "num_tokens": 387431938.0, "step": 10201 }, { "epoch": 1.2977992621803842, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.310123920440674, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.871568500995636, "num_tokens": 387472724.0, "step": 10202 }, { "epoch": 1.2979264724589747, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.448103666305542, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8709491491317749, "num_tokens": 387513666.0, "step": 10203 }, { "epoch": 1.2980536827375653, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2677574157714844, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8774169683456421, "num_tokens": 387544691.0, "step": 10204 }, { "epoch": 1.2981808930161556, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.462615966796875, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8794506788253784, "num_tokens": 387580839.0, "step": 10205 }, { "epoch": 1.298308103294746, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.22387957572937, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.862966775894165, "num_tokens": 387616258.0, "step": 10206 }, { "epoch": 1.2984353135733366, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5040738582611084, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8618935346603394, "num_tokens": 387654058.0, "step": 10207 }, { "epoch": 1.2985625238519272, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1617796421051025, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.864145040512085, "num_tokens": 387690036.0, "step": 10208 }, { "epoch": 1.2986897341305177, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.590487003326416, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8699243664741516, "num_tokens": 387729114.0, "step": 10209 }, { "epoch": 1.2988169444091082, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.083711862564087, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8728644847869873, "num_tokens": 387768225.0, "step": 10210 }, { "epoch": 1.2989441546876987, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.764098882675171, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8624504804611206, "num_tokens": 387802093.0, "step": 10211 }, { "epoch": 1.2990713649662893, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0347700119018555, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8727547526359558, "num_tokens": 387838857.0, "step": 10212 }, { "epoch": 1.2991985752448798, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.005954265594482, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8706865310668945, "num_tokens": 387878974.0, "step": 10213 }, { "epoch": 1.2993257855234703, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1153931617736816, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8626143932342529, "num_tokens": 387920157.0, "step": 10214 }, { "epoch": 1.2994529958020609, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.253462791442871, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.874646782875061, "num_tokens": 387953345.0, "step": 10215 }, { "epoch": 1.2995802060806514, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.400768280029297, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8788785934448242, "num_tokens": 387990696.0, "step": 10216 }, { "epoch": 1.299707416359242, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.268899440765381, "learning_rate": 1e-05, "loss": 0.5002, "mean_token_accuracy": 0.8610990047454834, "num_tokens": 388028142.0, "step": 10217 }, { "epoch": 1.2998346266378324, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.689786672592163, "learning_rate": 1e-05, "loss": 0.5253, "mean_token_accuracy": 0.84479820728302, "num_tokens": 388058095.0, "step": 10218 }, { "epoch": 1.299961836916423, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.046713829040527, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8687882423400879, "num_tokens": 388097141.0, "step": 10219 }, { "epoch": 1.3000890471950133, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6895523071289062, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8720014095306396, "num_tokens": 388132138.0, "step": 10220 }, { "epoch": 1.3002162574736038, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.834282636642456, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8718352317810059, "num_tokens": 388172837.0, "step": 10221 }, { "epoch": 1.3003434677521943, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5908291339874268, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8591555953025818, "num_tokens": 388214797.0, "step": 10222 }, { "epoch": 1.3004706780307849, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00022983551025390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6756205558776855, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.862004280090332, "num_tokens": 388256374.0, "step": 10223 }, { "epoch": 1.3005978883093754, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4897565841674805, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8741307258605957, "num_tokens": 388291227.0, "step": 10224 }, { "epoch": 1.300725098587966, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.576434850692749, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8794375658035278, "num_tokens": 388332963.0, "step": 10225 }, { "epoch": 1.3008523088665565, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3958969116210938, "learning_rate": 1e-05, "loss": 0.3956, "mean_token_accuracy": 0.885826826095581, "num_tokens": 388372825.0, "step": 10226 }, { "epoch": 1.300979519145147, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.542833089828491, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8723869323730469, "num_tokens": 388411850.0, "step": 10227 }, { "epoch": 1.3011067294237375, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.322222948074341, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.866007924079895, "num_tokens": 388449509.0, "step": 10228 }, { "epoch": 1.3012339397023278, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5644679069519043, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8631329536437988, "num_tokens": 388483698.0, "step": 10229 }, { "epoch": 1.3013611499809183, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2442030906677246, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8689478635787964, "num_tokens": 388516832.0, "step": 10230 }, { "epoch": 1.3014883602595089, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6191465854644775, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8798400163650513, "num_tokens": 388553272.0, "step": 10231 }, { "epoch": 1.3016155705380994, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1681385040283203, "learning_rate": 1e-05, "loss": 0.392, "mean_token_accuracy": 0.8870875835418701, "num_tokens": 388589163.0, "step": 10232 }, { "epoch": 1.30174278081669, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7450761795043945, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8617959022521973, "num_tokens": 388628305.0, "step": 10233 }, { "epoch": 1.3018699910952805, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1160991191864014, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8628313541412354, "num_tokens": 388673462.0, "step": 10234 }, { "epoch": 1.301997201373871, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9472174644470215, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8626549243927002, "num_tokens": 388709357.0, "step": 10235 }, { "epoch": 1.3021244116524615, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.184764862060547, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.8878922462463379, "num_tokens": 388746550.0, "step": 10236 }, { "epoch": 1.302251621931052, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.081820964813232, "learning_rate": 1e-05, "loss": 0.5172, "mean_token_accuracy": 0.8545597195625305, "num_tokens": 388795139.0, "step": 10237 }, { "epoch": 1.3023788322096426, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.39103364944458, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8751248717308044, "num_tokens": 388832282.0, "step": 10238 }, { "epoch": 1.302506042488233, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.037985324859619, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8784387111663818, "num_tokens": 388873974.0, "step": 10239 }, { "epoch": 1.3026332527668236, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5395619869232178, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8832831382751465, "num_tokens": 388914562.0, "step": 10240 }, { "epoch": 1.3027604630454142, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8755531311035156, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8601489663124084, "num_tokens": 388950293.0, "step": 10241 }, { "epoch": 1.3028876733240047, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5799756050109863, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8688490390777588, "num_tokens": 388986649.0, "step": 10242 }, { "epoch": 1.3030148836025952, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6413652896881104, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8823537230491638, "num_tokens": 389027779.0, "step": 10243 }, { "epoch": 1.3031420938811857, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5632007122039795, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8546974658966064, "num_tokens": 389062760.0, "step": 10244 }, { "epoch": 1.303269304159776, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.468951463699341, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8567452430725098, "num_tokens": 389101892.0, "step": 10245 }, { "epoch": 1.3033965144383666, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5286262035369873, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8714803457260132, "num_tokens": 389140155.0, "step": 10246 }, { "epoch": 1.303523724716957, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.338475227355957, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8635638952255249, "num_tokens": 389181563.0, "step": 10247 }, { "epoch": 1.3036509349955476, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.523165702819824, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.875578761100769, "num_tokens": 389217415.0, "step": 10248 }, { "epoch": 1.3037781452741382, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2448923587799072, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8838762044906616, "num_tokens": 389255889.0, "step": 10249 }, { "epoch": 1.3039053555527287, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5587549209594727, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8765575885772705, "num_tokens": 389297650.0, "step": 10250 }, { "epoch": 1.3040325658313192, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.176536798477173, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8542031049728394, "num_tokens": 389339647.0, "step": 10251 }, { "epoch": 1.3041597761099097, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6493937969207764, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8735082149505615, "num_tokens": 389376135.0, "step": 10252 }, { "epoch": 1.3042869863885003, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1304972171783447, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8616610765457153, "num_tokens": 389410516.0, "step": 10253 }, { "epoch": 1.3044141966670906, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.772027015686035, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8730154633522034, "num_tokens": 389452712.0, "step": 10254 }, { "epoch": 1.304541406945681, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.099743127822876, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8720231056213379, "num_tokens": 389489257.0, "step": 10255 }, { "epoch": 1.3046686172242716, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.965463161468506, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8565965890884399, "num_tokens": 389524344.0, "step": 10256 }, { "epoch": 1.3047958275028622, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.166050672531128, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8513824939727783, "num_tokens": 389564351.0, "step": 10257 }, { "epoch": 1.3049230377814527, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.11605978012085, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8709185719490051, "num_tokens": 389604606.0, "step": 10258 }, { "epoch": 1.3050502480600432, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.366238594055176, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8716267347335815, "num_tokens": 389639600.0, "step": 10259 }, { "epoch": 1.3051774583386337, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.114087104797363, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.8583757281303406, "num_tokens": 389677770.0, "step": 10260 }, { "epoch": 1.3053046686172243, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.549170970916748, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8564203381538391, "num_tokens": 389715366.0, "step": 10261 }, { "epoch": 1.3054318788958148, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9451122283935547, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8767855167388916, "num_tokens": 389754745.0, "step": 10262 }, { "epoch": 1.3055590891744053, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.573880672454834, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8602125644683838, "num_tokens": 389798599.0, "step": 10263 }, { "epoch": 1.3056862994529959, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.769127130508423, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8677752017974854, "num_tokens": 389831861.0, "step": 10264 }, { "epoch": 1.3058135097315864, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5303568840026855, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.858721137046814, "num_tokens": 389877785.0, "step": 10265 }, { "epoch": 1.305940720010177, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6020379066467285, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8852771520614624, "num_tokens": 389921969.0, "step": 10266 }, { "epoch": 1.3060679302887674, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4732112884521484, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8731012344360352, "num_tokens": 389959281.0, "step": 10267 }, { "epoch": 1.306195140567358, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5043556690216064, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8828153610229492, "num_tokens": 389992580.0, "step": 10268 }, { "epoch": 1.3063223508459483, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4194939136505127, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8614755868911743, "num_tokens": 390029377.0, "step": 10269 }, { "epoch": 1.3064495611245388, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.448885917663574, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8742586374282837, "num_tokens": 390059679.0, "step": 10270 }, { "epoch": 1.3065767714031293, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.361736297607422, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8692747354507446, "num_tokens": 390096751.0, "step": 10271 }, { "epoch": 1.3067039816817199, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4259212017059326, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8597984910011292, "num_tokens": 390138648.0, "step": 10272 }, { "epoch": 1.3068311919603104, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3098227977752686, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8697596788406372, "num_tokens": 390176008.0, "step": 10273 }, { "epoch": 1.306958402238901, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.423281669616699, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8677363395690918, "num_tokens": 390218120.0, "step": 10274 }, { "epoch": 1.3070856125174914, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.245591402053833, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8741244673728943, "num_tokens": 390264776.0, "step": 10275 }, { "epoch": 1.307212822796082, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4838461875915527, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8793889284133911, "num_tokens": 390298798.0, "step": 10276 }, { "epoch": 1.3073400330746725, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1975648403167725, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8662992715835571, "num_tokens": 390333332.0, "step": 10277 }, { "epoch": 1.3074672433532628, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5637168884277344, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8626079559326172, "num_tokens": 390374128.0, "step": 10278 }, { "epoch": 1.3075944536318533, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1205642223358154, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8684306740760803, "num_tokens": 390416509.0, "step": 10279 }, { "epoch": 1.3077216639104439, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.730531930923462, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8708515167236328, "num_tokens": 390448129.0, "step": 10280 }, { "epoch": 1.3078488741890344, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0632526874542236, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8692755699157715, "num_tokens": 390489329.0, "step": 10281 }, { "epoch": 1.307976084467625, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.943101644515991, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8714190125465393, "num_tokens": 390529839.0, "step": 10282 }, { "epoch": 1.3081032947462155, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1043381690979004, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8709448575973511, "num_tokens": 390568888.0, "step": 10283 }, { "epoch": 1.308230505024806, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.174344539642334, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8777393102645874, "num_tokens": 390606709.0, "step": 10284 }, { "epoch": 1.3083577153033965, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3191497325897217, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8680545091629028, "num_tokens": 390645027.0, "step": 10285 }, { "epoch": 1.308484925581987, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.243039131164551, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8701940774917603, "num_tokens": 390685923.0, "step": 10286 }, { "epoch": 1.3086121358605776, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.5842299461364746, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8847030997276306, "num_tokens": 390728410.0, "step": 10287 }, { "epoch": 1.308739346139168, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.092464447021484, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.862429678440094, "num_tokens": 390764146.0, "step": 10288 }, { "epoch": 1.3088665564177586, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.673593759536743, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8688391447067261, "num_tokens": 390800363.0, "step": 10289 }, { "epoch": 1.3089937666963491, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8641111850738525, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8714020252227783, "num_tokens": 390840213.0, "step": 10290 }, { "epoch": 1.3091209769749397, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002307891845703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.609931230545044, "learning_rate": 1e-05, "loss": 0.4038, "mean_token_accuracy": 0.8859868049621582, "num_tokens": 390879087.0, "step": 10291 }, { "epoch": 1.3092481872535302, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6908888816833496, "learning_rate": 1e-05, "loss": 0.3907, "mean_token_accuracy": 0.8876544833183289, "num_tokens": 390914074.0, "step": 10292 }, { "epoch": 1.3093753975321207, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.513228178024292, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8793896436691284, "num_tokens": 390956283.0, "step": 10293 }, { "epoch": 1.309502607810711, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.582345724105835, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.871009111404419, "num_tokens": 390996898.0, "step": 10294 }, { "epoch": 1.3096298180893016, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.424440622329712, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8820788860321045, "num_tokens": 391033280.0, "step": 10295 }, { "epoch": 1.309757028367892, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.523796796798706, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.8876659870147705, "num_tokens": 391074159.0, "step": 10296 }, { "epoch": 1.3098842386464826, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3384592533111572, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8735854625701904, "num_tokens": 391115649.0, "step": 10297 }, { "epoch": 1.3100114489250732, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5291507244110107, "learning_rate": 1e-05, "loss": 0.3864, "mean_token_accuracy": 0.8917094469070435, "num_tokens": 391147601.0, "step": 10298 }, { "epoch": 1.3101386592036637, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2631452083587646, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8704839944839478, "num_tokens": 391183483.0, "step": 10299 }, { "epoch": 1.3102658694822542, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5819737911224365, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8867402076721191, "num_tokens": 391217346.0, "step": 10300 }, { "epoch": 1.3103930797608447, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.172307252883911, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8794967532157898, "num_tokens": 391255575.0, "step": 10301 }, { "epoch": 1.3105202900394353, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.7046055793762207, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8766595125198364, "num_tokens": 391291680.0, "step": 10302 }, { "epoch": 1.3106475003180256, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1110661029815674, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8736876249313354, "num_tokens": 391330034.0, "step": 10303 }, { "epoch": 1.310774710596616, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8880507946014404, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8809936046600342, "num_tokens": 391370558.0, "step": 10304 }, { "epoch": 1.3109019208752066, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1284162998199463, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8680428862571716, "num_tokens": 391414091.0, "step": 10305 }, { "epoch": 1.3110291311537972, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.079200267791748, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8598334789276123, "num_tokens": 391458057.0, "step": 10306 }, { "epoch": 1.3111563414323877, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3104732036590576, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8734171390533447, "num_tokens": 391495419.0, "step": 10307 }, { "epoch": 1.3112835517109782, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.126407146453857, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8694608211517334, "num_tokens": 391529353.0, "step": 10308 }, { "epoch": 1.3114107619895687, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5319247245788574, "learning_rate": 1e-05, "loss": 0.5412, "mean_token_accuracy": 0.8463395833969116, "num_tokens": 391571111.0, "step": 10309 }, { "epoch": 1.3115379722681593, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.967022180557251, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8673793077468872, "num_tokens": 391612284.0, "step": 10310 }, { "epoch": 1.3116651825467498, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.605713129043579, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8631354570388794, "num_tokens": 391646934.0, "step": 10311 }, { "epoch": 1.3117923928253403, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7317914962768555, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8794738054275513, "num_tokens": 391686871.0, "step": 10312 }, { "epoch": 1.3119196031039309, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5759267807006836, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8696470260620117, "num_tokens": 391728188.0, "step": 10313 }, { "epoch": 1.3120468133825214, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.54203724861145, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8640507459640503, "num_tokens": 391767434.0, "step": 10314 }, { "epoch": 1.312174023661112, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5311434268951416, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8766716718673706, "num_tokens": 391809581.0, "step": 10315 }, { "epoch": 1.3123012339397024, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4001238346099854, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8653202056884766, "num_tokens": 391849877.0, "step": 10316 }, { "epoch": 1.312428444218293, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.524017095565796, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8610221147537231, "num_tokens": 391885995.0, "step": 10317 }, { "epoch": 1.3125556544968833, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3041670322418213, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8808151483535767, "num_tokens": 391919100.0, "step": 10318 }, { "epoch": 1.3126828647754738, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5302281379699707, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8674063682556152, "num_tokens": 391953246.0, "step": 10319 }, { "epoch": 1.3128100750540643, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.238403558731079, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.87725830078125, "num_tokens": 391986618.0, "step": 10320 }, { "epoch": 1.3129372853326549, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.556528329849243, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.866073489189148, "num_tokens": 392029663.0, "step": 10321 }, { "epoch": 1.3130644956112454, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.179945945739746, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8693037033081055, "num_tokens": 392063329.0, "step": 10322 }, { "epoch": 1.313191705889836, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.645949125289917, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8754887580871582, "num_tokens": 392101400.0, "step": 10323 }, { "epoch": 1.3133189161684264, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1106011867523193, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8670669794082642, "num_tokens": 392140146.0, "step": 10324 }, { "epoch": 1.313446126447017, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7879929542541504, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.880709171295166, "num_tokens": 392176427.0, "step": 10325 }, { "epoch": 1.3135733367256075, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.091146945953369, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8652126789093018, "num_tokens": 392221105.0, "step": 10326 }, { "epoch": 1.3137005470041978, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9893875122070312, "learning_rate": 1e-05, "loss": 0.5113, "mean_token_accuracy": 0.8540269732475281, "num_tokens": 392257301.0, "step": 10327 }, { "epoch": 1.3138277572827883, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.1726226806640625, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8737820386886597, "num_tokens": 392295849.0, "step": 10328 }, { "epoch": 1.3139549675613789, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.150223731994629, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8604884147644043, "num_tokens": 392336480.0, "step": 10329 }, { "epoch": 1.3140821778399694, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.398469924926758, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8633859157562256, "num_tokens": 392372562.0, "step": 10330 }, { "epoch": 1.31420938811856, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1238627433776855, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8816481232643127, "num_tokens": 392408843.0, "step": 10331 }, { "epoch": 1.3143365983971504, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.5727758407592773, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8724054098129272, "num_tokens": 392440725.0, "step": 10332 }, { "epoch": 1.314463808675741, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9488022327423096, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.8583272695541382, "num_tokens": 392484643.0, "step": 10333 }, { "epoch": 1.3145910189543315, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5902841091156006, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8735243082046509, "num_tokens": 392523849.0, "step": 10334 }, { "epoch": 1.314718229232922, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.758200168609619, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.871944785118103, "num_tokens": 392559598.0, "step": 10335 }, { "epoch": 1.3148454395115126, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5378313064575195, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.8516374826431274, "num_tokens": 392603856.0, "step": 10336 }, { "epoch": 1.314972649790103, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5990710258483887, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8789010047912598, "num_tokens": 392645317.0, "step": 10337 }, { "epoch": 1.3150998600686936, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.481395959854126, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8682588338851929, "num_tokens": 392681650.0, "step": 10338 }, { "epoch": 1.3152270703472841, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5118939876556396, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8620138168334961, "num_tokens": 392711588.0, "step": 10339 }, { "epoch": 1.3153542806258747, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4011497497558594, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8751472234725952, "num_tokens": 392758504.0, "step": 10340 }, { "epoch": 1.3154814909044652, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4682676792144775, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.874991238117218, "num_tokens": 392792153.0, "step": 10341 }, { "epoch": 1.3156087011830557, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.32057785987854, "learning_rate": 1e-05, "loss": 0.4046, "mean_token_accuracy": 0.8830599784851074, "num_tokens": 392840971.0, "step": 10342 }, { "epoch": 1.315735911461646, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.462799549102783, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8649539947509766, "num_tokens": 392882760.0, "step": 10343 }, { "epoch": 1.3158631217402366, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.275861978530884, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8584866523742676, "num_tokens": 392922165.0, "step": 10344 }, { "epoch": 1.315990332018827, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.503858804702759, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8619517683982849, "num_tokens": 392957181.0, "step": 10345 }, { "epoch": 1.3161175422974176, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2098796367645264, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8637401461601257, "num_tokens": 392998405.0, "step": 10346 }, { "epoch": 1.3162447525760081, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.570603132247925, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8748458623886108, "num_tokens": 393036437.0, "step": 10347 }, { "epoch": 1.3163719628545987, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.140681028366089, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.877918541431427, "num_tokens": 393070711.0, "step": 10348 }, { "epoch": 1.3164991731331892, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.708777666091919, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8550035357475281, "num_tokens": 393109052.0, "step": 10349 }, { "epoch": 1.3166263834117797, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.092684268951416, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8612748980522156, "num_tokens": 393146244.0, "step": 10350 }, { "epoch": 1.3167535936903703, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.894688129425049, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.873031497001648, "num_tokens": 393180630.0, "step": 10351 }, { "epoch": 1.3168808039689606, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.11245059967041, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8695184588432312, "num_tokens": 393222687.0, "step": 10352 }, { "epoch": 1.317008014247551, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0889716148376465, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.865626335144043, "num_tokens": 393267383.0, "step": 10353 }, { "epoch": 1.3171352245261416, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.303131580352783, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8781093955039978, "num_tokens": 393302055.0, "step": 10354 }, { "epoch": 1.3172624348047322, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.142494201660156, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8820264339447021, "num_tokens": 393336863.0, "step": 10355 }, { "epoch": 1.3173896450833227, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.528074264526367, "learning_rate": 1e-05, "loss": 0.5199, "mean_token_accuracy": 0.8486464023590088, "num_tokens": 393373602.0, "step": 10356 }, { "epoch": 1.3175168553619132, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9998619556427, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8604484796524048, "num_tokens": 393409858.0, "step": 10357 }, { "epoch": 1.3176440656405037, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5860774517059326, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8767917156219482, "num_tokens": 393451198.0, "step": 10358 }, { "epoch": 1.3177712759190943, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7949864864349365, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8814991116523743, "num_tokens": 393486916.0, "step": 10359 }, { "epoch": 1.3178984861976848, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5658652782440186, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8838295936584473, "num_tokens": 393520039.0, "step": 10360 }, { "epoch": 1.3180256964762753, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.609729290008545, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.859671413898468, "num_tokens": 393557960.0, "step": 10361 }, { "epoch": 1.3181529067548658, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5205323696136475, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.858553409576416, "num_tokens": 393592697.0, "step": 10362 }, { "epoch": 1.3182801170334564, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.465380907058716, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8611581325531006, "num_tokens": 393635423.0, "step": 10363 }, { "epoch": 1.318407327312047, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.468083143234253, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.877204418182373, "num_tokens": 393673508.0, "step": 10364 }, { "epoch": 1.3185345375906374, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.364743232727051, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8861932754516602, "num_tokens": 393711974.0, "step": 10365 }, { "epoch": 1.318661747869228, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.453253746032715, "learning_rate": 1e-05, "loss": 0.5165, "mean_token_accuracy": 0.8492481708526611, "num_tokens": 393752670.0, "step": 10366 }, { "epoch": 1.3187889581478183, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.305217981338501, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8669572472572327, "num_tokens": 393794312.0, "step": 10367 }, { "epoch": 1.3189161684264088, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4479801654815674, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8570945858955383, "num_tokens": 393833830.0, "step": 10368 }, { "epoch": 1.3190433787049993, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.257411003112793, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8725332021713257, "num_tokens": 393874844.0, "step": 10369 }, { "epoch": 1.3191705889835899, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4796862602233887, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8552950620651245, "num_tokens": 393913861.0, "step": 10370 }, { "epoch": 1.3192977992621804, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2100679874420166, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8636167049407959, "num_tokens": 393953014.0, "step": 10371 }, { "epoch": 1.319425009540771, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5306396484375, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8706965446472168, "num_tokens": 393995603.0, "step": 10372 }, { "epoch": 1.3195522198193614, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1353542804718018, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8773976564407349, "num_tokens": 394032127.0, "step": 10373 }, { "epoch": 1.319679430097952, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.643002986907959, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.8870307207107544, "num_tokens": 394067153.0, "step": 10374 }, { "epoch": 1.3198066403765425, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0689697265625, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8713787794113159, "num_tokens": 394103484.0, "step": 10375 }, { "epoch": 1.3199338506551328, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.835714817047119, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8761510848999023, "num_tokens": 394144302.0, "step": 10376 }, { "epoch": 1.3200610609337233, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.050307273864746, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.867169976234436, "num_tokens": 394179258.0, "step": 10377 }, { "epoch": 1.3201882712123139, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.090112209320068, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8615854382514954, "num_tokens": 394221498.0, "step": 10378 }, { "epoch": 1.3203154814909044, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.1898550987243652, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8812553882598877, "num_tokens": 394255985.0, "step": 10379 }, { "epoch": 1.320442691769495, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.276232719421387, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8562084436416626, "num_tokens": 394291020.0, "step": 10380 }, { "epoch": 1.3205699020480854, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.4932031631469727, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8654384613037109, "num_tokens": 394329415.0, "step": 10381 }, { "epoch": 1.320697112326676, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.211874961853027, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8676866292953491, "num_tokens": 394365287.0, "step": 10382 }, { "epoch": 1.3208243226052665, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6872780323028564, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8737635612487793, "num_tokens": 394403749.0, "step": 10383 }, { "epoch": 1.320951532883857, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9798641204833984, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8878161311149597, "num_tokens": 394445443.0, "step": 10384 }, { "epoch": 1.3210787431624476, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.646679401397705, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8862415552139282, "num_tokens": 394487650.0, "step": 10385 }, { "epoch": 1.321205953441038, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.80709171295166, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8793487548828125, "num_tokens": 394524606.0, "step": 10386 }, { "epoch": 1.3213331637196286, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.581462860107422, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8622516393661499, "num_tokens": 394561177.0, "step": 10387 }, { "epoch": 1.3214603739982191, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7095119953155518, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8810553550720215, "num_tokens": 394605212.0, "step": 10388 }, { "epoch": 1.3215875842768097, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.484299421310425, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8717460632324219, "num_tokens": 394637662.0, "step": 10389 }, { "epoch": 1.3217147945554002, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.582439661026001, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8780323266983032, "num_tokens": 394675551.0, "step": 10390 }, { "epoch": 1.3218420048339907, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3862082958221436, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8547316193580627, "num_tokens": 394715225.0, "step": 10391 }, { "epoch": 1.321969215112581, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5509390830993652, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8804500699043274, "num_tokens": 394749872.0, "step": 10392 }, { "epoch": 1.3220964253911716, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3108136653900146, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8577046394348145, "num_tokens": 394791674.0, "step": 10393 }, { "epoch": 1.322223635669762, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.571500301361084, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.859972357749939, "num_tokens": 394829556.0, "step": 10394 }, { "epoch": 1.3223508459483526, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2306902408599854, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8736398220062256, "num_tokens": 394864421.0, "step": 10395 }, { "epoch": 1.3224780562269431, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.645216941833496, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8741477131843567, "num_tokens": 394901008.0, "step": 10396 }, { "epoch": 1.3226052665055337, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1574831008911133, "learning_rate": 1e-05, "loss": 0.5435, "mean_token_accuracy": 0.8440480828285217, "num_tokens": 394947172.0, "step": 10397 }, { "epoch": 1.3227324767841242, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.782600164413452, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8617079854011536, "num_tokens": 394984515.0, "step": 10398 }, { "epoch": 1.3228596870627147, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1484501361846924, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8592153787612915, "num_tokens": 395019819.0, "step": 10399 }, { "epoch": 1.3229868973413053, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9461874961853027, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.868471622467041, "num_tokens": 395058606.0, "step": 10400 }, { "epoch": 1.3231141076198956, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.212257146835327, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8786019086837769, "num_tokens": 395094354.0, "step": 10401 }, { "epoch": 1.323241317898486, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.070782661437988, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8693970441818237, "num_tokens": 395136034.0, "step": 10402 }, { "epoch": 1.3233685281770766, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3855981826782227, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8863466382026672, "num_tokens": 395176335.0, "step": 10403 }, { "epoch": 1.3234957384556671, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0301666259765625, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.880567193031311, "num_tokens": 395216297.0, "step": 10404 }, { "epoch": 1.3236229487342577, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.536907196044922, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.872710108757019, "num_tokens": 395255685.0, "step": 10405 }, { "epoch": 1.3237501590128482, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8430135250091553, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8617215752601624, "num_tokens": 395300484.0, "step": 10406 }, { "epoch": 1.3238773692914387, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5883400440216064, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8656272888183594, "num_tokens": 395333760.0, "step": 10407 }, { "epoch": 1.3240045795700293, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6158149242401123, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.868046224117279, "num_tokens": 395376264.0, "step": 10408 }, { "epoch": 1.3241317898486198, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5544137954711914, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8577052354812622, "num_tokens": 395417758.0, "step": 10409 }, { "epoch": 1.3242590001272103, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.430975914001465, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8648250102996826, "num_tokens": 395458657.0, "step": 10410 }, { "epoch": 1.3243862104058008, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.53711199760437, "learning_rate": 1e-05, "loss": 0.4956, "mean_token_accuracy": 0.8569327592849731, "num_tokens": 395500838.0, "step": 10411 }, { "epoch": 1.3245134206843914, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.300102472305298, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8673532605171204, "num_tokens": 395544912.0, "step": 10412 }, { "epoch": 1.324640630962982, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.554448127746582, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8714380860328674, "num_tokens": 395584661.0, "step": 10413 }, { "epoch": 1.3247678412415724, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2125794887542725, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8701592683792114, "num_tokens": 395623374.0, "step": 10414 }, { "epoch": 1.324895051520163, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6241371631622314, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8803812861442566, "num_tokens": 395658854.0, "step": 10415 }, { "epoch": 1.3250222617987533, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.146944522857666, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8772659301757812, "num_tokens": 395695298.0, "step": 10416 }, { "epoch": 1.3251494720773438, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7444052696228027, "learning_rate": 1e-05, "loss": 0.522, "mean_token_accuracy": 0.8470968008041382, "num_tokens": 395731800.0, "step": 10417 }, { "epoch": 1.3252766823559343, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.102987766265869, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.86077880859375, "num_tokens": 395773811.0, "step": 10418 }, { "epoch": 1.3254038926345248, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9099738597869873, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8745061755180359, "num_tokens": 395817280.0, "step": 10419 }, { "epoch": 1.3255311029131154, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.142835855484009, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8632855415344238, "num_tokens": 395857950.0, "step": 10420 }, { "epoch": 1.325658313191706, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.088298320770264, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8603768944740295, "num_tokens": 395893048.0, "step": 10421 }, { "epoch": 1.3257855234702964, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3089423179626465, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8748979568481445, "num_tokens": 395932678.0, "step": 10422 }, { "epoch": 1.325912733748887, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.129036903381348, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8747813701629639, "num_tokens": 395967253.0, "step": 10423 }, { "epoch": 1.3260399440274775, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5063672065734863, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8864637613296509, "num_tokens": 396000935.0, "step": 10424 }, { "epoch": 1.3261671543060678, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.005799293518066, "learning_rate": 1e-05, "loss": 0.5066, "mean_token_accuracy": 0.8549473285675049, "num_tokens": 396039148.0, "step": 10425 }, { "epoch": 1.3262943645846583, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.587960958480835, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.883438766002655, "num_tokens": 396072035.0, "step": 10426 }, { "epoch": 1.3264215748632489, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8101603984832764, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8646196126937866, "num_tokens": 396112047.0, "step": 10427 }, { "epoch": 1.3265487851418394, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.557121992111206, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8693543076515198, "num_tokens": 396149844.0, "step": 10428 }, { "epoch": 1.32667599542043, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.632237434387207, "learning_rate": 1e-05, "loss": 0.3857, "mean_token_accuracy": 0.8928428888320923, "num_tokens": 396187493.0, "step": 10429 }, { "epoch": 1.3268032056990204, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.510683536529541, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8708443641662598, "num_tokens": 396220590.0, "step": 10430 }, { "epoch": 1.326930415977611, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.509798526763916, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8664882183074951, "num_tokens": 396262590.0, "step": 10431 }, { "epoch": 1.3270576262562015, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4454874992370605, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8738622665405273, "num_tokens": 396296491.0, "step": 10432 }, { "epoch": 1.327184836534792, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4327430725097656, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8726307153701782, "num_tokens": 396334798.0, "step": 10433 }, { "epoch": 1.3273120468133826, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3946404457092285, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8678842782974243, "num_tokens": 396368278.0, "step": 10434 }, { "epoch": 1.327439257091973, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3952038288116455, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.873842716217041, "num_tokens": 396405034.0, "step": 10435 }, { "epoch": 1.3275664673705636, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3511672019958496, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8556661605834961, "num_tokens": 396445503.0, "step": 10436 }, { "epoch": 1.3276936776491541, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3894433975219727, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8748981952667236, "num_tokens": 396482462.0, "step": 10437 }, { "epoch": 1.3278208879277447, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.29585337638855, "learning_rate": 1e-05, "loss": 0.3819, "mean_token_accuracy": 0.8884149193763733, "num_tokens": 396519186.0, "step": 10438 }, { "epoch": 1.3279480982063352, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3972156047821045, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8696297407150269, "num_tokens": 396558996.0, "step": 10439 }, { "epoch": 1.3280753084849257, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.279989004135132, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8522194623947144, "num_tokens": 396593158.0, "step": 10440 }, { "epoch": 1.328202518763516, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4205408096313477, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8621299266815186, "num_tokens": 396629189.0, "step": 10441 }, { "epoch": 1.3283297290421066, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.231776237487793, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8520603775978088, "num_tokens": 396665932.0, "step": 10442 }, { "epoch": 1.328456939320697, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023174285888671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4689950942993164, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8567186594009399, "num_tokens": 396706638.0, "step": 10443 }, { "epoch": 1.3285841495992876, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.166973114013672, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8742426633834839, "num_tokens": 396744354.0, "step": 10444 }, { "epoch": 1.3287113598778781, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.590256929397583, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8721630573272705, "num_tokens": 396783806.0, "step": 10445 }, { "epoch": 1.3288385701564687, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0799264907836914, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8705976009368896, "num_tokens": 396822337.0, "step": 10446 }, { "epoch": 1.3289657804350592, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.774149179458618, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8553496599197388, "num_tokens": 396861674.0, "step": 10447 }, { "epoch": 1.3290929907136497, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0269994735717773, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8646576404571533, "num_tokens": 396897686.0, "step": 10448 }, { "epoch": 1.3292202009922403, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.046331882476807, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8627923727035522, "num_tokens": 396936261.0, "step": 10449 }, { "epoch": 1.3293474112708306, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.1182875633239746, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8640316724777222, "num_tokens": 396967452.0, "step": 10450 }, { "epoch": 1.329474621549421, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.297320365905762, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8849552273750305, "num_tokens": 397004672.0, "step": 10451 }, { "epoch": 1.3296018318280116, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.421302556991577, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.867652416229248, "num_tokens": 397039310.0, "step": 10452 }, { "epoch": 1.3297290421066021, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.328316688537598, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8774861097335815, "num_tokens": 397077465.0, "step": 10453 }, { "epoch": 1.3298562523851927, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.7229881286621094, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8762071132659912, "num_tokens": 397108606.0, "step": 10454 }, { "epoch": 1.3299834626637832, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.106295108795166, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8655829429626465, "num_tokens": 397146628.0, "step": 10455 }, { "epoch": 1.3301106729423737, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7386531829833984, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8702445030212402, "num_tokens": 397182345.0, "step": 10456 }, { "epoch": 1.3302378832209643, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.890986442565918, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8705980777740479, "num_tokens": 397218673.0, "step": 10457 }, { "epoch": 1.3303650934995548, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6443471908569336, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8783054351806641, "num_tokens": 397250186.0, "step": 10458 }, { "epoch": 1.3304923037781453, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7248635292053223, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8681380152702332, "num_tokens": 397287947.0, "step": 10459 }, { "epoch": 1.3306195140567358, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5193562507629395, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8676052093505859, "num_tokens": 397328523.0, "step": 10460 }, { "epoch": 1.3307467243353264, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.629175901412964, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8715026378631592, "num_tokens": 397366238.0, "step": 10461 }, { "epoch": 1.330873934613917, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.408024549484253, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8840270042419434, "num_tokens": 397404324.0, "step": 10462 }, { "epoch": 1.3310011448925074, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.594733238220215, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.86952805519104, "num_tokens": 397441074.0, "step": 10463 }, { "epoch": 1.331128355171098, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.310096502304077, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8590468168258667, "num_tokens": 397476913.0, "step": 10464 }, { "epoch": 1.3312555654496883, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6362149715423584, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8743557929992676, "num_tokens": 397507569.0, "step": 10465 }, { "epoch": 1.3313827757282788, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2194252014160156, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.876464307308197, "num_tokens": 397544131.0, "step": 10466 }, { "epoch": 1.3315099860068693, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.722155809402466, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8865257501602173, "num_tokens": 397585042.0, "step": 10467 }, { "epoch": 1.3316371962854598, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.167917013168335, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8639862537384033, "num_tokens": 397620169.0, "step": 10468 }, { "epoch": 1.3317644065640504, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.888554096221924, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8666619062423706, "num_tokens": 397657349.0, "step": 10469 }, { "epoch": 1.331891616842641, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.191828966140747, "learning_rate": 1e-05, "loss": 0.5087, "mean_token_accuracy": 0.8530608415603638, "num_tokens": 397696309.0, "step": 10470 }, { "epoch": 1.3320188271212314, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.034815788269043, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8763204216957092, "num_tokens": 397734520.0, "step": 10471 }, { "epoch": 1.332146037399822, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.329225778579712, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.861854076385498, "num_tokens": 397773078.0, "step": 10472 }, { "epoch": 1.3322732476784125, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.054080486297607, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8805657625198364, "num_tokens": 397814180.0, "step": 10473 }, { "epoch": 1.3324004579570028, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.498661756515503, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8795589208602905, "num_tokens": 397850618.0, "step": 10474 }, { "epoch": 1.3325276682355933, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.925976037979126, "learning_rate": 1e-05, "loss": 0.513, "mean_token_accuracy": 0.8542104959487915, "num_tokens": 397888538.0, "step": 10475 }, { "epoch": 1.3326548785141838, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.58920955657959, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8804343938827515, "num_tokens": 397917861.0, "step": 10476 }, { "epoch": 1.3327820887927744, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6825528144836426, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8872936964035034, "num_tokens": 397954052.0, "step": 10477 }, { "epoch": 1.332909299071365, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5585076808929443, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8700941801071167, "num_tokens": 397996458.0, "step": 10478 }, { "epoch": 1.3330365093499554, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.485597610473633, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8807053565979004, "num_tokens": 398034164.0, "step": 10479 }, { "epoch": 1.333163719628546, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5393030643463135, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8802571296691895, "num_tokens": 398075828.0, "step": 10480 }, { "epoch": 1.3332909299071365, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.349674701690674, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8701716661453247, "num_tokens": 398112941.0, "step": 10481 }, { "epoch": 1.333418140185727, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5580620765686035, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8803240060806274, "num_tokens": 398146255.0, "step": 10482 }, { "epoch": 1.3335453504643175, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.236987352371216, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8816877007484436, "num_tokens": 398178552.0, "step": 10483 }, { "epoch": 1.333672560742908, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.600023031234741, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8818351030349731, "num_tokens": 398216310.0, "step": 10484 }, { "epoch": 1.3337997710214986, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1692955493927, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8707376718521118, "num_tokens": 398249756.0, "step": 10485 }, { "epoch": 1.3339269813000891, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.697937488555908, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8692351579666138, "num_tokens": 398285264.0, "step": 10486 }, { "epoch": 1.3340541915786797, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1198019981384277, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8534091711044312, "num_tokens": 398320964.0, "step": 10487 }, { "epoch": 1.3341814018572702, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.851402759552002, "learning_rate": 1e-05, "loss": 0.5158, "mean_token_accuracy": 0.8526648283004761, "num_tokens": 398355392.0, "step": 10488 }, { "epoch": 1.3343086121358605, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1233859062194824, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8679540157318115, "num_tokens": 398393551.0, "step": 10489 }, { "epoch": 1.334435822414451, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.023875713348389, "learning_rate": 1e-05, "loss": 0.5085, "mean_token_accuracy": 0.8550307750701904, "num_tokens": 398433812.0, "step": 10490 }, { "epoch": 1.3345630326930416, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2440996170043945, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8695884943008423, "num_tokens": 398469790.0, "step": 10491 }, { "epoch": 1.334690242971632, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.106088161468506, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8755375742912292, "num_tokens": 398512253.0, "step": 10492 }, { "epoch": 1.3348174532502226, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.4373011589050293, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.849907636642456, "num_tokens": 398549790.0, "step": 10493 }, { "epoch": 1.3349446635288131, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.034664154052734, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8844039440155029, "num_tokens": 398586844.0, "step": 10494 }, { "epoch": 1.3350718738074037, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.5509159564971924, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.867689847946167, "num_tokens": 398624177.0, "step": 10495 }, { "epoch": 1.3351990840859942, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8672237396240234, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8808873891830444, "num_tokens": 398661840.0, "step": 10496 }, { "epoch": 1.3353262943645847, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5517399311065674, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8815531134605408, "num_tokens": 398699535.0, "step": 10497 }, { "epoch": 1.3354535046431752, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.691581964492798, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8601610660552979, "num_tokens": 398736294.0, "step": 10498 }, { "epoch": 1.3355807149217656, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5085384845733643, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8773544430732727, "num_tokens": 398776372.0, "step": 10499 }, { "epoch": 1.335707925200356, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5474517345428467, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8658007383346558, "num_tokens": 398816000.0, "step": 10500 }, { "epoch": 1.3358351354789466, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4618940353393555, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8801699876785278, "num_tokens": 398851960.0, "step": 10501 }, { "epoch": 1.3359623457575371, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.46268367767334, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8683152198791504, "num_tokens": 398886485.0, "step": 10502 }, { "epoch": 1.3360895560361277, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.416893720626831, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.862701416015625, "num_tokens": 398923800.0, "step": 10503 }, { "epoch": 1.3362167663147182, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4003567695617676, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8738957643508911, "num_tokens": 398963578.0, "step": 10504 }, { "epoch": 1.3363439765933087, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3555562496185303, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.873401403427124, "num_tokens": 399009791.0, "step": 10505 }, { "epoch": 1.3364711868718993, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.387916088104248, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.879170835018158, "num_tokens": 399048701.0, "step": 10506 }, { "epoch": 1.3365983971504898, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3197715282440186, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8796089291572571, "num_tokens": 399083441.0, "step": 10507 }, { "epoch": 1.3367256074290803, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.410921335220337, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8760969638824463, "num_tokens": 399113762.0, "step": 10508 }, { "epoch": 1.3368528177076708, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.273972511291504, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8784840703010559, "num_tokens": 399149966.0, "step": 10509 }, { "epoch": 1.3369800279862614, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4270284175872803, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.868102490901947, "num_tokens": 399188777.0, "step": 10510 }, { "epoch": 1.337107238264852, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.000232696533203125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.210289478302002, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8798812627792358, "num_tokens": 399229618.0, "step": 10511 }, { "epoch": 1.3372344485434424, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.497758626937866, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8767924308776855, "num_tokens": 399270558.0, "step": 10512 }, { "epoch": 1.337361658822033, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1537842750549316, "learning_rate": 1e-05, "loss": 0.5425, "mean_token_accuracy": 0.8401280641555786, "num_tokens": 399310660.0, "step": 10513 }, { "epoch": 1.3374888691006233, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6142711639404297, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8671343326568604, "num_tokens": 399347838.0, "step": 10514 }, { "epoch": 1.3376160793792138, "ewc_loss": 0.06689453125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0790045261383057, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8756400942802429, "num_tokens": 399384790.0, "step": 10515 }, { "epoch": 1.3377432896578043, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7867655754089355, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8787372708320618, "num_tokens": 399425479.0, "step": 10516 }, { "epoch": 1.3378704999363948, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.040501356124878, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.871961772441864, "num_tokens": 399467161.0, "step": 10517 }, { "epoch": 1.3379977102149854, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.057124614715576, "learning_rate": 1e-05, "loss": 0.5022, "mean_token_accuracy": 0.8606597185134888, "num_tokens": 399500204.0, "step": 10518 }, { "epoch": 1.338124920493576, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1377532482147217, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8557248115539551, "num_tokens": 399546142.0, "step": 10519 }, { "epoch": 1.3382521307721664, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.299101829528809, "learning_rate": 1e-05, "loss": 0.525, "mean_token_accuracy": 0.8527380228042603, "num_tokens": 399582446.0, "step": 10520 }, { "epoch": 1.338379341050757, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4898366928100586, "learning_rate": 1e-05, "loss": 0.5549, "mean_token_accuracy": 0.8417412042617798, "num_tokens": 399616998.0, "step": 10521 }, { "epoch": 1.3385065513293475, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.246745586395264, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8749834895133972, "num_tokens": 399648649.0, "step": 10522 }, { "epoch": 1.3386337616079378, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6670053005218506, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8846591711044312, "num_tokens": 399690417.0, "step": 10523 }, { "epoch": 1.3387609718865283, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.043292045593262, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8641691207885742, "num_tokens": 399732346.0, "step": 10524 }, { "epoch": 1.3388881821651188, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6827239990234375, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8602602481842041, "num_tokens": 399770516.0, "step": 10525 }, { "epoch": 1.3390153924437094, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.857090950012207, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8680499792098999, "num_tokens": 399803266.0, "step": 10526 }, { "epoch": 1.3391426027223, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.589890241622925, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.867599606513977, "num_tokens": 399841940.0, "step": 10527 }, { "epoch": 1.3392698130008904, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.710552215576172, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.859355092048645, "num_tokens": 399880207.0, "step": 10528 }, { "epoch": 1.339397023279481, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4940574169158936, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8607080578804016, "num_tokens": 399915731.0, "step": 10529 }, { "epoch": 1.3395242335580715, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6238059997558594, "learning_rate": 1e-05, "loss": 0.3767, "mean_token_accuracy": 0.8941185474395752, "num_tokens": 399953398.0, "step": 10530 }, { "epoch": 1.339651443836662, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.420393705368042, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8536137938499451, "num_tokens": 399991132.0, "step": 10531 }, { "epoch": 1.3397786541152525, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.577526807785034, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8637607097625732, "num_tokens": 400031615.0, "step": 10532 }, { "epoch": 1.339905864393843, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3066728115081787, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8509047031402588, "num_tokens": 400070804.0, "step": 10533 }, { "epoch": 1.3400330746724336, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.604551076889038, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8636488318443298, "num_tokens": 400114021.0, "step": 10534 }, { "epoch": 1.3401602849510241, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2081143856048584, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8740381598472595, "num_tokens": 400156526.0, "step": 10535 }, { "epoch": 1.3402874952296147, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7283060550689697, "learning_rate": 1e-05, "loss": 0.3929, "mean_token_accuracy": 0.889527440071106, "num_tokens": 400188487.0, "step": 10536 }, { "epoch": 1.3404147055082052, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.167896032333374, "learning_rate": 1e-05, "loss": 0.5208, "mean_token_accuracy": 0.8486648797988892, "num_tokens": 400222223.0, "step": 10537 }, { "epoch": 1.3405419157867955, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8658745288848877, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8852874636650085, "num_tokens": 400259181.0, "step": 10538 }, { "epoch": 1.340669126065386, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1599693298339844, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8820983171463013, "num_tokens": 400295202.0, "step": 10539 }, { "epoch": 1.3407963363439765, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.043262958526611, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8816734552383423, "num_tokens": 400330936.0, "step": 10540 }, { "epoch": 1.340923546622567, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.31575608253479, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8722766637802124, "num_tokens": 400367971.0, "step": 10541 }, { "epoch": 1.3410507569011576, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.083799362182617, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8655598163604736, "num_tokens": 400406868.0, "step": 10542 }, { "epoch": 1.3411779671797481, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.521622657775879, "learning_rate": 1e-05, "loss": 0.503, "mean_token_accuracy": 0.8553944230079651, "num_tokens": 400442907.0, "step": 10543 }, { "epoch": 1.3413051774583387, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9295403957366943, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8559539914131165, "num_tokens": 400479485.0, "step": 10544 }, { "epoch": 1.3414323877369292, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5828561782836914, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.865312397480011, "num_tokens": 400513790.0, "step": 10545 }, { "epoch": 1.3415595980155197, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7007102966308594, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8693825006484985, "num_tokens": 400557638.0, "step": 10546 }, { "epoch": 1.3416868082941102, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5715200901031494, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8807334303855896, "num_tokens": 400594121.0, "step": 10547 }, { "epoch": 1.3418140185727006, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.50488018989563, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8686152696609497, "num_tokens": 400632427.0, "step": 10548 }, { "epoch": 1.341941228851291, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5602169036865234, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8743500709533691, "num_tokens": 400666938.0, "step": 10549 }, { "epoch": 1.3420684391298816, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.345883369445801, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8827289342880249, "num_tokens": 400703510.0, "step": 10550 }, { "epoch": 1.3421956494084721, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5775182247161865, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8742935657501221, "num_tokens": 400737082.0, "step": 10551 }, { "epoch": 1.3423228596870627, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2179317474365234, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8697620034217834, "num_tokens": 400787254.0, "step": 10552 }, { "epoch": 1.3424500699656532, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6510350704193115, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.872035026550293, "num_tokens": 400824384.0, "step": 10553 }, { "epoch": 1.3425772802442437, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.156177520751953, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8635122776031494, "num_tokens": 400859536.0, "step": 10554 }, { "epoch": 1.3427044905228342, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7727270126342773, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8814659118652344, "num_tokens": 400896022.0, "step": 10555 }, { "epoch": 1.3428317008014248, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.115204334259033, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8783004283905029, "num_tokens": 400936291.0, "step": 10556 }, { "epoch": 1.3429589110800153, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9365246295928955, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8804831504821777, "num_tokens": 400980749.0, "step": 10557 }, { "epoch": 1.3430861213586058, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.1787703037261963, "learning_rate": 1e-05, "loss": 0.5073, "mean_token_accuracy": 0.8534328937530518, "num_tokens": 401016591.0, "step": 10558 }, { "epoch": 1.3432133316371964, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.086215496063232, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8684222102165222, "num_tokens": 401062520.0, "step": 10559 }, { "epoch": 1.343340541915787, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.326850414276123, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8744586706161499, "num_tokens": 401105968.0, "step": 10560 }, { "epoch": 1.3434677521943774, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1131367683410645, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8840763568878174, "num_tokens": 401145290.0, "step": 10561 }, { "epoch": 1.343594962472968, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.5166263580322266, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8782250285148621, "num_tokens": 401181253.0, "step": 10562 }, { "epoch": 1.3437221727515583, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.985285520553589, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.872094988822937, "num_tokens": 401218191.0, "step": 10563 }, { "epoch": 1.3438493830301488, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5784103870391846, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8700367212295532, "num_tokens": 401252049.0, "step": 10564 }, { "epoch": 1.3439765933087393, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7957825660705566, "learning_rate": 1e-05, "loss": 0.5189, "mean_token_accuracy": 0.8481094837188721, "num_tokens": 401294819.0, "step": 10565 }, { "epoch": 1.3441038035873298, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.558164358139038, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8749228715896606, "num_tokens": 401332007.0, "step": 10566 }, { "epoch": 1.3442310138659204, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.620220184326172, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8802528381347656, "num_tokens": 401371442.0, "step": 10567 }, { "epoch": 1.344358224144511, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4883694648742676, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8767076730728149, "num_tokens": 401409956.0, "step": 10568 }, { "epoch": 1.3444854344231014, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5171473026275635, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8798152804374695, "num_tokens": 401446445.0, "step": 10569 }, { "epoch": 1.344612644701692, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4199156761169434, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8603192567825317, "num_tokens": 401489366.0, "step": 10570 }, { "epoch": 1.3447398549802825, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.467369318008423, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8645589351654053, "num_tokens": 401533326.0, "step": 10571 }, { "epoch": 1.3448670652588728, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.360233783721924, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.875524640083313, "num_tokens": 401571607.0, "step": 10572 }, { "epoch": 1.3449942755374633, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4651455879211426, "learning_rate": 1e-05, "loss": 0.5083, "mean_token_accuracy": 0.8511124849319458, "num_tokens": 401607930.0, "step": 10573 }, { "epoch": 1.3451214858160538, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.291743516921997, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8838486671447754, "num_tokens": 401645841.0, "step": 10574 }, { "epoch": 1.3452486960946444, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.488267660140991, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8617637157440186, "num_tokens": 401680795.0, "step": 10575 }, { "epoch": 1.345375906373235, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2355172634124756, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8715312480926514, "num_tokens": 401719710.0, "step": 10576 }, { "epoch": 1.3455031166518254, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.551187038421631, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8729759454727173, "num_tokens": 401758956.0, "step": 10577 }, { "epoch": 1.345630326930416, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1629788875579834, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.8600724935531616, "num_tokens": 401798108.0, "step": 10578 }, { "epoch": 1.3457575372090065, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6702322959899902, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8784486055374146, "num_tokens": 401833495.0, "step": 10579 }, { "epoch": 1.345884747487597, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.094519853591919, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8688174486160278, "num_tokens": 401875399.0, "step": 10580 }, { "epoch": 1.3460119577661875, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.856407880783081, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8713244795799255, "num_tokens": 401915540.0, "step": 10581 }, { "epoch": 1.346139168044778, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1026222705841064, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.855280339717865, "num_tokens": 401960013.0, "step": 10582 }, { "epoch": 1.3462663783233686, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0773420333862305, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.859379768371582, "num_tokens": 401990709.0, "step": 10583 }, { "epoch": 1.3463935886019591, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.250220775604248, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8763284683227539, "num_tokens": 402031277.0, "step": 10584 }, { "epoch": 1.3465207988805497, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.174285411834717, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8745102882385254, "num_tokens": 402073248.0, "step": 10585 }, { "epoch": 1.3466480091591402, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.494006395339966, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8660956025123596, "num_tokens": 402116072.0, "step": 10586 }, { "epoch": 1.3467752194377305, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.070959091186523, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8705509901046753, "num_tokens": 402153092.0, "step": 10587 }, { "epoch": 1.346902429716321, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6099162101745605, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8761980533599854, "num_tokens": 402192959.0, "step": 10588 }, { "epoch": 1.3470296399949115, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8637075424194336, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8696987628936768, "num_tokens": 402232245.0, "step": 10589 }, { "epoch": 1.347156850273502, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.59576416015625, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8828498125076294, "num_tokens": 402265862.0, "step": 10590 }, { "epoch": 1.3472840605520926, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6704514026641846, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8774288892745972, "num_tokens": 402303747.0, "step": 10591 }, { "epoch": 1.3474112708306831, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.533737897872925, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8727045059204102, "num_tokens": 402342952.0, "step": 10592 }, { "epoch": 1.3475384811092737, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5371744632720947, "learning_rate": 1e-05, "loss": 0.5315, "mean_token_accuracy": 0.8454147577285767, "num_tokens": 402382388.0, "step": 10593 }, { "epoch": 1.3476656913878642, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.464207649230957, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.876118004322052, "num_tokens": 402418937.0, "step": 10594 }, { "epoch": 1.3477929016664547, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.457454204559326, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8607402443885803, "num_tokens": 402454347.0, "step": 10595 }, { "epoch": 1.3479201119450452, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.418203353881836, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8540984392166138, "num_tokens": 402490887.0, "step": 10596 }, { "epoch": 1.3480473222236355, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.409123182296753, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8625799417495728, "num_tokens": 402531293.0, "step": 10597 }, { "epoch": 1.348174532502226, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.355262517929077, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.8577373027801514, "num_tokens": 402574396.0, "step": 10598 }, { "epoch": 1.3483017427808166, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.383181571960449, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8700910210609436, "num_tokens": 402615609.0, "step": 10599 }, { "epoch": 1.3484289530594071, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3205363750457764, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8668956756591797, "num_tokens": 402650627.0, "step": 10600 }, { "epoch": 1.3485561633379977, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.391404390335083, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8663920164108276, "num_tokens": 402693036.0, "step": 10601 }, { "epoch": 1.3486833736165882, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2806663513183594, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.8806610703468323, "num_tokens": 402730748.0, "step": 10602 }, { "epoch": 1.3488105838951787, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4342169761657715, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8599426746368408, "num_tokens": 402762642.0, "step": 10603 }, { "epoch": 1.3489377941737692, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2382729053497314, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8629792928695679, "num_tokens": 402794265.0, "step": 10604 }, { "epoch": 1.3490650044523598, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4856770038604736, "learning_rate": 1e-05, "loss": 0.5208, "mean_token_accuracy": 0.8502165079116821, "num_tokens": 402830609.0, "step": 10605 }, { "epoch": 1.3491922147309503, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1645145416259766, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8763771653175354, "num_tokens": 402866608.0, "step": 10606 }, { "epoch": 1.3493194250095408, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.590047836303711, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8632919788360596, "num_tokens": 402904030.0, "step": 10607 }, { "epoch": 1.3494466352881314, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0954909324645996, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8702143430709839, "num_tokens": 402939552.0, "step": 10608 }, { "epoch": 1.3495738455667219, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7466115951538086, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8766602873802185, "num_tokens": 402981683.0, "step": 10609 }, { "epoch": 1.3497010558453124, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0421931743621826, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.8569825291633606, "num_tokens": 403019497.0, "step": 10610 }, { "epoch": 1.349828266123903, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.029179096221924, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8678386211395264, "num_tokens": 403053415.0, "step": 10611 }, { "epoch": 1.3499554764024932, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.0984508991241455, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8706902265548706, "num_tokens": 403089676.0, "step": 10612 }, { "epoch": 1.3500826866810838, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.285849571228027, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8662010431289673, "num_tokens": 403130429.0, "step": 10613 }, { "epoch": 1.3502098969596743, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.387378454208374, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8835769295692444, "num_tokens": 403171420.0, "step": 10614 }, { "epoch": 1.3503371072382648, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.343140602111816, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8672115802764893, "num_tokens": 403210700.0, "step": 10615 }, { "epoch": 1.3504643175168554, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.708159923553467, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8711454272270203, "num_tokens": 403249572.0, "step": 10616 }, { "epoch": 1.350591527795446, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.133468151092529, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8637558221817017, "num_tokens": 403288176.0, "step": 10617 }, { "epoch": 1.3507187380740364, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00023365020751953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7359676361083984, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8825737237930298, "num_tokens": 403329403.0, "step": 10618 }, { "epoch": 1.350845948352627, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.928457021713257, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.8672058582305908, "num_tokens": 403362276.0, "step": 10619 }, { "epoch": 1.3509731586312175, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.650291919708252, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8700560927391052, "num_tokens": 403397797.0, "step": 10620 }, { "epoch": 1.3511003689098078, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7524337768554688, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8693009614944458, "num_tokens": 403437943.0, "step": 10621 }, { "epoch": 1.3512275791883983, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.548102378845215, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8751988410949707, "num_tokens": 403473575.0, "step": 10622 }, { "epoch": 1.3513547894669888, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.659595012664795, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8648435473442078, "num_tokens": 403504720.0, "step": 10623 }, { "epoch": 1.3514819997455794, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.466752529144287, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8593246340751648, "num_tokens": 403549244.0, "step": 10624 }, { "epoch": 1.35160921002417, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5970170497894287, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8527785539627075, "num_tokens": 403582881.0, "step": 10625 }, { "epoch": 1.3517364203027604, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.329648017883301, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.881291389465332, "num_tokens": 403623477.0, "step": 10626 }, { "epoch": 1.351863630581351, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.614234209060669, "learning_rate": 1e-05, "loss": 0.4975, "mean_token_accuracy": 0.8558272123336792, "num_tokens": 403662399.0, "step": 10627 }, { "epoch": 1.3519908408599415, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.23685884475708, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8705110549926758, "num_tokens": 403703908.0, "step": 10628 }, { "epoch": 1.352118051138532, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.701742649078369, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8725401759147644, "num_tokens": 403746385.0, "step": 10629 }, { "epoch": 1.3522452614171225, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.168311357498169, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8747166395187378, "num_tokens": 403785728.0, "step": 10630 }, { "epoch": 1.352372471695713, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8579835891723633, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.876176118850708, "num_tokens": 403827991.0, "step": 10631 }, { "epoch": 1.3524996819743036, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.172760486602783, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8656600713729858, "num_tokens": 403861963.0, "step": 10632 }, { "epoch": 1.3526268922528941, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.024001598358154, "learning_rate": 1e-05, "loss": 0.4063, "mean_token_accuracy": 0.885159969329834, "num_tokens": 403899744.0, "step": 10633 }, { "epoch": 1.3527541025314846, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.282653331756592, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.8869590759277344, "num_tokens": 403939482.0, "step": 10634 }, { "epoch": 1.3528813128100752, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.095606803894043, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8835076689720154, "num_tokens": 403976233.0, "step": 10635 }, { "epoch": 1.3530085230886655, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5003278255462646, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.879197895526886, "num_tokens": 404011181.0, "step": 10636 }, { "epoch": 1.353135733367256, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.970226526260376, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8824582099914551, "num_tokens": 404050519.0, "step": 10637 }, { "epoch": 1.3532629436458465, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.594878673553467, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8781464099884033, "num_tokens": 404089407.0, "step": 10638 }, { "epoch": 1.353390153924437, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7304534912109375, "learning_rate": 1e-05, "loss": 0.3848, "mean_token_accuracy": 0.8923578858375549, "num_tokens": 404126564.0, "step": 10639 }, { "epoch": 1.3535173642030276, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5879552364349365, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8694283366203308, "num_tokens": 404166113.0, "step": 10640 }, { "epoch": 1.3536445744816181, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5216336250305176, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8697772026062012, "num_tokens": 404204513.0, "step": 10641 }, { "epoch": 1.3537717847602087, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.603116989135742, "learning_rate": 1e-05, "loss": 0.5341, "mean_token_accuracy": 0.8531221747398376, "num_tokens": 404244475.0, "step": 10642 }, { "epoch": 1.3538989950387992, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3525550365448, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8620748519897461, "num_tokens": 404278974.0, "step": 10643 }, { "epoch": 1.3540262053173897, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6064703464508057, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8624507188796997, "num_tokens": 404313123.0, "step": 10644 }, { "epoch": 1.3541534155959802, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.229039430618286, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.874666690826416, "num_tokens": 404350195.0, "step": 10645 }, { "epoch": 1.3542806258745705, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6605803966522217, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8723145723342896, "num_tokens": 404387044.0, "step": 10646 }, { "epoch": 1.354407836153161, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1538445949554443, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8718480467796326, "num_tokens": 404423877.0, "step": 10647 }, { "epoch": 1.3545350464317516, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.773362636566162, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8766589760780334, "num_tokens": 404466428.0, "step": 10648 }, { "epoch": 1.3546622567103421, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1278514862060547, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.867242157459259, "num_tokens": 404503341.0, "step": 10649 }, { "epoch": 1.3547894669889327, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9411048889160156, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8689150810241699, "num_tokens": 404543650.0, "step": 10650 }, { "epoch": 1.3549166772675232, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.174433946609497, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8648597002029419, "num_tokens": 404588834.0, "step": 10651 }, { "epoch": 1.3550438875461137, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.104779243469238, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8655529022216797, "num_tokens": 404625903.0, "step": 10652 }, { "epoch": 1.3551710978247042, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.390704393386841, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8740397691726685, "num_tokens": 404666299.0, "step": 10653 }, { "epoch": 1.3552983081032948, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.067508220672607, "learning_rate": 1e-05, "loss": 0.3959, "mean_token_accuracy": 0.890282154083252, "num_tokens": 404702743.0, "step": 10654 }, { "epoch": 1.3554255183818853, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.492375612258911, "learning_rate": 1e-05, "loss": 0.5108, "mean_token_accuracy": 0.852349042892456, "num_tokens": 404748010.0, "step": 10655 }, { "epoch": 1.3555527286604758, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.964890480041504, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8701829314231873, "num_tokens": 404785886.0, "step": 10656 }, { "epoch": 1.3556799389390664, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5448837280273438, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.8884367942810059, "num_tokens": 404820982.0, "step": 10657 }, { "epoch": 1.3558071492176569, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.800004720687866, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8683353662490845, "num_tokens": 404863943.0, "step": 10658 }, { "epoch": 1.3559343594962474, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5191891193389893, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.864824116230011, "num_tokens": 404905168.0, "step": 10659 }, { "epoch": 1.356061569774838, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.667830228805542, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8693540692329407, "num_tokens": 404941863.0, "step": 10660 }, { "epoch": 1.3561887800534282, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.464876890182495, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8667287826538086, "num_tokens": 404979780.0, "step": 10661 }, { "epoch": 1.3563159903320188, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5653960704803467, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.862557053565979, "num_tokens": 405019194.0, "step": 10662 }, { "epoch": 1.3564432006106093, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3949732780456543, "learning_rate": 1e-05, "loss": 0.394, "mean_token_accuracy": 0.8885709047317505, "num_tokens": 405056021.0, "step": 10663 }, { "epoch": 1.3565704108891998, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5297181606292725, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8600345849990845, "num_tokens": 405099195.0, "step": 10664 }, { "epoch": 1.3566976211677904, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.321915864944458, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8672423958778381, "num_tokens": 405142821.0, "step": 10665 }, { "epoch": 1.3568248314463809, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5406570434570312, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8659033179283142, "num_tokens": 405178434.0, "step": 10666 }, { "epoch": 1.3569520417249714, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2715771198272705, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8553168177604675, "num_tokens": 405211197.0, "step": 10667 }, { "epoch": 1.357079252003562, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5719497203826904, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8748189806938171, "num_tokens": 405250330.0, "step": 10668 }, { "epoch": 1.3572064622821525, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1971933841705322, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8723684549331665, "num_tokens": 405291694.0, "step": 10669 }, { "epoch": 1.3573336725607428, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6694657802581787, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8803461790084839, "num_tokens": 405326162.0, "step": 10670 }, { "epoch": 1.3574608828393333, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1546196937561035, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8792591691017151, "num_tokens": 405361862.0, "step": 10671 }, { "epoch": 1.3575880931179238, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8084769248962402, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8712939023971558, "num_tokens": 405402934.0, "step": 10672 }, { "epoch": 1.3577153033965144, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1542863845825195, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8633567690849304, "num_tokens": 405442206.0, "step": 10673 }, { "epoch": 1.357842513675105, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9569621086120605, "learning_rate": 1e-05, "loss": 0.544, "mean_token_accuracy": 0.8436263799667358, "num_tokens": 405483351.0, "step": 10674 }, { "epoch": 1.3579697239536954, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2531213760375977, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.8637220859527588, "num_tokens": 405523763.0, "step": 10675 }, { "epoch": 1.358096934232286, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.013209819793701, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8835746049880981, "num_tokens": 405559640.0, "step": 10676 }, { "epoch": 1.3582241445108765, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.4069221019744873, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8752579092979431, "num_tokens": 405594907.0, "step": 10677 }, { "epoch": 1.358351354789467, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9294605255126953, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.8847607374191284, "num_tokens": 405630346.0, "step": 10678 }, { "epoch": 1.3584785650680575, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.515300989151001, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.863426923751831, "num_tokens": 405661603.0, "step": 10679 }, { "epoch": 1.358605775346648, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.739272117614746, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8695998191833496, "num_tokens": 405698907.0, "step": 10680 }, { "epoch": 1.3587329856252386, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.516547441482544, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8606950044631958, "num_tokens": 405744780.0, "step": 10681 }, { "epoch": 1.3588601959038291, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.552828311920166, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.878487229347229, "num_tokens": 405777883.0, "step": 10682 }, { "epoch": 1.3589874061824196, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.517179012298584, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8674404621124268, "num_tokens": 405814680.0, "step": 10683 }, { "epoch": 1.3591146164610102, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.407628297805786, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.8539676666259766, "num_tokens": 405847734.0, "step": 10684 }, { "epoch": 1.3592418267396005, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5177440643310547, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8678925037384033, "num_tokens": 405882365.0, "step": 10685 }, { "epoch": 1.359369037018191, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2891855239868164, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8736821413040161, "num_tokens": 405920641.0, "step": 10686 }, { "epoch": 1.3594962472967815, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5430078506469727, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.864985466003418, "num_tokens": 405958523.0, "step": 10687 }, { "epoch": 1.359623457575372, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.202625036239624, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8745871782302856, "num_tokens": 405994338.0, "step": 10688 }, { "epoch": 1.3597506678539626, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.611280918121338, "learning_rate": 1e-05, "loss": 0.5011, "mean_token_accuracy": 0.8549020290374756, "num_tokens": 406033285.0, "step": 10689 }, { "epoch": 1.3598778781325531, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.127256393432617, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8789629340171814, "num_tokens": 406076933.0, "step": 10690 }, { "epoch": 1.3600050884111436, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7362406253814697, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.871402382850647, "num_tokens": 406118258.0, "step": 10691 }, { "epoch": 1.3601322986897342, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.09618878364563, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8791303634643555, "num_tokens": 406155445.0, "step": 10692 }, { "epoch": 1.3602595089683247, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.921388626098633, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8677142858505249, "num_tokens": 406195128.0, "step": 10693 }, { "epoch": 1.3603867192469152, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.145216226577759, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8767952919006348, "num_tokens": 406229311.0, "step": 10694 }, { "epoch": 1.3605139295255055, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.08978796005249, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8700146079063416, "num_tokens": 406266252.0, "step": 10695 }, { "epoch": 1.360641139804096, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.2988991737365723, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8775538802146912, "num_tokens": 406310555.0, "step": 10696 }, { "epoch": 1.3607683500826866, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.14517879486084, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8812787532806396, "num_tokens": 406349620.0, "step": 10697 }, { "epoch": 1.3608955603612771, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5165634155273438, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8722624182701111, "num_tokens": 406380956.0, "step": 10698 }, { "epoch": 1.3610227706398677, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.014780044555664, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8840752243995667, "num_tokens": 406417274.0, "step": 10699 }, { "epoch": 1.3611499809184582, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.584024667739868, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8569380044937134, "num_tokens": 406453809.0, "step": 10700 }, { "epoch": 1.3612771911970487, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8333680629730225, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8747037649154663, "num_tokens": 406489227.0, "step": 10701 }, { "epoch": 1.3614044014756392, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.544698476791382, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8751585483551025, "num_tokens": 406532736.0, "step": 10702 }, { "epoch": 1.3615316117542298, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.674290895462036, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8697472810745239, "num_tokens": 406574409.0, "step": 10703 }, { "epoch": 1.3616588220328203, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4968953132629395, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8583092093467712, "num_tokens": 406612519.0, "step": 10704 }, { "epoch": 1.3617860323114108, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.555140495300293, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.862449049949646, "num_tokens": 406652578.0, "step": 10705 }, { "epoch": 1.3619132425900013, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4367456436157227, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.857593297958374, "num_tokens": 406688739.0, "step": 10706 }, { "epoch": 1.3620404528685919, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4839630126953125, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.877825140953064, "num_tokens": 406727721.0, "step": 10707 }, { "epoch": 1.3621676631471824, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3737008571624756, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.864189088344574, "num_tokens": 406762864.0, "step": 10708 }, { "epoch": 1.362294873425773, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4469704627990723, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8805096745491028, "num_tokens": 406803333.0, "step": 10709 }, { "epoch": 1.3624220837043632, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3121938705444336, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8765543699264526, "num_tokens": 406840942.0, "step": 10710 }, { "epoch": 1.3625492939829538, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.470947742462158, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8724456429481506, "num_tokens": 406877293.0, "step": 10711 }, { "epoch": 1.3626765042615443, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2483718395233154, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.872739315032959, "num_tokens": 406915622.0, "step": 10712 }, { "epoch": 1.3628037145401348, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.51602840423584, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8823766708374023, "num_tokens": 406955378.0, "step": 10713 }, { "epoch": 1.3629309248187254, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002346038818359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1797938346862793, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.8627163171768188, "num_tokens": 406997305.0, "step": 10714 }, { "epoch": 1.3630581350973159, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6185054779052734, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8850660920143127, "num_tokens": 407039121.0, "step": 10715 }, { "epoch": 1.3631853453759064, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.100734233856201, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8630157709121704, "num_tokens": 407081007.0, "step": 10716 }, { "epoch": 1.363312555654497, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00023555755615234375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.793065309524536, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8708351850509644, "num_tokens": 407122498.0, "step": 10717 }, { "epoch": 1.3634397659330875, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0659687519073486, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8808627128601074, "num_tokens": 407169304.0, "step": 10718 }, { "epoch": 1.3635669762116778, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.0394415855407715, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8577331304550171, "num_tokens": 407204364.0, "step": 10719 }, { "epoch": 1.3636941864902683, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1928439140319824, "learning_rate": 1e-05, "loss": 0.5217, "mean_token_accuracy": 0.8472990989685059, "num_tokens": 407242467.0, "step": 10720 }, { "epoch": 1.3638213967688588, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.200911045074463, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8727269768714905, "num_tokens": 407279965.0, "step": 10721 }, { "epoch": 1.3639486070474494, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.44937801361084, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8809047937393188, "num_tokens": 407316086.0, "step": 10722 }, { "epoch": 1.3640758173260399, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.141615867614746, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8609235286712646, "num_tokens": 407355433.0, "step": 10723 }, { "epoch": 1.3642030276046304, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6059629917144775, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.876807689666748, "num_tokens": 407392919.0, "step": 10724 }, { "epoch": 1.364330237883221, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9697818756103516, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8662947416305542, "num_tokens": 407428089.0, "step": 10725 }, { "epoch": 1.3644574481618115, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.622375965118408, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8779645562171936, "num_tokens": 407467232.0, "step": 10726 }, { "epoch": 1.364584658440402, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7697088718414307, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8627245426177979, "num_tokens": 407504975.0, "step": 10727 }, { "epoch": 1.3647118687189925, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5644798278808594, "learning_rate": 1e-05, "loss": 0.3839, "mean_token_accuracy": 0.8917026519775391, "num_tokens": 407535308.0, "step": 10728 }, { "epoch": 1.364839078997583, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6057231426239014, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8728453516960144, "num_tokens": 407578956.0, "step": 10729 }, { "epoch": 1.3649662892761736, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4829137325286865, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8739962577819824, "num_tokens": 407617633.0, "step": 10730 }, { "epoch": 1.3650934995547641, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.529271364212036, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8748654127120972, "num_tokens": 407646322.0, "step": 10731 }, { "epoch": 1.3652207098333546, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4080698490142822, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8619124889373779, "num_tokens": 407684957.0, "step": 10732 }, { "epoch": 1.3653479201119452, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4750161170959473, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8754628300666809, "num_tokens": 407722917.0, "step": 10733 }, { "epoch": 1.3654751303905355, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3349852561950684, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8709403276443481, "num_tokens": 407759599.0, "step": 10734 }, { "epoch": 1.365602340669126, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.48555850982666, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8671015501022339, "num_tokens": 407796416.0, "step": 10735 }, { "epoch": 1.3657295509477165, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.268751859664917, "learning_rate": 1e-05, "loss": 0.5298, "mean_token_accuracy": 0.8475185632705688, "num_tokens": 407841211.0, "step": 10736 }, { "epoch": 1.365856761226307, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.522977352142334, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8688711524009705, "num_tokens": 407879409.0, "step": 10737 }, { "epoch": 1.3659839715048976, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.205453395843506, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8666604161262512, "num_tokens": 407915275.0, "step": 10738 }, { "epoch": 1.3661111817834881, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.610243797302246, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8734121322631836, "num_tokens": 407953641.0, "step": 10739 }, { "epoch": 1.3662383920620786, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1241607666015625, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.876314103603363, "num_tokens": 407992662.0, "step": 10740 }, { "epoch": 1.3663656023406692, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7751333713531494, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8783978223800659, "num_tokens": 408026005.0, "step": 10741 }, { "epoch": 1.3664928126192597, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0912294387817383, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8706368207931519, "num_tokens": 408065250.0, "step": 10742 }, { "epoch": 1.3666200228978502, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.97646164894104, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8695828914642334, "num_tokens": 408109333.0, "step": 10743 }, { "epoch": 1.3667472331764405, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.171424627304077, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8756102323532104, "num_tokens": 408142416.0, "step": 10744 }, { "epoch": 1.366874443455031, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.129727840423584, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8866349458694458, "num_tokens": 408184335.0, "step": 10745 }, { "epoch": 1.3670016537336216, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3837790489196777, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8703823089599609, "num_tokens": 408221585.0, "step": 10746 }, { "epoch": 1.3671288640122121, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1289753913879395, "learning_rate": 1e-05, "loss": 0.5202, "mean_token_accuracy": 0.848950207233429, "num_tokens": 408262105.0, "step": 10747 }, { "epoch": 1.3672560742908026, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.567216396331787, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8734942078590393, "num_tokens": 408302365.0, "step": 10748 }, { "epoch": 1.3673832845693932, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00023651123046875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.948441982269287, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.864037036895752, "num_tokens": 408342033.0, "step": 10749 }, { "epoch": 1.3675104948479837, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5981807708740234, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8743524551391602, "num_tokens": 408381512.0, "step": 10750 }, { "epoch": 1.3676377051265742, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.755507707595825, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8702301979064941, "num_tokens": 408423867.0, "step": 10751 }, { "epoch": 1.3677649154051648, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5465831756591797, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8729771375656128, "num_tokens": 408468417.0, "step": 10752 }, { "epoch": 1.3678921256837553, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.590818405151367, "learning_rate": 1e-05, "loss": 0.5211, "mean_token_accuracy": 0.8508110642433167, "num_tokens": 408514628.0, "step": 10753 }, { "epoch": 1.3680193359623458, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4969189167022705, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8574069142341614, "num_tokens": 408551936.0, "step": 10754 }, { "epoch": 1.3681465462409363, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.498920440673828, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8672298192977905, "num_tokens": 408581713.0, "step": 10755 }, { "epoch": 1.3682737565195269, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4217159748077393, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8783077001571655, "num_tokens": 408618265.0, "step": 10756 }, { "epoch": 1.3684009667981174, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.437492847442627, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8717050552368164, "num_tokens": 408655701.0, "step": 10757 }, { "epoch": 1.368528177076708, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.369483470916748, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8610578775405884, "num_tokens": 408690281.0, "step": 10758 }, { "epoch": 1.3686553873552982, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4089791774749756, "learning_rate": 1e-05, "loss": 0.5149, "mean_token_accuracy": 0.8520214557647705, "num_tokens": 408733483.0, "step": 10759 }, { "epoch": 1.3687825976338888, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3202335834503174, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8716970086097717, "num_tokens": 408770596.0, "step": 10760 }, { "epoch": 1.3689098079124793, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.41569447517395, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.880151093006134, "num_tokens": 408808265.0, "step": 10761 }, { "epoch": 1.3690370181910698, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.264802932739258, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.880908727645874, "num_tokens": 408847205.0, "step": 10762 }, { "epoch": 1.3691642284696603, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4545373916625977, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8656663298606873, "num_tokens": 408885744.0, "step": 10763 }, { "epoch": 1.3692914387482509, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2037789821624756, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.877703845500946, "num_tokens": 408925393.0, "step": 10764 }, { "epoch": 1.3694186490268414, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5334692001342773, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8675282001495361, "num_tokens": 408964995.0, "step": 10765 }, { "epoch": 1.369545859305432, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.161534070968628, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8793731331825256, "num_tokens": 408996336.0, "step": 10766 }, { "epoch": 1.3696730695840225, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6412758827209473, "learning_rate": 1e-05, "loss": 0.5065, "mean_token_accuracy": 0.8538036346435547, "num_tokens": 409034972.0, "step": 10767 }, { "epoch": 1.3698002798626128, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.089629888534546, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8837807774543762, "num_tokens": 409069421.0, "step": 10768 }, { "epoch": 1.3699274901412033, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8206064701080322, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.864678680896759, "num_tokens": 409104763.0, "step": 10769 }, { "epoch": 1.3700547004197938, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.058988094329834, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.875008761882782, "num_tokens": 409142799.0, "step": 10770 }, { "epoch": 1.3701819106983844, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.068665027618408, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8624219298362732, "num_tokens": 409178861.0, "step": 10771 }, { "epoch": 1.3703091209769749, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.173187017440796, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8727012276649475, "num_tokens": 409218628.0, "step": 10772 }, { "epoch": 1.3704363312555654, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.273601055145264, "learning_rate": 1e-05, "loss": 0.522, "mean_token_accuracy": 0.8521313667297363, "num_tokens": 409255209.0, "step": 10773 }, { "epoch": 1.370563541534156, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4735336303710938, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8682844042778015, "num_tokens": 409295807.0, "step": 10774 }, { "epoch": 1.3706907518127465, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.21753454208374, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8732764720916748, "num_tokens": 409329263.0, "step": 10775 }, { "epoch": 1.370817962091337, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6767566204071045, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8893629312515259, "num_tokens": 409365571.0, "step": 10776 }, { "epoch": 1.3709451723699275, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.005488872528076, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8672388792037964, "num_tokens": 409406096.0, "step": 10777 }, { "epoch": 1.371072382648518, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6622142791748047, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8585742115974426, "num_tokens": 409450436.0, "step": 10778 }, { "epoch": 1.3711995929271086, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.820976972579956, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8702760934829712, "num_tokens": 409485134.0, "step": 10779 }, { "epoch": 1.371326803205699, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5809924602508545, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8715990781784058, "num_tokens": 409522890.0, "step": 10780 }, { "epoch": 1.3714540134842896, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.67384672164917, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8670413494110107, "num_tokens": 409561537.0, "step": 10781 }, { "epoch": 1.3715812237628802, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4880173206329346, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8806132078170776, "num_tokens": 409595259.0, "step": 10782 }, { "epoch": 1.3717084340414705, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.588252067565918, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8778015375137329, "num_tokens": 409633978.0, "step": 10783 }, { "epoch": 1.371835644320061, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3942339420318604, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8712291121482849, "num_tokens": 409671320.0, "step": 10784 }, { "epoch": 1.3719628545986515, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.563318967819214, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8738081455230713, "num_tokens": 409712816.0, "step": 10785 }, { "epoch": 1.372090064877242, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3060073852539062, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8685109615325928, "num_tokens": 409751918.0, "step": 10786 }, { "epoch": 1.3722172751558326, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6055119037628174, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8858169317245483, "num_tokens": 409786905.0, "step": 10787 }, { "epoch": 1.3723444854344231, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2201013565063477, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8699762225151062, "num_tokens": 409827634.0, "step": 10788 }, { "epoch": 1.3724716957130136, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6951985359191895, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8743821382522583, "num_tokens": 409863060.0, "step": 10789 }, { "epoch": 1.3725989059916042, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.166593551635742, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8771152496337891, "num_tokens": 409896723.0, "step": 10790 }, { "epoch": 1.3727261162701947, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.841651439666748, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8705862760543823, "num_tokens": 409935912.0, "step": 10791 }, { "epoch": 1.3728533265487852, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.161092758178711, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8732240796089172, "num_tokens": 409973381.0, "step": 10792 }, { "epoch": 1.3729805368273755, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.011947154998779, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.86402827501297, "num_tokens": 410010548.0, "step": 10793 }, { "epoch": 1.373107747105966, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.277193784713745, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8853800296783447, "num_tokens": 410046456.0, "step": 10794 }, { "epoch": 1.3732349573845566, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.08074951171875, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8868733048439026, "num_tokens": 410080568.0, "step": 10795 }, { "epoch": 1.3733621676631471, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.4562714099884033, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8762433528900146, "num_tokens": 410121008.0, "step": 10796 }, { "epoch": 1.3734893779417376, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.967841386795044, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8744941353797913, "num_tokens": 410165872.0, "step": 10797 }, { "epoch": 1.3736165882203282, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.565772771835327, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8761463165283203, "num_tokens": 410207379.0, "step": 10798 }, { "epoch": 1.3737437984989187, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.774815320968628, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8585875034332275, "num_tokens": 410246729.0, "step": 10799 }, { "epoch": 1.3738710087775092, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.578951835632324, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8767969012260437, "num_tokens": 410281237.0, "step": 10800 }, { "epoch": 1.3739982190560998, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5663723945617676, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8654273152351379, "num_tokens": 410318485.0, "step": 10801 }, { "epoch": 1.3741254293346903, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5562210083007812, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8712188601493835, "num_tokens": 410357995.0, "step": 10802 }, { "epoch": 1.3742526396132808, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.403146743774414, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8700106739997864, "num_tokens": 410395672.0, "step": 10803 }, { "epoch": 1.3743798498918713, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.553027629852295, "learning_rate": 1e-05, "loss": 0.5165, "mean_token_accuracy": 0.8490983247756958, "num_tokens": 410433566.0, "step": 10804 }, { "epoch": 1.3745070601704619, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.285297393798828, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8694981932640076, "num_tokens": 410473622.0, "step": 10805 }, { "epoch": 1.3746342704490524, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.578057050704956, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8777709007263184, "num_tokens": 410507155.0, "step": 10806 }, { "epoch": 1.374761480727643, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2131474018096924, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8667477369308472, "num_tokens": 410543352.0, "step": 10807 }, { "epoch": 1.3748886910062332, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6364119052886963, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8776776790618896, "num_tokens": 410584560.0, "step": 10808 }, { "epoch": 1.3750159012848238, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.160207509994507, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8796412944793701, "num_tokens": 410615948.0, "step": 10809 }, { "epoch": 1.3751431115634143, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7522966861724854, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8729507327079773, "num_tokens": 410649513.0, "step": 10810 }, { "epoch": 1.3752703218420048, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.108525276184082, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8684530854225159, "num_tokens": 410693087.0, "step": 10811 }, { "epoch": 1.3753975321205953, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.920365333557129, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8822310566902161, "num_tokens": 410729200.0, "step": 10812 }, { "epoch": 1.3755247423991859, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.154609203338623, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8652153611183167, "num_tokens": 410769830.0, "step": 10813 }, { "epoch": 1.3756519526777764, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0845465660095215, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8670956492424011, "num_tokens": 410815570.0, "step": 10814 }, { "epoch": 1.375779162956367, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3064842224121094, "learning_rate": 1e-05, "loss": 0.3813, "mean_token_accuracy": 0.8947823643684387, "num_tokens": 410852777.0, "step": 10815 }, { "epoch": 1.3759063732349575, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.148454666137695, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8752185702323914, "num_tokens": 410886809.0, "step": 10816 }, { "epoch": 1.3760335835135478, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5141713619232178, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.87476646900177, "num_tokens": 410924631.0, "step": 10817 }, { "epoch": 1.3761607937921383, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.020013332366943, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.871668815612793, "num_tokens": 410965212.0, "step": 10818 }, { "epoch": 1.3762880040707288, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.595968723297119, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.8589736223220825, "num_tokens": 411003429.0, "step": 10819 }, { "epoch": 1.3764152143493193, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.827482223510742, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8565254211425781, "num_tokens": 411043121.0, "step": 10820 }, { "epoch": 1.3765424246279099, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.571938991546631, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.860267698764801, "num_tokens": 411082146.0, "step": 10821 }, { "epoch": 1.3766696349065004, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6534197330474854, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8663138747215271, "num_tokens": 411116070.0, "step": 10822 }, { "epoch": 1.376796845185091, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5049562454223633, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8647561073303223, "num_tokens": 411162066.0, "step": 10823 }, { "epoch": 1.3769240554636815, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.529141664505005, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8744059205055237, "num_tokens": 411200939.0, "step": 10824 }, { "epoch": 1.377051265742272, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4522767066955566, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.865760326385498, "num_tokens": 411241161.0, "step": 10825 }, { "epoch": 1.3771784760208625, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.453920602798462, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8683596849441528, "num_tokens": 411284316.0, "step": 10826 }, { "epoch": 1.377305686299453, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.40313720703125, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8545156717300415, "num_tokens": 411320795.0, "step": 10827 }, { "epoch": 1.3774328965780436, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4140501022338867, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8747075796127319, "num_tokens": 411357513.0, "step": 10828 }, { "epoch": 1.377560106856634, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3432748317718506, "learning_rate": 1e-05, "loss": 0.384, "mean_token_accuracy": 0.8912222385406494, "num_tokens": 411395310.0, "step": 10829 }, { "epoch": 1.3776873171352246, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4170851707458496, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8703992366790771, "num_tokens": 411437774.0, "step": 10830 }, { "epoch": 1.3778145274138152, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3043463230133057, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8812949061393738, "num_tokens": 411471399.0, "step": 10831 }, { "epoch": 1.3779417376924055, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4521467685699463, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8599364161491394, "num_tokens": 411505289.0, "step": 10832 }, { "epoch": 1.378068947970996, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2386364936828613, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8812016248703003, "num_tokens": 411541463.0, "step": 10833 }, { "epoch": 1.3781961582495865, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.5135843753814697, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8690917491912842, "num_tokens": 411577276.0, "step": 10834 }, { "epoch": 1.378323368528177, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1893374919891357, "learning_rate": 1e-05, "loss": 0.5234, "mean_token_accuracy": 0.8429406881332397, "num_tokens": 411610417.0, "step": 10835 }, { "epoch": 1.3784505788067676, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6150569915771484, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8520348072052002, "num_tokens": 411646539.0, "step": 10836 }, { "epoch": 1.378577789085358, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1167383193969727, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8639261722564697, "num_tokens": 411685890.0, "step": 10837 }, { "epoch": 1.3787049993639486, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7545154094696045, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8797827959060669, "num_tokens": 411725566.0, "step": 10838 }, { "epoch": 1.3788322096425392, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0521631240844727, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8765737414360046, "num_tokens": 411767256.0, "step": 10839 }, { "epoch": 1.3789594199211297, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.007570266723633, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8693115711212158, "num_tokens": 411802210.0, "step": 10840 }, { "epoch": 1.3790866301997202, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.1153225898742676, "learning_rate": 1e-05, "loss": 0.386, "mean_token_accuracy": 0.8898710012435913, "num_tokens": 411842011.0, "step": 10841 }, { "epoch": 1.3792138404783105, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2561869621276855, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8768004775047302, "num_tokens": 411877137.0, "step": 10842 }, { "epoch": 1.379341050756901, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4114279747009277, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.877842903137207, "num_tokens": 411909130.0, "step": 10843 }, { "epoch": 1.3794682610354916, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.269617080688477, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8665493130683899, "num_tokens": 411942542.0, "step": 10844 }, { "epoch": 1.3795954713140821, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6662795543670654, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8629821538925171, "num_tokens": 411974272.0, "step": 10845 }, { "epoch": 1.3797226815926726, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0561370849609375, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8798298835754395, "num_tokens": 412012748.0, "step": 10846 }, { "epoch": 1.3798498918712632, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6807713508605957, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8659489154815674, "num_tokens": 412051171.0, "step": 10847 }, { "epoch": 1.3799771021498537, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.868245840072632, "learning_rate": 1e-05, "loss": 0.514, "mean_token_accuracy": 0.8577501773834229, "num_tokens": 412090267.0, "step": 10848 }, { "epoch": 1.3801043124284442, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6307249069213867, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8636452555656433, "num_tokens": 412118471.0, "step": 10849 }, { "epoch": 1.3802315227070348, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6829135417938232, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8760764598846436, "num_tokens": 412163639.0, "step": 10850 }, { "epoch": 1.3803587329856253, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.523026943206787, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8596525192260742, "num_tokens": 412203685.0, "step": 10851 }, { "epoch": 1.3804859432642158, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5849709510803223, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8716074228286743, "num_tokens": 412238810.0, "step": 10852 }, { "epoch": 1.3806131535428063, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4505937099456787, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8679001331329346, "num_tokens": 412273840.0, "step": 10853 }, { "epoch": 1.3807403638213969, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5136878490448, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8689475059509277, "num_tokens": 412309931.0, "step": 10854 }, { "epoch": 1.3808675740999874, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3795628547668457, "learning_rate": 1e-05, "loss": 0.5384, "mean_token_accuracy": 0.8391851186752319, "num_tokens": 412346934.0, "step": 10855 }, { "epoch": 1.380994784378578, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4897091388702393, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.868019700050354, "num_tokens": 412387942.0, "step": 10856 }, { "epoch": 1.3811219946571682, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2968266010284424, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8716058135032654, "num_tokens": 412429726.0, "step": 10857 }, { "epoch": 1.3812492049357588, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5075628757476807, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8743011951446533, "num_tokens": 412472711.0, "step": 10858 }, { "epoch": 1.3813764152143493, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2255196571350098, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8792904615402222, "num_tokens": 412511384.0, "step": 10859 }, { "epoch": 1.3815036254929398, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5934574604034424, "learning_rate": 1e-05, "loss": 0.5072, "mean_token_accuracy": 0.8554468154907227, "num_tokens": 412552991.0, "step": 10860 }, { "epoch": 1.3816308357715303, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.174729824066162, "learning_rate": 1e-05, "loss": 0.5501, "mean_token_accuracy": 0.8349555730819702, "num_tokens": 412593794.0, "step": 10861 }, { "epoch": 1.3817580460501209, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7171075344085693, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8788937926292419, "num_tokens": 412625569.0, "step": 10862 }, { "epoch": 1.3818852563287114, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.109937906265259, "learning_rate": 1e-05, "loss": 0.3932, "mean_token_accuracy": 0.8867223262786865, "num_tokens": 412658086.0, "step": 10863 }, { "epoch": 1.382012466607302, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.905078172683716, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8701006770133972, "num_tokens": 412698642.0, "step": 10864 }, { "epoch": 1.3821396768858925, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1191742420196533, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8735771179199219, "num_tokens": 412738327.0, "step": 10865 }, { "epoch": 1.3822668871644828, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.123212814331055, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8709617853164673, "num_tokens": 412773420.0, "step": 10866 }, { "epoch": 1.3823940974430733, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3116116523742676, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8763231635093689, "num_tokens": 412805383.0, "step": 10867 }, { "epoch": 1.3825213077216638, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.183467864990234, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8766899704933167, "num_tokens": 412846924.0, "step": 10868 }, { "epoch": 1.3826485180002543, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.548149824142456, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8607233762741089, "num_tokens": 412883232.0, "step": 10869 }, { "epoch": 1.3827757282788449, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0440897941589355, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8779363036155701, "num_tokens": 412922110.0, "step": 10870 }, { "epoch": 1.3829029385574354, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6299219131469727, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8665143847465515, "num_tokens": 412963661.0, "step": 10871 }, { "epoch": 1.383030148836026, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8284530639648438, "learning_rate": 1e-05, "loss": 0.3795, "mean_token_accuracy": 0.8959716558456421, "num_tokens": 413002468.0, "step": 10872 }, { "epoch": 1.3831573591146165, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5945098400115967, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.8553134202957153, "num_tokens": 413048135.0, "step": 10873 }, { "epoch": 1.383284569393207, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.658398151397705, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8711971044540405, "num_tokens": 413082552.0, "step": 10874 }, { "epoch": 1.3834117796717975, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5334835052490234, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8632022738456726, "num_tokens": 413129230.0, "step": 10875 }, { "epoch": 1.383538989950388, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5177576541900635, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8848368525505066, "num_tokens": 413168830.0, "step": 10876 }, { "epoch": 1.3836662002289786, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002384185791015625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4623312950134277, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.871918261051178, "num_tokens": 413208423.0, "step": 10877 }, { "epoch": 1.383793410507569, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4673259258270264, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8708042502403259, "num_tokens": 413238133.0, "step": 10878 }, { "epoch": 1.3839206207861596, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.406647205352783, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8768091797828674, "num_tokens": 413273653.0, "step": 10879 }, { "epoch": 1.3840478310647502, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4094531536102295, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8728420734405518, "num_tokens": 413316492.0, "step": 10880 }, { "epoch": 1.3841750413433405, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3446555137634277, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8798226714134216, "num_tokens": 413359101.0, "step": 10881 }, { "epoch": 1.384302251621931, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.39984393119812, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.881942868232727, "num_tokens": 413396618.0, "step": 10882 }, { "epoch": 1.3844294619005215, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3110454082489014, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.881580114364624, "num_tokens": 413431863.0, "step": 10883 }, { "epoch": 1.384556672179112, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4177627563476562, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8672467470169067, "num_tokens": 413470137.0, "step": 10884 }, { "epoch": 1.3846838824577026, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2668416500091553, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8818081617355347, "num_tokens": 413508475.0, "step": 10885 }, { "epoch": 1.384811092736293, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.454536199569702, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.869267463684082, "num_tokens": 413548846.0, "step": 10886 }, { "epoch": 1.3849383030148836, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2095093727111816, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8686276078224182, "num_tokens": 413587375.0, "step": 10887 }, { "epoch": 1.3850655132934742, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5353188514709473, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8719323873519897, "num_tokens": 413626701.0, "step": 10888 }, { "epoch": 1.3851927235720647, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1378610134124756, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8562523722648621, "num_tokens": 413665207.0, "step": 10889 }, { "epoch": 1.385319933850655, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.671377182006836, "learning_rate": 1e-05, "loss": 0.4039, "mean_token_accuracy": 0.8849706649780273, "num_tokens": 413700549.0, "step": 10890 }, { "epoch": 1.3854471441292455, "ewc_loss": 0.0673828125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.3392181396484375e-05, "grad_norm": 3.110196352005005, "learning_rate": 1e-05, "loss": 0.5336, "mean_token_accuracy": 0.8404777646064758, "num_tokens": 413738308.0, "step": 10891 }, { "epoch": 1.385574354407836, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8352551460266113, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8735694885253906, "num_tokens": 413779540.0, "step": 10892 }, { "epoch": 1.3857015646864266, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0444250106811523, "learning_rate": 1e-05, "loss": 0.3858, "mean_token_accuracy": 0.888516902923584, "num_tokens": 413818677.0, "step": 10893 }, { "epoch": 1.385828774965017, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.10126256942749, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8676947951316833, "num_tokens": 413861397.0, "step": 10894 }, { "epoch": 1.3859559852436076, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.172349214553833, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8765569925308228, "num_tokens": 413897629.0, "step": 10895 }, { "epoch": 1.3860831955221982, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.328150272369385, "learning_rate": 1e-05, "loss": 0.5218, "mean_token_accuracy": 0.8542163372039795, "num_tokens": 413935711.0, "step": 10896 }, { "epoch": 1.3862104058007887, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5099048614501953, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.879217267036438, "num_tokens": 413977340.0, "step": 10897 }, { "epoch": 1.3863376160793792, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.267241954803467, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8799169063568115, "num_tokens": 414014429.0, "step": 10898 }, { "epoch": 1.3864648263579697, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7293002605438232, "learning_rate": 1e-05, "loss": 0.5058, "mean_token_accuracy": 0.8592247366905212, "num_tokens": 414055426.0, "step": 10899 }, { "epoch": 1.3865920366365603, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.055237770080566, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8623526692390442, "num_tokens": 414086549.0, "step": 10900 }, { "epoch": 1.3867192469151508, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7144856452941895, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8764685392379761, "num_tokens": 414121767.0, "step": 10901 }, { "epoch": 1.3868464571937413, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.860222339630127, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8636789917945862, "num_tokens": 414154201.0, "step": 10902 }, { "epoch": 1.3869736674723319, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6111605167388916, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.8588720560073853, "num_tokens": 414189898.0, "step": 10903 }, { "epoch": 1.3871008777509224, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.71001935005188, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8734934329986572, "num_tokens": 414226507.0, "step": 10904 }, { "epoch": 1.387228088029513, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4962995052337646, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8669863939285278, "num_tokens": 414261909.0, "step": 10905 }, { "epoch": 1.3873552983081032, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6401615142822266, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8628044128417969, "num_tokens": 414303296.0, "step": 10906 }, { "epoch": 1.3874825085866938, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3964085578918457, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8684312701225281, "num_tokens": 414340638.0, "step": 10907 }, { "epoch": 1.3876097188652843, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.63600492477417, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8679279088973999, "num_tokens": 414376900.0, "step": 10908 }, { "epoch": 1.3877369291438748, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.293121814727783, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8732015490531921, "num_tokens": 414410657.0, "step": 10909 }, { "epoch": 1.3878641394224653, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.7018072605133057, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8546663522720337, "num_tokens": 414443165.0, "step": 10910 }, { "epoch": 1.3879913497010559, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2096071243286133, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8569190502166748, "num_tokens": 414482185.0, "step": 10911 }, { "epoch": 1.3881185599796464, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8258819580078125, "learning_rate": 1e-05, "loss": 0.516, "mean_token_accuracy": 0.8574935793876648, "num_tokens": 414517509.0, "step": 10912 }, { "epoch": 1.388245770258237, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2735066413879395, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8729003071784973, "num_tokens": 414558650.0, "step": 10913 }, { "epoch": 1.3883729805368275, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8929193019866943, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8753563761711121, "num_tokens": 414594301.0, "step": 10914 }, { "epoch": 1.3885001908154178, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2482082843780518, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8773659467697144, "num_tokens": 414626895.0, "step": 10915 }, { "epoch": 1.3886274010940083, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.972193956375122, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8717318773269653, "num_tokens": 414664708.0, "step": 10916 }, { "epoch": 1.3887546113725988, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.348109245300293, "learning_rate": 1e-05, "loss": 0.511, "mean_token_accuracy": 0.8514142036437988, "num_tokens": 414706243.0, "step": 10917 }, { "epoch": 1.3888818216511893, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.967188596725464, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8706182241439819, "num_tokens": 414740681.0, "step": 10918 }, { "epoch": 1.3890090319297799, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.475471019744873, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8731880187988281, "num_tokens": 414774942.0, "step": 10919 }, { "epoch": 1.3891362422083704, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8180785179138184, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8798618316650391, "num_tokens": 414817369.0, "step": 10920 }, { "epoch": 1.389263452486961, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5173704624176025, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8663153648376465, "num_tokens": 414855503.0, "step": 10921 }, { "epoch": 1.3893906627655515, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.650865316390991, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8650141954421997, "num_tokens": 414891974.0, "step": 10922 }, { "epoch": 1.389517873044142, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.506427764892578, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8808640837669373, "num_tokens": 414926773.0, "step": 10923 }, { "epoch": 1.3896450833227325, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4939463138580322, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8789403438568115, "num_tokens": 414962848.0, "step": 10924 }, { "epoch": 1.389772293601323, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4868111610412598, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.874811589717865, "num_tokens": 415003341.0, "step": 10925 }, { "epoch": 1.3898995038799136, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.36833119392395, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8804709911346436, "num_tokens": 415036836.0, "step": 10926 }, { "epoch": 1.390026714158504, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4806647300720215, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8712942600250244, "num_tokens": 415078589.0, "step": 10927 }, { "epoch": 1.3901539244370946, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.281358480453491, "learning_rate": 1e-05, "loss": 0.4038, "mean_token_accuracy": 0.8851292133331299, "num_tokens": 415114916.0, "step": 10928 }, { "epoch": 1.3902811347156852, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5137174129486084, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8673681616783142, "num_tokens": 415153774.0, "step": 10929 }, { "epoch": 1.3904083449942755, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.230334520339966, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.857729434967041, "num_tokens": 415188359.0, "step": 10930 }, { "epoch": 1.390535555272866, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5590884685516357, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8756288290023804, "num_tokens": 415224423.0, "step": 10931 }, { "epoch": 1.3906627655514565, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1588315963745117, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.864174485206604, "num_tokens": 415263219.0, "step": 10932 }, { "epoch": 1.390789975830047, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6632518768310547, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8690568208694458, "num_tokens": 415301043.0, "step": 10933 }, { "epoch": 1.3909171861086376, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0988962650299072, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8696159720420837, "num_tokens": 415339128.0, "step": 10934 }, { "epoch": 1.391044396387228, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8333656787872314, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8702701330184937, "num_tokens": 415374574.0, "step": 10935 }, { "epoch": 1.3911716066658186, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1023457050323486, "learning_rate": 1e-05, "loss": 0.5265, "mean_token_accuracy": 0.8435795307159424, "num_tokens": 415410133.0, "step": 10936 }, { "epoch": 1.3912988169444092, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.034209251403809, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8581568002700806, "num_tokens": 415451329.0, "step": 10937 }, { "epoch": 1.3914260272229997, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.190443277359009, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8702083826065063, "num_tokens": 415491448.0, "step": 10938 }, { "epoch": 1.39155323750159, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.1890106201171875, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8706698417663574, "num_tokens": 415533583.0, "step": 10939 }, { "epoch": 1.3916804477801805, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.4902167320251465, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8593124151229858, "num_tokens": 415567985.0, "step": 10940 }, { "epoch": 1.391807658058771, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.106175422668457, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.873051643371582, "num_tokens": 415603130.0, "step": 10941 }, { "epoch": 1.3919348683373616, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5725901126861572, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8611177206039429, "num_tokens": 415641284.0, "step": 10942 }, { "epoch": 1.392062078615952, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00023937225341796875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9458675384521484, "learning_rate": 1e-05, "loss": 0.5056, "mean_token_accuracy": 0.8572397232055664, "num_tokens": 415682483.0, "step": 10943 }, { "epoch": 1.3921892888945426, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.577455520629883, "learning_rate": 1e-05, "loss": 0.5166, "mean_token_accuracy": 0.8537646532058716, "num_tokens": 415723364.0, "step": 10944 }, { "epoch": 1.3923164991731332, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7768633365631104, "learning_rate": 1e-05, "loss": 0.5021, "mean_token_accuracy": 0.8597526550292969, "num_tokens": 415762750.0, "step": 10945 }, { "epoch": 1.3924437094517237, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5315659046173096, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8586205244064331, "num_tokens": 415801363.0, "step": 10946 }, { "epoch": 1.3925709197303142, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6301209926605225, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8733013868331909, "num_tokens": 415836121.0, "step": 10947 }, { "epoch": 1.3926981300089047, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.47917103767395, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8578687906265259, "num_tokens": 415880145.0, "step": 10948 }, { "epoch": 1.3928253402874953, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5307679176330566, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8685996532440186, "num_tokens": 415921244.0, "step": 10949 }, { "epoch": 1.3929525505660858, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.413422107696533, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8703030347824097, "num_tokens": 415957166.0, "step": 10950 }, { "epoch": 1.3930797608446763, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4754960536956787, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8612290620803833, "num_tokens": 415999462.0, "step": 10951 }, { "epoch": 1.3932069711232669, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.351839542388916, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8696881532669067, "num_tokens": 416038619.0, "step": 10952 }, { "epoch": 1.3933341814018574, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.457068681716919, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8664305210113525, "num_tokens": 416084885.0, "step": 10953 }, { "epoch": 1.393461391680448, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.297907590866089, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.858870804309845, "num_tokens": 416126863.0, "step": 10954 }, { "epoch": 1.3935886019590382, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.480456590652466, "learning_rate": 1e-05, "loss": 0.3771, "mean_token_accuracy": 0.8931912779808044, "num_tokens": 416160888.0, "step": 10955 }, { "epoch": 1.3937158122376287, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2373597621917725, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8752024173736572, "num_tokens": 416199012.0, "step": 10956 }, { "epoch": 1.3938430225162193, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5558693408966064, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8793371319770813, "num_tokens": 416234078.0, "step": 10957 }, { "epoch": 1.3939702327948098, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1625614166259766, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8764083981513977, "num_tokens": 416272912.0, "step": 10958 }, { "epoch": 1.3940974430734003, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.677150249481201, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8779388666152954, "num_tokens": 416312829.0, "step": 10959 }, { "epoch": 1.3942246533519909, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1001429557800293, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8693356513977051, "num_tokens": 416353337.0, "step": 10960 }, { "epoch": 1.3943518636305814, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8698039054870605, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8651961088180542, "num_tokens": 416392425.0, "step": 10961 }, { "epoch": 1.394479073909172, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1001439094543457, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8596722483634949, "num_tokens": 416434641.0, "step": 10962 }, { "epoch": 1.3946062841877624, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.096299171447754, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8671879172325134, "num_tokens": 416479107.0, "step": 10963 }, { "epoch": 1.3947334944663528, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.266592502593994, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.864099383354187, "num_tokens": 416512523.0, "step": 10964 }, { "epoch": 1.3948607047449433, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.212940216064453, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8699536323547363, "num_tokens": 416550176.0, "step": 10965 }, { "epoch": 1.3949879150235338, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.510007381439209, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8696627616882324, "num_tokens": 416583878.0, "step": 10966 }, { "epoch": 1.3951151253021243, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.126189231872559, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8656454086303711, "num_tokens": 416612863.0, "step": 10967 }, { "epoch": 1.3952423355807149, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.646505117416382, "learning_rate": 1e-05, "loss": 0.5234, "mean_token_accuracy": 0.855744481086731, "num_tokens": 416650691.0, "step": 10968 }, { "epoch": 1.3953695458593054, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.904237985610962, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8735045790672302, "num_tokens": 416687227.0, "step": 10969 }, { "epoch": 1.395496756137896, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6209912300109863, "learning_rate": 1e-05, "loss": 0.4878, "mean_token_accuracy": 0.8616131544113159, "num_tokens": 416725712.0, "step": 10970 }, { "epoch": 1.3956239664164865, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.725105047225952, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8774733543395996, "num_tokens": 416756169.0, "step": 10971 }, { "epoch": 1.395751176695077, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5595157146453857, "learning_rate": 1e-05, "loss": 0.5162, "mean_token_accuracy": 0.847671627998352, "num_tokens": 416794351.0, "step": 10972 }, { "epoch": 1.3958783869736675, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5751047134399414, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8840707540512085, "num_tokens": 416829201.0, "step": 10973 }, { "epoch": 1.396005597252258, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.482044219970703, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.875758171081543, "num_tokens": 416862171.0, "step": 10974 }, { "epoch": 1.3961328075308486, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.507011651992798, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8583210706710815, "num_tokens": 416895923.0, "step": 10975 }, { "epoch": 1.396260017809439, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4155657291412354, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8637972474098206, "num_tokens": 416927441.0, "step": 10976 }, { "epoch": 1.3963872280880296, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4545810222625732, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8723897933959961, "num_tokens": 416962911.0, "step": 10977 }, { "epoch": 1.3965144383666201, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.347700595855713, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8640286922454834, "num_tokens": 417001403.0, "step": 10978 }, { "epoch": 1.3966416486452105, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.445866584777832, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8608323931694031, "num_tokens": 417038043.0, "step": 10979 }, { "epoch": 1.396768858923801, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3005762100219727, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8632768392562866, "num_tokens": 417073516.0, "step": 10980 }, { "epoch": 1.3968960692023915, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4759366512298584, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8680967092514038, "num_tokens": 417112334.0, "step": 10981 }, { "epoch": 1.397023279480982, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2236926555633545, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8707566261291504, "num_tokens": 417155273.0, "step": 10982 }, { "epoch": 1.3971504897595726, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.549867630004883, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8775988817214966, "num_tokens": 417190145.0, "step": 10983 }, { "epoch": 1.397277700038163, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1546382904052734, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8833855390548706, "num_tokens": 417228184.0, "step": 10984 }, { "epoch": 1.3974049103167536, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.682990789413452, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8818237781524658, "num_tokens": 417258590.0, "step": 10985 }, { "epoch": 1.3975321205953442, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024127960205078125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0888748168945312, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8802229166030884, "num_tokens": 417295969.0, "step": 10986 }, { "epoch": 1.3976593308739347, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.896970510482788, "learning_rate": 1e-05, "loss": 0.5107, "mean_token_accuracy": 0.8536484837532043, "num_tokens": 417334428.0, "step": 10987 }, { "epoch": 1.397786541152525, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.099680185317993, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8734312057495117, "num_tokens": 417370074.0, "step": 10988 }, { "epoch": 1.3979137514311155, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.115819454193115, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8754546642303467, "num_tokens": 417413544.0, "step": 10989 }, { "epoch": 1.398040961709706, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2615158557891846, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8663308620452881, "num_tokens": 417445893.0, "step": 10990 }, { "epoch": 1.3981681719882966, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.246603488922119, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8806431889533997, "num_tokens": 417481469.0, "step": 10991 }, { "epoch": 1.398295382266887, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.536609649658203, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.863693118095398, "num_tokens": 417517070.0, "step": 10992 }, { "epoch": 1.3984225925454776, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.137336730957031, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8654958009719849, "num_tokens": 417558265.0, "step": 10993 }, { "epoch": 1.3985498028240682, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6552538871765137, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.873320460319519, "num_tokens": 417599083.0, "step": 10994 }, { "epoch": 1.3986770131026587, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9414596557617188, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8803511261940002, "num_tokens": 417632425.0, "step": 10995 }, { "epoch": 1.3988042233812492, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.631781578063965, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8764254450798035, "num_tokens": 417667209.0, "step": 10996 }, { "epoch": 1.3989314336598397, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7516212463378906, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8743470311164856, "num_tokens": 417706168.0, "step": 10997 }, { "epoch": 1.3990586439384303, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5496654510498047, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8779983520507812, "num_tokens": 417745225.0, "step": 10998 }, { "epoch": 1.3991858542170208, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6282472610473633, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8762272000312805, "num_tokens": 417782688.0, "step": 10999 }, { "epoch": 1.3993130644956113, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4797894954681396, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8684967756271362, "num_tokens": 417814157.0, "step": 11000 }, { "epoch": 1.3994402747742019, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5607407093048096, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8627215027809143, "num_tokens": 417845460.0, "step": 11001 }, { "epoch": 1.3995674850527924, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3862006664276123, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8729512691497803, "num_tokens": 417881460.0, "step": 11002 }, { "epoch": 1.399694695331383, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5235400199890137, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8866517543792725, "num_tokens": 417919690.0, "step": 11003 }, { "epoch": 1.3998219056099732, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2896790504455566, "learning_rate": 1e-05, "loss": 0.3887, "mean_token_accuracy": 0.8873630166053772, "num_tokens": 417959474.0, "step": 11004 }, { "epoch": 1.3999491158885637, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.586215019226074, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.86855548620224, "num_tokens": 417992342.0, "step": 11005 }, { "epoch": 1.4000763261671543, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.208531141281128, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8525332808494568, "num_tokens": 418032006.0, "step": 11006 }, { "epoch": 1.4002035364457448, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.683415412902832, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8573105335235596, "num_tokens": 418069318.0, "step": 11007 }, { "epoch": 1.4003307467243353, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.143552780151367, "learning_rate": 1e-05, "loss": 0.3986, "mean_token_accuracy": 0.8846457004547119, "num_tokens": 418103605.0, "step": 11008 }, { "epoch": 1.4004579570029259, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8544440269470215, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.863845944404602, "num_tokens": 418136363.0, "step": 11009 }, { "epoch": 1.4005851672815164, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1295034885406494, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8701784014701843, "num_tokens": 418178096.0, "step": 11010 }, { "epoch": 1.400712377560107, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0569071769714355, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8826313018798828, "num_tokens": 418211025.0, "step": 11011 }, { "epoch": 1.4008395878386974, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.265866756439209, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8719455003738403, "num_tokens": 418251210.0, "step": 11012 }, { "epoch": 1.4009667981172877, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.15410041809082, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8780107498168945, "num_tokens": 418286593.0, "step": 11013 }, { "epoch": 1.4010940083958783, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.492405652999878, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8816025853157043, "num_tokens": 418324794.0, "step": 11014 }, { "epoch": 1.4012212186744688, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.038942813873291, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8766107559204102, "num_tokens": 418366067.0, "step": 11015 }, { "epoch": 1.4013484289530593, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6068100929260254, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.8881310224533081, "num_tokens": 418406600.0, "step": 11016 }, { "epoch": 1.4014756392316499, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.83469557762146, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8747596740722656, "num_tokens": 418440168.0, "step": 11017 }, { "epoch": 1.4016028495102404, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6100199222564697, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.853312611579895, "num_tokens": 418482247.0, "step": 11018 }, { "epoch": 1.401730059788831, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6274664402008057, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8816518783569336, "num_tokens": 418517376.0, "step": 11019 }, { "epoch": 1.4018572700674214, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5723440647125244, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8654223680496216, "num_tokens": 418555841.0, "step": 11020 }, { "epoch": 1.401984480346012, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4529547691345215, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8720575571060181, "num_tokens": 418594232.0, "step": 11021 }, { "epoch": 1.4021116906246025, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.53326416015625, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8798319101333618, "num_tokens": 418631726.0, "step": 11022 }, { "epoch": 1.402238900903193, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.346458911895752, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.8934046626091003, "num_tokens": 418671172.0, "step": 11023 }, { "epoch": 1.4023661111817836, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.524580955505371, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8768140077590942, "num_tokens": 418706541.0, "step": 11024 }, { "epoch": 1.402493321460374, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2775533199310303, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8766398429870605, "num_tokens": 418743557.0, "step": 11025 }, { "epoch": 1.4026205317389646, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5380747318267822, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8690526485443115, "num_tokens": 418783741.0, "step": 11026 }, { "epoch": 1.4027477420175551, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2106001377105713, "learning_rate": 1e-05, "loss": 0.4157, "mean_token_accuracy": 0.8788102269172668, "num_tokens": 418822208.0, "step": 11027 }, { "epoch": 1.4028749522961454, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6029975414276123, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8757569789886475, "num_tokens": 418862717.0, "step": 11028 }, { "epoch": 1.403002162574736, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.165332794189453, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.874952495098114, "num_tokens": 418892961.0, "step": 11029 }, { "epoch": 1.4031293728533265, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.716095209121704, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.8545366525650024, "num_tokens": 418933384.0, "step": 11030 }, { "epoch": 1.403256583131917, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.104475736618042, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8763858675956726, "num_tokens": 418971042.0, "step": 11031 }, { "epoch": 1.4033837934105076, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8865926265716553, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8747768402099609, "num_tokens": 419009949.0, "step": 11032 }, { "epoch": 1.403511003689098, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1158254146575928, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8778337240219116, "num_tokens": 419044607.0, "step": 11033 }, { "epoch": 1.4036382139676886, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.0996527671813965, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8750386238098145, "num_tokens": 419081407.0, "step": 11034 }, { "epoch": 1.4037654242462791, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2562057971954346, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8795629739761353, "num_tokens": 419125007.0, "step": 11035 }, { "epoch": 1.4038926345248697, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.220062255859375, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.874747633934021, "num_tokens": 419166009.0, "step": 11036 }, { "epoch": 1.40401984480346, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5346860885620117, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8692783117294312, "num_tokens": 419194185.0, "step": 11037 }, { "epoch": 1.4041470550820505, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.088290691375732, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8761910796165466, "num_tokens": 419236742.0, "step": 11038 }, { "epoch": 1.404274265360641, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6304638385772705, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8711671829223633, "num_tokens": 419272477.0, "step": 11039 }, { "epoch": 1.4044014756392316, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8977034091949463, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8631728887557983, "num_tokens": 419309198.0, "step": 11040 }, { "epoch": 1.404528685917822, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6149935722351074, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8791799545288086, "num_tokens": 419343065.0, "step": 11041 }, { "epoch": 1.4046558961964126, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.695566415786743, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8612372875213623, "num_tokens": 419382826.0, "step": 11042 }, { "epoch": 1.4047831064750032, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5592620372772217, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.862059473991394, "num_tokens": 419416839.0, "step": 11043 }, { "epoch": 1.4049103167535937, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.564345121383667, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8730224370956421, "num_tokens": 419448041.0, "step": 11044 }, { "epoch": 1.4050375270321842, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4906795024871826, "learning_rate": 1e-05, "loss": 0.5248, "mean_token_accuracy": 0.8473821878433228, "num_tokens": 419484917.0, "step": 11045 }, { "epoch": 1.4051647373107747, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4636425971984863, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8709681630134583, "num_tokens": 419521329.0, "step": 11046 }, { "epoch": 1.4052919475893653, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.434852361679077, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8679224848747253, "num_tokens": 419556141.0, "step": 11047 }, { "epoch": 1.4054191578679558, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.410600185394287, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.874075174331665, "num_tokens": 419593027.0, "step": 11048 }, { "epoch": 1.4055463681465463, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3779213428497314, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.880927562713623, "num_tokens": 419632419.0, "step": 11049 }, { "epoch": 1.4056735784251368, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.388007164001465, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8670486211776733, "num_tokens": 419677625.0, "step": 11050 }, { "epoch": 1.4058007887037274, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.337259531021118, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8690844774246216, "num_tokens": 419721546.0, "step": 11051 }, { "epoch": 1.405927998982318, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.402988910675049, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.8534043431282043, "num_tokens": 419759345.0, "step": 11052 }, { "epoch": 1.4060552092609082, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3163015842437744, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8691509962081909, "num_tokens": 419789819.0, "step": 11053 }, { "epoch": 1.4061824195394987, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4033117294311523, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.867584228515625, "num_tokens": 419830525.0, "step": 11054 }, { "epoch": 1.4063096298180893, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.259633779525757, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8701988458633423, "num_tokens": 419875592.0, "step": 11055 }, { "epoch": 1.4064368400966798, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4402096271514893, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8882217407226562, "num_tokens": 419913530.0, "step": 11056 }, { "epoch": 1.4065640503752703, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2014963626861572, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8701801300048828, "num_tokens": 419955947.0, "step": 11057 }, { "epoch": 1.4066912606538609, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5338141918182373, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.882050633430481, "num_tokens": 419986931.0, "step": 11058 }, { "epoch": 1.4068184709324514, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1345653533935547, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8714392185211182, "num_tokens": 420024037.0, "step": 11059 }, { "epoch": 1.406945681211042, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000240325927734375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6822917461395264, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8660138249397278, "num_tokens": 420062843.0, "step": 11060 }, { "epoch": 1.4070728914896324, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0537681579589844, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8550086617469788, "num_tokens": 420102311.0, "step": 11061 }, { "epoch": 1.4072001017682227, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.913907766342163, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.864043116569519, "num_tokens": 420143364.0, "step": 11062 }, { "epoch": 1.4073273120468133, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.0335209369659424, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.877143383026123, "num_tokens": 420180930.0, "step": 11063 }, { "epoch": 1.4074545223254038, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.235791206359863, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8728386163711548, "num_tokens": 420216048.0, "step": 11064 }, { "epoch": 1.4075817326039943, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.2644431591033936, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8726029992103577, "num_tokens": 420250792.0, "step": 11065 }, { "epoch": 1.4077089428825849, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.441568374633789, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8746235370635986, "num_tokens": 420280518.0, "step": 11066 }, { "epoch": 1.4078361531611754, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.6832685470581055, "learning_rate": 1e-05, "loss": 0.5078, "mean_token_accuracy": 0.8547170758247375, "num_tokens": 420318274.0, "step": 11067 }, { "epoch": 1.407963363439766, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.264102935791016, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8736766576766968, "num_tokens": 420353277.0, "step": 11068 }, { "epoch": 1.4080905737183564, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.8225579261779785, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8775039911270142, "num_tokens": 420384457.0, "step": 11069 }, { "epoch": 1.408217783996947, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.014718532562256, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8609265089035034, "num_tokens": 420419229.0, "step": 11070 }, { "epoch": 1.4083449942755375, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.74063777923584, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8611936569213867, "num_tokens": 420457002.0, "step": 11071 }, { "epoch": 1.408472204554128, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.822267532348633, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8799080848693848, "num_tokens": 420490802.0, "step": 11072 }, { "epoch": 1.4085994148327186, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.585635185241699, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8788270950317383, "num_tokens": 420531782.0, "step": 11073 }, { "epoch": 1.408726625111309, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7061102390289307, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.8598950505256653, "num_tokens": 420575577.0, "step": 11074 }, { "epoch": 1.4088538353898996, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4676780700683594, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8737332224845886, "num_tokens": 420612963.0, "step": 11075 }, { "epoch": 1.4089810456684901, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6620028018951416, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8794964551925659, "num_tokens": 420649518.0, "step": 11076 }, { "epoch": 1.4091082559470804, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.368919610977173, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8610947132110596, "num_tokens": 420684377.0, "step": 11077 }, { "epoch": 1.409235466225671, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.67450213432312, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8627486228942871, "num_tokens": 420721566.0, "step": 11078 }, { "epoch": 1.4093626765042615, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.264296770095825, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8729238510131836, "num_tokens": 420758591.0, "step": 11079 }, { "epoch": 1.409489886782852, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.775113582611084, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8624438047409058, "num_tokens": 420796864.0, "step": 11080 }, { "epoch": 1.4096170970614426, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1927101612091064, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8800497055053711, "num_tokens": 420834689.0, "step": 11081 }, { "epoch": 1.409744307340033, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9313547611236572, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8717451691627502, "num_tokens": 420877795.0, "step": 11082 }, { "epoch": 1.4098715176186236, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.222102403640747, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.874066948890686, "num_tokens": 420917217.0, "step": 11083 }, { "epoch": 1.4099987278972141, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.094366073608398, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8691613674163818, "num_tokens": 420952607.0, "step": 11084 }, { "epoch": 1.4101259381758047, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4111106395721436, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8626469373703003, "num_tokens": 420986570.0, "step": 11085 }, { "epoch": 1.410253148454395, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.068309307098389, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8712471723556519, "num_tokens": 421024482.0, "step": 11086 }, { "epoch": 1.4103803587329855, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.558013439178467, "learning_rate": 1e-05, "loss": 0.3845, "mean_token_accuracy": 0.892193615436554, "num_tokens": 421058185.0, "step": 11087 }, { "epoch": 1.410507569011576, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9020233154296875, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8693827390670776, "num_tokens": 421089796.0, "step": 11088 }, { "epoch": 1.4106347792901666, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.609355926513672, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8658126592636108, "num_tokens": 421132644.0, "step": 11089 }, { "epoch": 1.410761989568757, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6745383739471436, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8675739765167236, "num_tokens": 421167790.0, "step": 11090 }, { "epoch": 1.4108891998473476, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5874407291412354, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8644112944602966, "num_tokens": 421206765.0, "step": 11091 }, { "epoch": 1.4110164101259381, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4627997875213623, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8787808418273926, "num_tokens": 421249997.0, "step": 11092 }, { "epoch": 1.4111436204045287, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.569553852081299, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.873375654220581, "num_tokens": 421287674.0, "step": 11093 }, { "epoch": 1.4112708306831192, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3157429695129395, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8755363821983337, "num_tokens": 421325494.0, "step": 11094 }, { "epoch": 1.4113980409617097, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5965821743011475, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8804116249084473, "num_tokens": 421359934.0, "step": 11095 }, { "epoch": 1.4115252512403003, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.223220109939575, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8563704490661621, "num_tokens": 421399195.0, "step": 11096 }, { "epoch": 1.4116524615188908, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6716203689575195, "learning_rate": 1e-05, "loss": 0.3893, "mean_token_accuracy": 0.8890067934989929, "num_tokens": 421428822.0, "step": 11097 }, { "epoch": 1.4117796717974813, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.143500328063965, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8659883737564087, "num_tokens": 421468409.0, "step": 11098 }, { "epoch": 1.4119068820760718, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.820131540298462, "learning_rate": 1e-05, "loss": 0.5319, "mean_token_accuracy": 0.8497229814529419, "num_tokens": 421506905.0, "step": 11099 }, { "epoch": 1.4120340923546624, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.123706817626953, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8575029373168945, "num_tokens": 421545347.0, "step": 11100 }, { "epoch": 1.412161302633253, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.002972602844238, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8713222742080688, "num_tokens": 421580819.0, "step": 11101 }, { "epoch": 1.4122885129118432, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1862220764160156, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8737235069274902, "num_tokens": 421628474.0, "step": 11102 }, { "epoch": 1.4124157231904337, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.168859004974365, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8631154298782349, "num_tokens": 421670437.0, "step": 11103 }, { "epoch": 1.4125429334690243, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4273152351379395, "learning_rate": 1e-05, "loss": 0.5157, "mean_token_accuracy": 0.8531680107116699, "num_tokens": 421709157.0, "step": 11104 }, { "epoch": 1.4126701437476148, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.140423774719238, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8777921199798584, "num_tokens": 421743783.0, "step": 11105 }, { "epoch": 1.4127973540262053, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5898680686950684, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.8850656151771545, "num_tokens": 421781376.0, "step": 11106 }, { "epoch": 1.4129245643047958, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9710540771484375, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8819105625152588, "num_tokens": 421819367.0, "step": 11107 }, { "epoch": 1.4130517745833864, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.610281229019165, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8816851377487183, "num_tokens": 421862830.0, "step": 11108 }, { "epoch": 1.413178984861977, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.782820224761963, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8793999552726746, "num_tokens": 421898353.0, "step": 11109 }, { "epoch": 1.4133061951405674, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.57008695602417, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8699682950973511, "num_tokens": 421932911.0, "step": 11110 }, { "epoch": 1.4134334054191577, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6299242973327637, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8742587566375732, "num_tokens": 421974409.0, "step": 11111 }, { "epoch": 1.4135606156977483, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5155465602874756, "learning_rate": 1e-05, "loss": 0.5056, "mean_token_accuracy": 0.8510009050369263, "num_tokens": 422011913.0, "step": 11112 }, { "epoch": 1.4136878259763388, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5131430625915527, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8717613816261292, "num_tokens": 422054487.0, "step": 11113 }, { "epoch": 1.4138150362549293, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.444610357284546, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8709510564804077, "num_tokens": 422092980.0, "step": 11114 }, { "epoch": 1.4139422465335199, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4527764320373535, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.861452579498291, "num_tokens": 422139368.0, "step": 11115 }, { "epoch": 1.4140694568121104, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.374401569366455, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8826772570610046, "num_tokens": 422177342.0, "step": 11116 }, { "epoch": 1.414196667090701, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.43835711479187, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8767328262329102, "num_tokens": 422221378.0, "step": 11117 }, { "epoch": 1.4143238773692914, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3213601112365723, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8702834844589233, "num_tokens": 422260503.0, "step": 11118 }, { "epoch": 1.414451087647882, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4525644779205322, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8632261157035828, "num_tokens": 422301218.0, "step": 11119 }, { "epoch": 1.4145782979264725, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002422332763671875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.264813184738159, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8793464303016663, "num_tokens": 422337366.0, "step": 11120 }, { "epoch": 1.414705508205063, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4981744289398193, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8835887312889099, "num_tokens": 422376087.0, "step": 11121 }, { "epoch": 1.4148327184836536, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.202526092529297, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8716514706611633, "num_tokens": 422413871.0, "step": 11122 }, { "epoch": 1.414959928762244, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.595487594604492, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8727172613143921, "num_tokens": 422450834.0, "step": 11123 }, { "epoch": 1.4150871390408346, "ewc_loss": 0.06787109375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.115955352783203, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8796112537384033, "num_tokens": 422494566.0, "step": 11124 }, { "epoch": 1.4152143493194251, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7711920738220215, "learning_rate": 1e-05, "loss": 0.4941, "mean_token_accuracy": 0.8644139170646667, "num_tokens": 422530107.0, "step": 11125 }, { "epoch": 1.4153415595980154, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.082458734512329, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8510243892669678, "num_tokens": 422567598.0, "step": 11126 }, { "epoch": 1.415468769876606, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.988959789276123, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8709989190101624, "num_tokens": 422611705.0, "step": 11127 }, { "epoch": 1.4155959801551965, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1425838470458984, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8640581965446472, "num_tokens": 422648082.0, "step": 11128 }, { "epoch": 1.415723190433787, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.191349983215332, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8732779622077942, "num_tokens": 422691974.0, "step": 11129 }, { "epoch": 1.4158504007123776, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.369603157043457, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8625764846801758, "num_tokens": 422728245.0, "step": 11130 }, { "epoch": 1.415977610990968, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.234219074249268, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8734793663024902, "num_tokens": 422760920.0, "step": 11131 }, { "epoch": 1.4161048212695586, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6186070442199707, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8682241439819336, "num_tokens": 422796188.0, "step": 11132 }, { "epoch": 1.4162320315481491, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.034727573394775, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8815964460372925, "num_tokens": 422836232.0, "step": 11133 }, { "epoch": 1.4163592418267397, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6636862754821777, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8762831687927246, "num_tokens": 422872855.0, "step": 11134 }, { "epoch": 1.41648645210533, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8275299072265625, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8683432340621948, "num_tokens": 422912826.0, "step": 11135 }, { "epoch": 1.4166136623839205, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6137146949768066, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8618901968002319, "num_tokens": 422950034.0, "step": 11136 }, { "epoch": 1.416740872662511, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6539151668548584, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8751572370529175, "num_tokens": 422987764.0, "step": 11137 }, { "epoch": 1.4168680829411016, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.514848232269287, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.8872543573379517, "num_tokens": 423024274.0, "step": 11138 }, { "epoch": 1.416995293219692, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5566203594207764, "learning_rate": 1e-05, "loss": 0.4824, "mean_token_accuracy": 0.8599590063095093, "num_tokens": 423058630.0, "step": 11139 }, { "epoch": 1.4171225034982826, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4282479286193848, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8800008893013, "num_tokens": 423099829.0, "step": 11140 }, { "epoch": 1.4172497137768731, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5167076587677, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.872560977935791, "num_tokens": 423130369.0, "step": 11141 }, { "epoch": 1.4173769240554637, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.353442907333374, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8747084736824036, "num_tokens": 423166437.0, "step": 11142 }, { "epoch": 1.4175041343340542, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5005345344543457, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.8846638798713684, "num_tokens": 423202220.0, "step": 11143 }, { "epoch": 1.4176313446126447, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2804386615753174, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8684677481651306, "num_tokens": 423243417.0, "step": 11144 }, { "epoch": 1.4177585548912353, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.547638177871704, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8686202764511108, "num_tokens": 423282337.0, "step": 11145 }, { "epoch": 1.4178857651698258, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.219552993774414, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8664771318435669, "num_tokens": 423318666.0, "step": 11146 }, { "epoch": 1.4180129754484163, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6350109577178955, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8653859496116638, "num_tokens": 423357895.0, "step": 11147 }, { "epoch": 1.4181401857270068, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1417086124420166, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8704412579536438, "num_tokens": 423395330.0, "step": 11148 }, { "epoch": 1.4182673960055974, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7858073711395264, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8704686760902405, "num_tokens": 423436824.0, "step": 11149 }, { "epoch": 1.418394606284188, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1173102855682373, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8684534430503845, "num_tokens": 423471906.0, "step": 11150 }, { "epoch": 1.4185218165627782, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9796195030212402, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8735678195953369, "num_tokens": 423510774.0, "step": 11151 }, { "epoch": 1.4186490268413687, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.182837963104248, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8713443279266357, "num_tokens": 423549825.0, "step": 11152 }, { "epoch": 1.4187762371199593, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.161736965179443, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8756273984909058, "num_tokens": 423577546.0, "step": 11153 }, { "epoch": 1.4189034473985498, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3994390964508057, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8804218173027039, "num_tokens": 423613167.0, "step": 11154 }, { "epoch": 1.4190306576771403, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.133461952209473, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8588557243347168, "num_tokens": 423649601.0, "step": 11155 }, { "epoch": 1.4191578679557308, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5996789932250977, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8624509572982788, "num_tokens": 423685429.0, "step": 11156 }, { "epoch": 1.4192850782343214, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9613163471221924, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8787565231323242, "num_tokens": 423724270.0, "step": 11157 }, { "epoch": 1.419412288512912, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6217665672302246, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.881189227104187, "num_tokens": 423760935.0, "step": 11158 }, { "epoch": 1.4195394987915024, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.742327928543091, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8693442940711975, "num_tokens": 423801949.0, "step": 11159 }, { "epoch": 1.4196667090700927, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5696518421173096, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.88160240650177, "num_tokens": 423840532.0, "step": 11160 }, { "epoch": 1.4197939193486833, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.574352502822876, "learning_rate": 1e-05, "loss": 0.4794, "mean_token_accuracy": 0.8625420331954956, "num_tokens": 423877912.0, "step": 11161 }, { "epoch": 1.4199211296272738, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.510977029800415, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.878211259841919, "num_tokens": 423917836.0, "step": 11162 }, { "epoch": 1.4200483399058643, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.457052230834961, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8639234304428101, "num_tokens": 423957307.0, "step": 11163 }, { "epoch": 1.4201755501844548, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.478644609451294, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8710407614707947, "num_tokens": 423994280.0, "step": 11164 }, { "epoch": 1.4203027604630454, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.379453659057617, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8699233531951904, "num_tokens": 424033682.0, "step": 11165 }, { "epoch": 1.420429970741636, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4473068714141846, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8669348359107971, "num_tokens": 424072432.0, "step": 11166 }, { "epoch": 1.4205571810202264, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.328052282333374, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8625246286392212, "num_tokens": 424110074.0, "step": 11167 }, { "epoch": 1.420684391298817, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.446610927581787, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8667161464691162, "num_tokens": 424148263.0, "step": 11168 }, { "epoch": 1.4208116015774075, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2713847160339355, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8610481023788452, "num_tokens": 424190342.0, "step": 11169 }, { "epoch": 1.420938811855998, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4694652557373047, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8618646264076233, "num_tokens": 424236136.0, "step": 11170 }, { "epoch": 1.4210660221345885, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2243237495422363, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8588309288024902, "num_tokens": 424274436.0, "step": 11171 }, { "epoch": 1.421193232413179, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.525644540786743, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8890016674995422, "num_tokens": 424309661.0, "step": 11172 }, { "epoch": 1.4213204426917696, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.165785074234009, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8606804013252258, "num_tokens": 424347127.0, "step": 11173 }, { "epoch": 1.4214476529703601, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6245405673980713, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8782455921173096, "num_tokens": 424391287.0, "step": 11174 }, { "epoch": 1.4215748632489504, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0916004180908203, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.8818944692611694, "num_tokens": 424428091.0, "step": 11175 }, { "epoch": 1.421702073527541, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8108370304107666, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8680828213691711, "num_tokens": 424469270.0, "step": 11176 }, { "epoch": 1.4218292838061315, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.0487091541290283, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8571615219116211, "num_tokens": 424509116.0, "step": 11177 }, { "epoch": 1.421956494084722, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.07633638381958, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8737006187438965, "num_tokens": 424542777.0, "step": 11178 }, { "epoch": 1.4220837043633126, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.142482042312622, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8732144832611084, "num_tokens": 424582299.0, "step": 11179 }, { "epoch": 1.422210914641903, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.326200008392334, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8850482702255249, "num_tokens": 424617543.0, "step": 11180 }, { "epoch": 1.4223381249204936, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.4720780849456787, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8826232552528381, "num_tokens": 424657512.0, "step": 11181 }, { "epoch": 1.4224653351990841, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.320308208465576, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8614968657493591, "num_tokens": 424695954.0, "step": 11182 }, { "epoch": 1.4225925454776747, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.740664482116699, "learning_rate": 1e-05, "loss": 0.501, "mean_token_accuracy": 0.859367847442627, "num_tokens": 424736174.0, "step": 11183 }, { "epoch": 1.422719755756265, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.087827205657959, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.870415449142456, "num_tokens": 424776371.0, "step": 11184 }, { "epoch": 1.4228469660348555, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.740203380584717, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.86213219165802, "num_tokens": 424816874.0, "step": 11185 }, { "epoch": 1.422974176313446, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.877610445022583, "learning_rate": 1e-05, "loss": 0.4023, "mean_token_accuracy": 0.8863540887832642, "num_tokens": 424847924.0, "step": 11186 }, { "epoch": 1.4231013865920366, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.634817600250244, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8851934671401978, "num_tokens": 424887950.0, "step": 11187 }, { "epoch": 1.423228596870627, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7272815704345703, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.862583577632904, "num_tokens": 424922608.0, "step": 11188 }, { "epoch": 1.4233558071492176, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.530454397201538, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.879425585269928, "num_tokens": 424958564.0, "step": 11189 }, { "epoch": 1.4234830174278081, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6304335594177246, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8679825067520142, "num_tokens": 424993351.0, "step": 11190 }, { "epoch": 1.4236102277063987, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.428663969039917, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8702088594436646, "num_tokens": 425028018.0, "step": 11191 }, { "epoch": 1.4237374379849892, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.597881317138672, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8728593587875366, "num_tokens": 425065023.0, "step": 11192 }, { "epoch": 1.4238646482635797, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3269131183624268, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8763678073883057, "num_tokens": 425101152.0, "step": 11193 }, { "epoch": 1.4239918585421703, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.632077217102051, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8650672435760498, "num_tokens": 425137756.0, "step": 11194 }, { "epoch": 1.4241190688207608, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.243671417236328, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8751729726791382, "num_tokens": 425171947.0, "step": 11195 }, { "epoch": 1.4242462790993513, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7265987396240234, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8702492117881775, "num_tokens": 425207219.0, "step": 11196 }, { "epoch": 1.4243734893779418, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.181420087814331, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8629322052001953, "num_tokens": 425242515.0, "step": 11197 }, { "epoch": 1.4245006996565324, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.871882677078247, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8579747676849365, "num_tokens": 425283409.0, "step": 11198 }, { "epoch": 1.424627909935123, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.183414936065674, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8642891645431519, "num_tokens": 425321862.0, "step": 11199 }, { "epoch": 1.4247551202137132, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.037075996398926, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.881173849105835, "num_tokens": 425362034.0, "step": 11200 }, { "epoch": 1.4248823304923037, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3261680603027344, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8546630144119263, "num_tokens": 425397900.0, "step": 11201 }, { "epoch": 1.4250095407708943, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.078670501708984, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8714883327484131, "num_tokens": 425435458.0, "step": 11202 }, { "epoch": 1.4251367510494848, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5012552738189697, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8805292844772339, "num_tokens": 425465687.0, "step": 11203 }, { "epoch": 1.4252639613280753, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.938626766204834, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8689624071121216, "num_tokens": 425500702.0, "step": 11204 }, { "epoch": 1.4253911716066658, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.567650556564331, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8721023797988892, "num_tokens": 425542827.0, "step": 11205 }, { "epoch": 1.4255183818852564, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7387473583221436, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8620363473892212, "num_tokens": 425585146.0, "step": 11206 }, { "epoch": 1.425645592163847, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.56561017036438, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.878261387348175, "num_tokens": 425623790.0, "step": 11207 }, { "epoch": 1.4257728024424374, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5423638820648193, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.871741771697998, "num_tokens": 425663290.0, "step": 11208 }, { "epoch": 1.4259000127210277, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5418689250946045, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8749957084655762, "num_tokens": 425701226.0, "step": 11209 }, { "epoch": 1.4260272229996183, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.393231153488159, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8708657026290894, "num_tokens": 425737524.0, "step": 11210 }, { "epoch": 1.4261544332782088, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5567915439605713, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.877844512462616, "num_tokens": 425771010.0, "step": 11211 }, { "epoch": 1.4262816435567993, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2750749588012695, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8645458221435547, "num_tokens": 425813077.0, "step": 11212 }, { "epoch": 1.4264088538353898, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5803167819976807, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8850229978561401, "num_tokens": 425852718.0, "step": 11213 }, { "epoch": 1.4265360641139804, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.199310302734375, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8708385825157166, "num_tokens": 425888196.0, "step": 11214 }, { "epoch": 1.426663274392571, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.672672748565674, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8794170618057251, "num_tokens": 425926774.0, "step": 11215 }, { "epoch": 1.4267904846711614, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.160484790802002, "learning_rate": 1e-05, "loss": 0.3935, "mean_token_accuracy": 0.8866899013519287, "num_tokens": 425966207.0, "step": 11216 }, { "epoch": 1.426917694949752, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7982258796691895, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8615908622741699, "num_tokens": 426008792.0, "step": 11217 }, { "epoch": 1.4270449052283425, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.110517978668213, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8705295324325562, "num_tokens": 426049650.0, "step": 11218 }, { "epoch": 1.427172115506933, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.974903106689453, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8803120255470276, "num_tokens": 426092143.0, "step": 11219 }, { "epoch": 1.4272993257855235, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00024318695068359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1750457286834717, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8587988615036011, "num_tokens": 426135393.0, "step": 11220 }, { "epoch": 1.427426536064114, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.153327465057373, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8718773126602173, "num_tokens": 426171858.0, "step": 11221 }, { "epoch": 1.4275537463427046, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.396465539932251, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8817999362945557, "num_tokens": 426206635.0, "step": 11222 }, { "epoch": 1.4276809566212951, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.140170097351074, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8743369579315186, "num_tokens": 426240942.0, "step": 11223 }, { "epoch": 1.4278081668998854, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5609371662139893, "learning_rate": 1e-05, "loss": 0.3932, "mean_token_accuracy": 0.8902921676635742, "num_tokens": 426279592.0, "step": 11224 }, { "epoch": 1.427935377178476, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.983334541320801, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8831459879875183, "num_tokens": 426314218.0, "step": 11225 }, { "epoch": 1.4280625874570665, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6156344413757324, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8714637756347656, "num_tokens": 426351514.0, "step": 11226 }, { "epoch": 1.428189797735657, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7860143184661865, "learning_rate": 1e-05, "loss": 0.5082, "mean_token_accuracy": 0.8602239489555359, "num_tokens": 426389553.0, "step": 11227 }, { "epoch": 1.4283170080142475, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5705759525299072, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8802046775817871, "num_tokens": 426424026.0, "step": 11228 }, { "epoch": 1.428444218292838, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6212642192840576, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8726727962493896, "num_tokens": 426466711.0, "step": 11229 }, { "epoch": 1.4285714285714286, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.501847267150879, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8670405745506287, "num_tokens": 426507636.0, "step": 11230 }, { "epoch": 1.4286986388500191, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5146028995513916, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8673044443130493, "num_tokens": 426551811.0, "step": 11231 }, { "epoch": 1.4288258491286097, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4451520442962646, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8694837093353271, "num_tokens": 426585814.0, "step": 11232 }, { "epoch": 1.4289530594072, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.463744640350342, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8780755996704102, "num_tokens": 426619327.0, "step": 11233 }, { "epoch": 1.4290802696857905, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.388751983642578, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.862920880317688, "num_tokens": 426656976.0, "step": 11234 }, { "epoch": 1.429207479964381, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.434046506881714, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8801252841949463, "num_tokens": 426693930.0, "step": 11235 }, { "epoch": 1.4293346902429716, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.331981658935547, "learning_rate": 1e-05, "loss": 0.3913, "mean_token_accuracy": 0.8891107439994812, "num_tokens": 426731655.0, "step": 11236 }, { "epoch": 1.429461900521562, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.447240114212036, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8665024638175964, "num_tokens": 426768384.0, "step": 11237 }, { "epoch": 1.4295891108001526, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2761805057525635, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8696194887161255, "num_tokens": 426806319.0, "step": 11238 }, { "epoch": 1.4297163210787431, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.499255657196045, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.8478330373764038, "num_tokens": 426841076.0, "step": 11239 }, { "epoch": 1.4298435313573337, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.209484100341797, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.87830650806427, "num_tokens": 426877647.0, "step": 11240 }, { "epoch": 1.4299707416359242, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.575950860977173, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8731168508529663, "num_tokens": 426917975.0, "step": 11241 }, { "epoch": 1.4300979519145147, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.130577802658081, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8713396787643433, "num_tokens": 426958975.0, "step": 11242 }, { "epoch": 1.4302251621931052, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.741459846496582, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8706200122833252, "num_tokens": 426997205.0, "step": 11243 }, { "epoch": 1.4303523724716958, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0759646892547607, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8714532256126404, "num_tokens": 427037455.0, "step": 11244 }, { "epoch": 1.4304795827502863, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.975306510925293, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8673678636550903, "num_tokens": 427071443.0, "step": 11245 }, { "epoch": 1.4306067930288768, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1084094047546387, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8739370703697205, "num_tokens": 427113258.0, "step": 11246 }, { "epoch": 1.4307340033074674, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.215449810028076, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.874449610710144, "num_tokens": 427153523.0, "step": 11247 }, { "epoch": 1.430861213586058, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.366351366043091, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8714684247970581, "num_tokens": 427186818.0, "step": 11248 }, { "epoch": 1.4309884238646482, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.267436504364014, "learning_rate": 1e-05, "loss": 0.5141, "mean_token_accuracy": 0.8566081523895264, "num_tokens": 427226817.0, "step": 11249 }, { "epoch": 1.4311156341432387, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.625527858734131, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8690846562385559, "num_tokens": 427268471.0, "step": 11250 }, { "epoch": 1.4312428444218293, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.100151538848877, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8828239440917969, "num_tokens": 427301242.0, "step": 11251 }, { "epoch": 1.4313700547004198, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6996219158172607, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8626059293746948, "num_tokens": 427335230.0, "step": 11252 }, { "epoch": 1.4314972649790103, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.882558584213257, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8630411624908447, "num_tokens": 427371993.0, "step": 11253 }, { "epoch": 1.4316244752576008, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.635732650756836, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8630225658416748, "num_tokens": 427408810.0, "step": 11254 }, { "epoch": 1.4317516855361914, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7052760124206543, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8575983047485352, "num_tokens": 427448740.0, "step": 11255 }, { "epoch": 1.431878895814782, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.541032552719116, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8729699850082397, "num_tokens": 427488234.0, "step": 11256 }, { "epoch": 1.4320061060933724, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5868284702301025, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8686715364456177, "num_tokens": 427528053.0, "step": 11257 }, { "epoch": 1.4321333163719627, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.445239305496216, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8771640062332153, "num_tokens": 427566882.0, "step": 11258 }, { "epoch": 1.4322605266505533, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.546013593673706, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8743025064468384, "num_tokens": 427600464.0, "step": 11259 }, { "epoch": 1.4323877369291438, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3759522438049316, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8772059679031372, "num_tokens": 427633394.0, "step": 11260 }, { "epoch": 1.4325149472077343, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5212109088897705, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8708794713020325, "num_tokens": 427670591.0, "step": 11261 }, { "epoch": 1.4326421574863248, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2811169624328613, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.8828051090240479, "num_tokens": 427711752.0, "step": 11262 }, { "epoch": 1.4327693677649154, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5841434001922607, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8711766004562378, "num_tokens": 427744655.0, "step": 11263 }, { "epoch": 1.432896578043506, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2021989822387695, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8765848875045776, "num_tokens": 427784388.0, "step": 11264 }, { "epoch": 1.4330237883220964, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6929028034210205, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8620204925537109, "num_tokens": 427823498.0, "step": 11265 }, { "epoch": 1.433150998600687, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1328632831573486, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8715616464614868, "num_tokens": 427868185.0, "step": 11266 }, { "epoch": 1.4332782088792775, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.87947416305542, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8626883029937744, "num_tokens": 427903315.0, "step": 11267 }, { "epoch": 1.433405419157868, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1278786659240723, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8588716983795166, "num_tokens": 427950206.0, "step": 11268 }, { "epoch": 1.4335326294364585, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.123812198638916, "learning_rate": 1e-05, "loss": 0.5598, "mean_token_accuracy": 0.8508713245391846, "num_tokens": 427983038.0, "step": 11269 }, { "epoch": 1.433659839715049, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.306706190109253, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8781638741493225, "num_tokens": 428015148.0, "step": 11270 }, { "epoch": 1.4337870499936396, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.185500621795654, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8713196516036987, "num_tokens": 428054300.0, "step": 11271 }, { "epoch": 1.4339142602722301, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.537705659866333, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8590717911720276, "num_tokens": 428093802.0, "step": 11272 }, { "epoch": 1.4340414705508204, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.052463531494141, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8803908824920654, "num_tokens": 428133283.0, "step": 11273 }, { "epoch": 1.434168680829411, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.639465093612671, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8624313473701477, "num_tokens": 428171125.0, "step": 11274 }, { "epoch": 1.4342958911080015, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.833651065826416, "learning_rate": 1e-05, "loss": 0.5049, "mean_token_accuracy": 0.8528918027877808, "num_tokens": 428211731.0, "step": 11275 }, { "epoch": 1.434423101386592, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6171834468841553, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8671577572822571, "num_tokens": 428243737.0, "step": 11276 }, { "epoch": 1.4345503116651825, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6412010192871094, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8794906139373779, "num_tokens": 428278761.0, "step": 11277 }, { "epoch": 1.434677521943773, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5448009967803955, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8885884284973145, "num_tokens": 428316158.0, "step": 11278 }, { "epoch": 1.4348047322223636, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5151314735412598, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8676877617835999, "num_tokens": 428347180.0, "step": 11279 }, { "epoch": 1.4349319425009541, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.485698938369751, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8794293999671936, "num_tokens": 428389962.0, "step": 11280 }, { "epoch": 1.4350591527795447, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4255318641662598, "learning_rate": 1e-05, "loss": 0.5069, "mean_token_accuracy": 0.8496019840240479, "num_tokens": 428432950.0, "step": 11281 }, { "epoch": 1.435186363058135, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.447044849395752, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8542232513427734, "num_tokens": 428470745.0, "step": 11282 }, { "epoch": 1.4353135733367255, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.376532554626465, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8664870262145996, "num_tokens": 428508836.0, "step": 11283 }, { "epoch": 1.435440783615316, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4109535217285156, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8723918199539185, "num_tokens": 428546371.0, "step": 11284 }, { "epoch": 1.4355679938939065, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3461079597473145, "learning_rate": 1e-05, "loss": 0.3981, "mean_token_accuracy": 0.8839167356491089, "num_tokens": 428578878.0, "step": 11285 }, { "epoch": 1.435695204172497, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.40718150138855, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8639299273490906, "num_tokens": 428611218.0, "step": 11286 }, { "epoch": 1.4358224144510876, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3201708793640137, "learning_rate": 1e-05, "loss": 0.5339, "mean_token_accuracy": 0.8438807129859924, "num_tokens": 428649381.0, "step": 11287 }, { "epoch": 1.4359496247296781, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.40291428565979, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8685368895530701, "num_tokens": 428685221.0, "step": 11288 }, { "epoch": 1.4360768350082687, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.279803991317749, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8711675405502319, "num_tokens": 428724194.0, "step": 11289 }, { "epoch": 1.4362040452868592, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4273669719696045, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8681305646896362, "num_tokens": 428760501.0, "step": 11290 }, { "epoch": 1.4363312555654497, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.252580404281616, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8589212894439697, "num_tokens": 428793917.0, "step": 11291 }, { "epoch": 1.4364584658440402, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4711661338806152, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8776726126670837, "num_tokens": 428828998.0, "step": 11292 }, { "epoch": 1.4365856761226308, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1843786239624023, "learning_rate": 1e-05, "loss": 0.4123, "mean_token_accuracy": 0.8854473233222961, "num_tokens": 428864739.0, "step": 11293 }, { "epoch": 1.4367128864012213, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5648045539855957, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8587382435798645, "num_tokens": 428901026.0, "step": 11294 }, { "epoch": 1.4368400966798118, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.11236572265625, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8784005641937256, "num_tokens": 428937698.0, "step": 11295 }, { "epoch": 1.4369673069584024, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7346837520599365, "learning_rate": 1e-05, "loss": 0.5138, "mean_token_accuracy": 0.8542214035987854, "num_tokens": 428978288.0, "step": 11296 }, { "epoch": 1.4370945172369929, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0307064056396484, "learning_rate": 1e-05, "loss": 0.3851, "mean_token_accuracy": 0.8872345685958862, "num_tokens": 429014299.0, "step": 11297 }, { "epoch": 1.4372217275155832, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.006425380706787, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.86335289478302, "num_tokens": 429052455.0, "step": 11298 }, { "epoch": 1.4373489377941737, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.058246374130249, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8698448538780212, "num_tokens": 429094168.0, "step": 11299 }, { "epoch": 1.4374761480727642, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.341027736663818, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8679190874099731, "num_tokens": 429137147.0, "step": 11300 }, { "epoch": 1.4376033583513548, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.3668692111968994, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8774551749229431, "num_tokens": 429181679.0, "step": 11301 }, { "epoch": 1.4377305686299453, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.46944522857666, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.870608925819397, "num_tokens": 429222378.0, "step": 11302 }, { "epoch": 1.4378577789085358, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.7933478355407715, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8692400455474854, "num_tokens": 429262044.0, "step": 11303 }, { "epoch": 1.4379849891871264, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.225396156311035, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8709445595741272, "num_tokens": 429301643.0, "step": 11304 }, { "epoch": 1.438112199465717, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.854532241821289, "learning_rate": 1e-05, "loss": 0.5426, "mean_token_accuracy": 0.8466668725013733, "num_tokens": 429339190.0, "step": 11305 }, { "epoch": 1.4382394097443074, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.979306936264038, "learning_rate": 1e-05, "loss": 0.4726, "mean_token_accuracy": 0.8682206869125366, "num_tokens": 429378569.0, "step": 11306 }, { "epoch": 1.4383666200228977, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7109384536743164, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8704413175582886, "num_tokens": 429416012.0, "step": 11307 }, { "epoch": 1.4384938303014883, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.816279888153076, "learning_rate": 1e-05, "loss": 0.3884, "mean_token_accuracy": 0.8922908306121826, "num_tokens": 429453242.0, "step": 11308 }, { "epoch": 1.4386210405800788, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.565765142440796, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8762878179550171, "num_tokens": 429493908.0, "step": 11309 }, { "epoch": 1.4387482508586693, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7360527515411377, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8739641308784485, "num_tokens": 429530335.0, "step": 11310 }, { "epoch": 1.4388754611372598, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.439091920852661, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8662585020065308, "num_tokens": 429567523.0, "step": 11311 }, { "epoch": 1.4390026714158504, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7083961963653564, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8784307241439819, "num_tokens": 429605887.0, "step": 11312 }, { "epoch": 1.439129881694441, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.323655605316162, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8823356628417969, "num_tokens": 429644834.0, "step": 11313 }, { "epoch": 1.4392570919730314, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7745707035064697, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8793023824691772, "num_tokens": 429681151.0, "step": 11314 }, { "epoch": 1.439384302251622, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.251913070678711, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8606052398681641, "num_tokens": 429719135.0, "step": 11315 }, { "epoch": 1.4395115125302125, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.906787633895874, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8798954486846924, "num_tokens": 429753443.0, "step": 11316 }, { "epoch": 1.439638722808803, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2490346431732178, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.875827968120575, "num_tokens": 429788762.0, "step": 11317 }, { "epoch": 1.4397659330873935, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.054910182952881, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8594957590103149, "num_tokens": 429824054.0, "step": 11318 }, { "epoch": 1.439893143365984, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3835785388946533, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8677400946617126, "num_tokens": 429857772.0, "step": 11319 }, { "epoch": 1.4400203536445746, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.074244022369385, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8756697177886963, "num_tokens": 429891184.0, "step": 11320 }, { "epoch": 1.4401475639231651, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.529308557510376, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.87836092710495, "num_tokens": 429930322.0, "step": 11321 }, { "epoch": 1.4402747742017554, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.940035343170166, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8748267889022827, "num_tokens": 429965708.0, "step": 11322 }, { "epoch": 1.440401984480346, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5958070755004883, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8693243265151978, "num_tokens": 430007266.0, "step": 11323 }, { "epoch": 1.4405291947589365, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.737227201461792, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8627194166183472, "num_tokens": 430044470.0, "step": 11324 }, { "epoch": 1.440656405037527, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6028807163238525, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.873358964920044, "num_tokens": 430079776.0, "step": 11325 }, { "epoch": 1.4407836153161175, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5291218757629395, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8782973289489746, "num_tokens": 430110697.0, "step": 11326 }, { "epoch": 1.440910825594708, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.570050001144409, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8716119527816772, "num_tokens": 430150350.0, "step": 11327 }, { "epoch": 1.4410380358732986, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.375732183456421, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8543840646743774, "num_tokens": 430185415.0, "step": 11328 }, { "epoch": 1.4411652461518891, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5749123096466064, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.866510272026062, "num_tokens": 430227865.0, "step": 11329 }, { "epoch": 1.4412924564304797, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2545628547668457, "learning_rate": 1e-05, "loss": 0.4167, "mean_token_accuracy": 0.8801393508911133, "num_tokens": 430267821.0, "step": 11330 }, { "epoch": 1.44141966670907, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6172852516174316, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8752858638763428, "num_tokens": 430307251.0, "step": 11331 }, { "epoch": 1.4415468769876605, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1849801540374756, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8570986390113831, "num_tokens": 430347842.0, "step": 11332 }, { "epoch": 1.441674087266251, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7311322689056396, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.863492488861084, "num_tokens": 430383175.0, "step": 11333 }, { "epoch": 1.4418012975448415, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1467931270599365, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8667010068893433, "num_tokens": 430414319.0, "step": 11334 }, { "epoch": 1.441928507823432, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8657960891723633, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8829054832458496, "num_tokens": 430454142.0, "step": 11335 }, { "epoch": 1.4420557181020226, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.132474660873413, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8595185279846191, "num_tokens": 430496980.0, "step": 11336 }, { "epoch": 1.4421829283806131, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.065688610076904, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.861360490322113, "num_tokens": 430533645.0, "step": 11337 }, { "epoch": 1.4423101386592037, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.268214464187622, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8717373609542847, "num_tokens": 430567744.0, "step": 11338 }, { "epoch": 1.4424373489377942, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1739020347595215, "learning_rate": 1e-05, "loss": 0.5476, "mean_token_accuracy": 0.8451350927352905, "num_tokens": 430603887.0, "step": 11339 }, { "epoch": 1.4425645592163847, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4865832328796387, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8708011507987976, "num_tokens": 430638748.0, "step": 11340 }, { "epoch": 1.4426917694949752, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.075331687927246, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.870541512966156, "num_tokens": 430677799.0, "step": 11341 }, { "epoch": 1.4428189797735658, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6050124168395996, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8658012747764587, "num_tokens": 430720816.0, "step": 11342 }, { "epoch": 1.4429461900521563, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8990159034729004, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8720199465751648, "num_tokens": 430755220.0, "step": 11343 }, { "epoch": 1.4430734003307468, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.609037399291992, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8720889091491699, "num_tokens": 430787737.0, "step": 11344 }, { "epoch": 1.4432006106093374, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7064614295959473, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8701394200325012, "num_tokens": 430823817.0, "step": 11345 }, { "epoch": 1.4433278208879279, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.54801344871521, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8650944232940674, "num_tokens": 430863416.0, "step": 11346 }, { "epoch": 1.4434550311665182, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5719077587127686, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8624581098556519, "num_tokens": 430894017.0, "step": 11347 }, { "epoch": 1.4435822414451087, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.476318120956421, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8664641380310059, "num_tokens": 430935023.0, "step": 11348 }, { "epoch": 1.4437094517236992, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.486360549926758, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8705552816390991, "num_tokens": 430971663.0, "step": 11349 }, { "epoch": 1.4438366620022898, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.403221845626831, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8656320571899414, "num_tokens": 431018919.0, "step": 11350 }, { "epoch": 1.4439638722808803, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4578733444213867, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8786875009536743, "num_tokens": 431050076.0, "step": 11351 }, { "epoch": 1.4440910825594708, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.35432505607605, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8574109077453613, "num_tokens": 431085535.0, "step": 11352 }, { "epoch": 1.4442182928380614, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4529223442077637, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8760653734207153, "num_tokens": 431123187.0, "step": 11353 }, { "epoch": 1.4443455031166519, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.305743932723999, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8597816228866577, "num_tokens": 431157867.0, "step": 11354 }, { "epoch": 1.4444727133952424, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.493684768676758, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8645789623260498, "num_tokens": 431191344.0, "step": 11355 }, { "epoch": 1.4445999236738327, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.233668804168701, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8724863529205322, "num_tokens": 431231019.0, "step": 11356 }, { "epoch": 1.4447271339524232, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.557708740234375, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8698912858963013, "num_tokens": 431265660.0, "step": 11357 }, { "epoch": 1.4448543442310138, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1683545112609863, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8804970979690552, "num_tokens": 431298666.0, "step": 11358 }, { "epoch": 1.4449815545096043, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.702439308166504, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8859175443649292, "num_tokens": 431327516.0, "step": 11359 }, { "epoch": 1.4451087647881948, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.0915822982788086, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.868552565574646, "num_tokens": 431366244.0, "step": 11360 }, { "epoch": 1.4452359750667854, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9042139053344727, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8623430728912354, "num_tokens": 431402440.0, "step": 11361 }, { "epoch": 1.445363185345376, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.0991299152374268, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8863990306854248, "num_tokens": 431436628.0, "step": 11362 }, { "epoch": 1.4454903956239664, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.140401840209961, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8832211494445801, "num_tokens": 431476343.0, "step": 11363 }, { "epoch": 1.445617605902557, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.27608323097229, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8781706094741821, "num_tokens": 431513741.0, "step": 11364 }, { "epoch": 1.4457448161811475, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.258736610412598, "learning_rate": 1e-05, "loss": 0.5011, "mean_token_accuracy": 0.8575084805488586, "num_tokens": 431556227.0, "step": 11365 }, { "epoch": 1.445872026459738, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5609586238861084, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8731205463409424, "num_tokens": 431592004.0, "step": 11366 }, { "epoch": 1.4459992367383285, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.126652717590332, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.867172122001648, "num_tokens": 431631286.0, "step": 11367 }, { "epoch": 1.446126447016919, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.681614637374878, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.8678622245788574, "num_tokens": 431669269.0, "step": 11368 }, { "epoch": 1.4462536572955096, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.905017852783203, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8625718355178833, "num_tokens": 431708061.0, "step": 11369 }, { "epoch": 1.4463808675741001, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.637667417526245, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.874449610710144, "num_tokens": 431745855.0, "step": 11370 }, { "epoch": 1.4465080778526904, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7143919467926025, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8712873458862305, "num_tokens": 431784481.0, "step": 11371 }, { "epoch": 1.446635288131281, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5575790405273438, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8633126616477966, "num_tokens": 431822724.0, "step": 11372 }, { "epoch": 1.4467624984098715, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.572248697280884, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8758136630058289, "num_tokens": 431859415.0, "step": 11373 }, { "epoch": 1.446889708688462, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4719114303588867, "learning_rate": 1e-05, "loss": 0.3967, "mean_token_accuracy": 0.8906700015068054, "num_tokens": 431899432.0, "step": 11374 }, { "epoch": 1.4470169189670525, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4943974018096924, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8760687112808228, "num_tokens": 431935546.0, "step": 11375 }, { "epoch": 1.447144129245643, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.409634590148926, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8723480105400085, "num_tokens": 431972366.0, "step": 11376 }, { "epoch": 1.4472713395242336, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4649815559387207, "learning_rate": 1e-05, "loss": 0.4788, "mean_token_accuracy": 0.8584190607070923, "num_tokens": 432007795.0, "step": 11377 }, { "epoch": 1.4473985498028241, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3526294231414795, "learning_rate": 1e-05, "loss": 0.5139, "mean_token_accuracy": 0.85007643699646, "num_tokens": 432047589.0, "step": 11378 }, { "epoch": 1.4475257600814146, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.461172342300415, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8796604871749878, "num_tokens": 432085615.0, "step": 11379 }, { "epoch": 1.447652970360005, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.288438081741333, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8751500844955444, "num_tokens": 432121112.0, "step": 11380 }, { "epoch": 1.4477801806385955, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5018837451934814, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8701140284538269, "num_tokens": 432156429.0, "step": 11381 }, { "epoch": 1.447907390917186, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2238175868988037, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8550077676773071, "num_tokens": 432195653.0, "step": 11382 }, { "epoch": 1.4480346011957765, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5892465114593506, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8707402944564819, "num_tokens": 432230997.0, "step": 11383 }, { "epoch": 1.448161811474367, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1353561878204346, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8621786832809448, "num_tokens": 432270868.0, "step": 11384 }, { "epoch": 1.4482890217529576, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.740628719329834, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8727655410766602, "num_tokens": 432309999.0, "step": 11385 }, { "epoch": 1.4484162320315481, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.074066400527954, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8823683261871338, "num_tokens": 432345240.0, "step": 11386 }, { "epoch": 1.4485434423101387, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9637296199798584, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8616254925727844, "num_tokens": 432391478.0, "step": 11387 }, { "epoch": 1.4486706525887292, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.111945629119873, "learning_rate": 1e-05, "loss": 0.3975, "mean_token_accuracy": 0.8855501413345337, "num_tokens": 432430601.0, "step": 11388 }, { "epoch": 1.4487978628673197, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.224949836730957, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8680596351623535, "num_tokens": 432467260.0, "step": 11389 }, { "epoch": 1.4489250731459102, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3575499057769775, "learning_rate": 1e-05, "loss": 0.4961, "mean_token_accuracy": 0.8541706204414368, "num_tokens": 432506270.0, "step": 11390 }, { "epoch": 1.4490522834245008, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2834343910217285, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8672356605529785, "num_tokens": 432543688.0, "step": 11391 }, { "epoch": 1.4491794937030913, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.63065505027771, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8787081241607666, "num_tokens": 432583522.0, "step": 11392 }, { "epoch": 1.4493067039816818, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.106125831604004, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.884760320186615, "num_tokens": 432622471.0, "step": 11393 }, { "epoch": 1.4494339142602723, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6941492557525635, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.881889820098877, "num_tokens": 432666689.0, "step": 11394 }, { "epoch": 1.4495611245388629, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.8938486576080322, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.875421404838562, "num_tokens": 432707816.0, "step": 11395 }, { "epoch": 1.4496883348174532, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.633751153945923, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.884975790977478, "num_tokens": 432746873.0, "step": 11396 }, { "epoch": 1.4498155450960437, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7284793853759766, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.883622407913208, "num_tokens": 432777825.0, "step": 11397 }, { "epoch": 1.4499427553746342, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5384864807128906, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8686254024505615, "num_tokens": 432816806.0, "step": 11398 }, { "epoch": 1.4500699656532248, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6137542724609375, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8797601461410522, "num_tokens": 432855726.0, "step": 11399 }, { "epoch": 1.4501971759318153, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4656336307525635, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8613613843917847, "num_tokens": 432891526.0, "step": 11400 }, { "epoch": 1.4503243862104058, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5528531074523926, "learning_rate": 1e-05, "loss": 0.3751, "mean_token_accuracy": 0.8933168649673462, "num_tokens": 432921995.0, "step": 11401 }, { "epoch": 1.4504515964889964, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.374642848968506, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8665176630020142, "num_tokens": 432962120.0, "step": 11402 }, { "epoch": 1.4505788067675869, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5458641052246094, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8686767816543579, "num_tokens": 433001944.0, "step": 11403 }, { "epoch": 1.4507060170461774, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.307279586791992, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8520392179489136, "num_tokens": 433040984.0, "step": 11404 }, { "epoch": 1.4508332273247677, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5796961784362793, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8673487901687622, "num_tokens": 433077914.0, "step": 11405 }, { "epoch": 1.4509604376033582, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2293667793273926, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8791148662567139, "num_tokens": 433111219.0, "step": 11406 }, { "epoch": 1.4510876478819488, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6609714031219482, "learning_rate": 1e-05, "loss": 0.4033, "mean_token_accuracy": 0.8863389492034912, "num_tokens": 433143204.0, "step": 11407 }, { "epoch": 1.4512148581605393, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.166022300720215, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8865437507629395, "num_tokens": 433175361.0, "step": 11408 }, { "epoch": 1.4513420684391298, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.794858694076538, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8803480863571167, "num_tokens": 433215303.0, "step": 11409 }, { "epoch": 1.4514692787177204, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.129281759262085, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8770806789398193, "num_tokens": 433255301.0, "step": 11410 }, { "epoch": 1.4515964889963109, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9908883571624756, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8618298768997192, "num_tokens": 433291110.0, "step": 11411 }, { "epoch": 1.4517236992749014, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2266693115234375, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8707795143127441, "num_tokens": 433330910.0, "step": 11412 }, { "epoch": 1.451850909553492, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.123251438140869, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8724029064178467, "num_tokens": 433364297.0, "step": 11413 }, { "epoch": 1.4519781198320825, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4157867431640625, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8686072826385498, "num_tokens": 433405900.0, "step": 11414 }, { "epoch": 1.452105330110673, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.086289882659912, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8621348142623901, "num_tokens": 433442639.0, "step": 11415 }, { "epoch": 1.4522325403892635, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.59515118598938, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8667654991149902, "num_tokens": 433471657.0, "step": 11416 }, { "epoch": 1.452359750667854, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.885324001312256, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.85191410779953, "num_tokens": 433510936.0, "step": 11417 }, { "epoch": 1.4524869609464446, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6006357669830322, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8689767122268677, "num_tokens": 433546940.0, "step": 11418 }, { "epoch": 1.4526141712250351, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.668239116668701, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8724113702774048, "num_tokens": 433588701.0, "step": 11419 }, { "epoch": 1.4527413815036254, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5633063316345215, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8697936534881592, "num_tokens": 433625678.0, "step": 11420 }, { "epoch": 1.452868591782216, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5110373497009277, "learning_rate": 1e-05, "loss": 0.5169, "mean_token_accuracy": 0.8547846674919128, "num_tokens": 433660624.0, "step": 11421 }, { "epoch": 1.4529958020608065, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.527071952819824, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8787695169448853, "num_tokens": 433702737.0, "step": 11422 }, { "epoch": 1.453123012339397, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.39426326751709, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.869200587272644, "num_tokens": 433741866.0, "step": 11423 }, { "epoch": 1.4532502226179875, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.506401777267456, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8622622489929199, "num_tokens": 433779869.0, "step": 11424 }, { "epoch": 1.453377432896578, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3135194778442383, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8821889162063599, "num_tokens": 433819629.0, "step": 11425 }, { "epoch": 1.4535046431751686, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5184366703033447, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8778927326202393, "num_tokens": 433855330.0, "step": 11426 }, { "epoch": 1.4536318534537591, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2604644298553467, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.863923966884613, "num_tokens": 433896130.0, "step": 11427 }, { "epoch": 1.4537590637323496, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.553093671798706, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8800548911094666, "num_tokens": 433928514.0, "step": 11428 }, { "epoch": 1.45388627401094, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1827449798583984, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8671970963478088, "num_tokens": 433969213.0, "step": 11429 }, { "epoch": 1.4540134842895305, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.654822587966919, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.8571604490280151, "num_tokens": 434007902.0, "step": 11430 }, { "epoch": 1.454140694568121, "ewc_loss": 0.068359375, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1280617713928223, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8578863739967346, "num_tokens": 434043243.0, "step": 11431 }, { "epoch": 1.4542679048467115, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7995352745056152, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8846768736839294, "num_tokens": 434081075.0, "step": 11432 }, { "epoch": 1.454395115125302, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000244140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0917105674743652, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8690566420555115, "num_tokens": 434121230.0, "step": 11433 }, { "epoch": 1.4545223254038926, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.016153812408447, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8674963712692261, "num_tokens": 434158324.0, "step": 11434 }, { "epoch": 1.4546495356824831, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.15999698638916, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8725730776786804, "num_tokens": 434203910.0, "step": 11435 }, { "epoch": 1.4547767459610736, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.224946022033691, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8819588422775269, "num_tokens": 434238915.0, "step": 11436 }, { "epoch": 1.4549039562396642, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4161858558654785, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8736535310745239, "num_tokens": 434280049.0, "step": 11437 }, { "epoch": 1.4550311665182547, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.21801233291626, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.880355954170227, "num_tokens": 434323980.0, "step": 11438 }, { "epoch": 1.4551583767968452, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.641378402709961, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8665390014648438, "num_tokens": 434363386.0, "step": 11439 }, { "epoch": 1.4552855870754358, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.02430534362793, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8813806772232056, "num_tokens": 434399480.0, "step": 11440 }, { "epoch": 1.4554127973540263, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.66719388961792, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8685951232910156, "num_tokens": 434445662.0, "step": 11441 }, { "epoch": 1.4555400076326168, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.82376766204834, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8729681968688965, "num_tokens": 434480926.0, "step": 11442 }, { "epoch": 1.4556672179112073, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6108319759368896, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.860793948173523, "num_tokens": 434520018.0, "step": 11443 }, { "epoch": 1.4557944281897979, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.652031421661377, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8589321970939636, "num_tokens": 434555163.0, "step": 11444 }, { "epoch": 1.4559216384683882, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5398313999176025, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8817338943481445, "num_tokens": 434586353.0, "step": 11445 }, { "epoch": 1.4560488487469787, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5259571075439453, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8776354789733887, "num_tokens": 434630558.0, "step": 11446 }, { "epoch": 1.4561760590255692, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.462296962738037, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8704157471656799, "num_tokens": 434672949.0, "step": 11447 }, { "epoch": 1.4563032693041598, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4624712467193604, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8658820390701294, "num_tokens": 434709207.0, "step": 11448 }, { "epoch": 1.4564304795827503, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.396205186843872, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8816372156143188, "num_tokens": 434745787.0, "step": 11449 }, { "epoch": 1.4565576898613408, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4429423809051514, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8716825246810913, "num_tokens": 434785419.0, "step": 11450 }, { "epoch": 1.4566849001399313, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3602914810180664, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8633257150650024, "num_tokens": 434818217.0, "step": 11451 }, { "epoch": 1.4568121104185219, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4334309101104736, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8614636063575745, "num_tokens": 434855654.0, "step": 11452 }, { "epoch": 1.4569393206971124, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3124334812164307, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8571648001670837, "num_tokens": 434890657.0, "step": 11453 }, { "epoch": 1.4570665309757027, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4488532543182373, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.881723940372467, "num_tokens": 434929164.0, "step": 11454 }, { "epoch": 1.4571937412542932, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2613418102264404, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8635096549987793, "num_tokens": 434966017.0, "step": 11455 }, { "epoch": 1.4573209515328838, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.494663715362549, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8796848058700562, "num_tokens": 435002076.0, "step": 11456 }, { "epoch": 1.4574481618114743, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2101972103118896, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8776427507400513, "num_tokens": 435030058.0, "step": 11457 }, { "epoch": 1.4575753720900648, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5805106163024902, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8652652502059937, "num_tokens": 435067816.0, "step": 11458 }, { "epoch": 1.4577025823686554, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1324331760406494, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8742126226425171, "num_tokens": 435106078.0, "step": 11459 }, { "epoch": 1.4578297926472459, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.720609664916992, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8814604878425598, "num_tokens": 435144558.0, "step": 11460 }, { "epoch": 1.4579570029258364, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.065415382385254, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.867757260799408, "num_tokens": 435183083.0, "step": 11461 }, { "epoch": 1.458084213204427, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.952317953109741, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8878465294837952, "num_tokens": 435223459.0, "step": 11462 }, { "epoch": 1.4582114234830175, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.086942434310913, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.877100944519043, "num_tokens": 435261491.0, "step": 11463 }, { "epoch": 1.458338633761608, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.234492778778076, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8612637519836426, "num_tokens": 435299634.0, "step": 11464 }, { "epoch": 1.4584658440401985, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3311805725097656, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8553887605667114, "num_tokens": 435339645.0, "step": 11465 }, { "epoch": 1.458593054318789, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.336906909942627, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8699409365653992, "num_tokens": 435382635.0, "step": 11466 }, { "epoch": 1.4587202645973796, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.649998188018799, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8616937398910522, "num_tokens": 435421653.0, "step": 11467 }, { "epoch": 1.45884747487597, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.169643402099609, "learning_rate": 1e-05, "loss": 0.4943, "mean_token_accuracy": 0.8608059883117676, "num_tokens": 435460135.0, "step": 11468 }, { "epoch": 1.4589746851545604, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7377684116363525, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8859761357307434, "num_tokens": 435497717.0, "step": 11469 }, { "epoch": 1.459101895433151, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9614248275756836, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.859619677066803, "num_tokens": 435535247.0, "step": 11470 }, { "epoch": 1.4592291057117415, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.659165382385254, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8778607845306396, "num_tokens": 435576841.0, "step": 11471 }, { "epoch": 1.459356315990332, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.785322904586792, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8737702369689941, "num_tokens": 435618197.0, "step": 11472 }, { "epoch": 1.4594835262689225, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5621719360351562, "learning_rate": 1e-05, "loss": 0.5359, "mean_token_accuracy": 0.847160816192627, "num_tokens": 435663327.0, "step": 11473 }, { "epoch": 1.459610736547513, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.679842710494995, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8702353239059448, "num_tokens": 435697795.0, "step": 11474 }, { "epoch": 1.4597379468261036, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4631102085113525, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8615962266921997, "num_tokens": 435734308.0, "step": 11475 }, { "epoch": 1.4598651571046941, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.61928129196167, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8646607995033264, "num_tokens": 435772575.0, "step": 11476 }, { "epoch": 1.4599923673832846, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.361236810684204, "learning_rate": 1e-05, "loss": 0.407, "mean_token_accuracy": 0.882361888885498, "num_tokens": 435807578.0, "step": 11477 }, { "epoch": 1.460119577661875, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6400418281555176, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8703691959381104, "num_tokens": 435849035.0, "step": 11478 }, { "epoch": 1.4602467879404655, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.271435022354126, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8730398416519165, "num_tokens": 435884912.0, "step": 11479 }, { "epoch": 1.460373998219056, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.7011852264404297, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8795942068099976, "num_tokens": 435928822.0, "step": 11480 }, { "epoch": 1.4605012084976465, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2057836055755615, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8795068860054016, "num_tokens": 435963906.0, "step": 11481 }, { "epoch": 1.460628418776237, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8407042026519775, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8549810647964478, "num_tokens": 436002826.0, "step": 11482 }, { "epoch": 1.4607556290548276, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.187852382659912, "learning_rate": 1e-05, "loss": 0.3964, "mean_token_accuracy": 0.8855798840522766, "num_tokens": 436039043.0, "step": 11483 }, { "epoch": 1.4608828393334181, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9929847717285156, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8799750804901123, "num_tokens": 436078081.0, "step": 11484 }, { "epoch": 1.4610100496120086, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.307314157485962, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8559665083885193, "num_tokens": 436110415.0, "step": 11485 }, { "epoch": 1.4611372598905992, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.047876358032227, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8797931671142578, "num_tokens": 436152473.0, "step": 11486 }, { "epoch": 1.4612644701691897, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.445281744003296, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8684644103050232, "num_tokens": 436192410.0, "step": 11487 }, { "epoch": 1.4613916804477802, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.969205141067505, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8611068725585938, "num_tokens": 436230605.0, "step": 11488 }, { "epoch": 1.4615188907263708, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5490200519561768, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8786178231239319, "num_tokens": 436267951.0, "step": 11489 }, { "epoch": 1.4616461010049613, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7774441242218018, "learning_rate": 1e-05, "loss": 0.4974, "mean_token_accuracy": 0.8540889620780945, "num_tokens": 436306685.0, "step": 11490 }, { "epoch": 1.4617733112835518, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5687923431396484, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8746129274368286, "num_tokens": 436345568.0, "step": 11491 }, { "epoch": 1.4619005215621423, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5621860027313232, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8720923662185669, "num_tokens": 436387580.0, "step": 11492 }, { "epoch": 1.4620277318407329, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5687148571014404, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8671054840087891, "num_tokens": 436423000.0, "step": 11493 }, { "epoch": 1.4621549421193232, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.40362548828125, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.873035192489624, "num_tokens": 436464366.0, "step": 11494 }, { "epoch": 1.4622821523979137, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.57318115234375, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8591891527175903, "num_tokens": 436505385.0, "step": 11495 }, { "epoch": 1.4624093626765042, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.283919334411621, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8606245517730713, "num_tokens": 436546251.0, "step": 11496 }, { "epoch": 1.4625365729550948, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6146554946899414, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.860365629196167, "num_tokens": 436582136.0, "step": 11497 }, { "epoch": 1.4626637832336853, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1984879970550537, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.872194766998291, "num_tokens": 436620027.0, "step": 11498 }, { "epoch": 1.4627909935122758, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.7000577449798584, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8679830431938171, "num_tokens": 436656720.0, "step": 11499 }, { "epoch": 1.4629182037908663, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1359128952026367, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8790870904922485, "num_tokens": 436693813.0, "step": 11500 }, { "epoch": 1.4630454140694569, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8496406078338623, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8736963868141174, "num_tokens": 436732023.0, "step": 11501 }, { "epoch": 1.4631726243480474, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.125392436981201, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8654931783676147, "num_tokens": 436772330.0, "step": 11502 }, { "epoch": 1.4632998346266377, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0291571617126465, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8666069507598877, "num_tokens": 436814593.0, "step": 11503 }, { "epoch": 1.4634270449052282, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.23970365524292, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8711725473403931, "num_tokens": 436849951.0, "step": 11504 }, { "epoch": 1.4635542551838188, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.159540176391602, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.856247067451477, "num_tokens": 436889113.0, "step": 11505 }, { "epoch": 1.4636814654624093, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4486091136932373, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.881007194519043, "num_tokens": 436921009.0, "step": 11506 }, { "epoch": 1.4638086757409998, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.095951557159424, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8729727268218994, "num_tokens": 436958425.0, "step": 11507 }, { "epoch": 1.4639358860195903, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5819404125213623, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8735047578811646, "num_tokens": 436992995.0, "step": 11508 }, { "epoch": 1.4640630962981809, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9303383827209473, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8776794075965881, "num_tokens": 437028884.0, "step": 11509 }, { "epoch": 1.4641903065767714, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.592790126800537, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8753356337547302, "num_tokens": 437071753.0, "step": 11510 }, { "epoch": 1.464317516855362, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7449774742126465, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8795682191848755, "num_tokens": 437108783.0, "step": 11511 }, { "epoch": 1.4644447271339525, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5467369556427, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8753777742385864, "num_tokens": 437143234.0, "step": 11512 }, { "epoch": 1.464571937412543, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6084096431732178, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8654850125312805, "num_tokens": 437182363.0, "step": 11513 }, { "epoch": 1.4646991476911335, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.477395534515381, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8760588765144348, "num_tokens": 437224680.0, "step": 11514 }, { "epoch": 1.464826357969724, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5312631130218506, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8723383545875549, "num_tokens": 437261154.0, "step": 11515 }, { "epoch": 1.4649535682483146, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4090726375579834, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8707578182220459, "num_tokens": 437301654.0, "step": 11516 }, { "epoch": 1.465080778526905, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4998414516448975, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8693815469741821, "num_tokens": 437338437.0, "step": 11517 }, { "epoch": 1.4652079888054954, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3389604091644287, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8657124638557434, "num_tokens": 437384258.0, "step": 11518 }, { "epoch": 1.465335199084086, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.518230676651001, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8574906587600708, "num_tokens": 437421762.0, "step": 11519 }, { "epoch": 1.4654624093626765, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.279045581817627, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8795080780982971, "num_tokens": 437456797.0, "step": 11520 }, { "epoch": 1.465589619641267, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5764381885528564, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8815402984619141, "num_tokens": 437484435.0, "step": 11521 }, { "epoch": 1.4657168299198575, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.211474895477295, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8707773685455322, "num_tokens": 437522573.0, "step": 11522 }, { "epoch": 1.465844040198448, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.661983013153076, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8589031100273132, "num_tokens": 437560204.0, "step": 11523 }, { "epoch": 1.4659712504770386, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1484549045562744, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8665628433227539, "num_tokens": 437595845.0, "step": 11524 }, { "epoch": 1.466098460755629, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8070318698883057, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8755216598510742, "num_tokens": 437631985.0, "step": 11525 }, { "epoch": 1.4662256710342196, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.12835693359375, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8622055053710938, "num_tokens": 437677024.0, "step": 11526 }, { "epoch": 1.46635288131281, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9840879440307617, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8867483139038086, "num_tokens": 437714980.0, "step": 11527 }, { "epoch": 1.4664800915914005, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2247579097747803, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8686560392379761, "num_tokens": 437752444.0, "step": 11528 }, { "epoch": 1.466607301869991, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.106786727905273, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8651711344718933, "num_tokens": 437787032.0, "step": 11529 }, { "epoch": 1.4667345121485815, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.402832508087158, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8690886497497559, "num_tokens": 437827560.0, "step": 11530 }, { "epoch": 1.466861722427172, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.066324234008789, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8666424751281738, "num_tokens": 437865339.0, "step": 11531 }, { "epoch": 1.4669889327057626, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.560289144515991, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8703179359436035, "num_tokens": 437899769.0, "step": 11532 }, { "epoch": 1.467116142984353, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8916289806365967, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8739449977874756, "num_tokens": 437934738.0, "step": 11533 }, { "epoch": 1.4672433532629436, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.576185464859009, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8649738430976868, "num_tokens": 437975948.0, "step": 11534 }, { "epoch": 1.4673705635415342, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7072033882141113, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8788695335388184, "num_tokens": 438010081.0, "step": 11535 }, { "epoch": 1.4674977738201247, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5570404529571533, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8745052814483643, "num_tokens": 438044069.0, "step": 11536 }, { "epoch": 1.4676249840987152, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5312340259552, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8743002414703369, "num_tokens": 438080866.0, "step": 11537 }, { "epoch": 1.4677521943773058, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5137314796447754, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8705281019210815, "num_tokens": 438120550.0, "step": 11538 }, { "epoch": 1.4678794046558963, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4202866554260254, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8674455881118774, "num_tokens": 438156968.0, "step": 11539 }, { "epoch": 1.4680066149344868, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4927072525024414, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8730474710464478, "num_tokens": 438191470.0, "step": 11540 }, { "epoch": 1.4681338252130773, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.356187343597412, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8583226203918457, "num_tokens": 438225944.0, "step": 11541 }, { "epoch": 1.4682610354916679, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.479386329650879, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8712947368621826, "num_tokens": 438263759.0, "step": 11542 }, { "epoch": 1.4683882457702582, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.514641046524048, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8734047412872314, "num_tokens": 438302371.0, "step": 11543 }, { "epoch": 1.4685154560488487, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.346437931060791, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8534165024757385, "num_tokens": 438342006.0, "step": 11544 }, { "epoch": 1.4686426663274392, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.424694776535034, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8629364371299744, "num_tokens": 438375649.0, "step": 11545 }, { "epoch": 1.4687698766060298, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.246335983276367, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8778420686721802, "num_tokens": 438417566.0, "step": 11546 }, { "epoch": 1.4688970868846203, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.488656997680664, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8780215978622437, "num_tokens": 438452219.0, "step": 11547 }, { "epoch": 1.4690242971632108, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.246389627456665, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.8834774494171143, "num_tokens": 438490503.0, "step": 11548 }, { "epoch": 1.4691515074418013, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.6237382888793945, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8846471309661865, "num_tokens": 438521792.0, "step": 11549 }, { "epoch": 1.4692787177203919, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1122372150421143, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8787478804588318, "num_tokens": 438557408.0, "step": 11550 }, { "epoch": 1.4694059279989824, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7423906326293945, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8838722705841064, "num_tokens": 438601424.0, "step": 11551 }, { "epoch": 1.4695331382775727, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.040599822998047, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8760590553283691, "num_tokens": 438638772.0, "step": 11552 }, { "epoch": 1.4696603485561632, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.023873805999756, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.867048442363739, "num_tokens": 438676906.0, "step": 11553 }, { "epoch": 1.4697875588347538, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1137335300445557, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8557594418525696, "num_tokens": 438715790.0, "step": 11554 }, { "epoch": 1.4699147691133443, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.273365020751953, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8756875991821289, "num_tokens": 438757739.0, "step": 11555 }, { "epoch": 1.4700419793919348, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.385690450668335, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8719875812530518, "num_tokens": 438797336.0, "step": 11556 }, { "epoch": 1.4701691896705253, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.347358226776123, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8679922819137573, "num_tokens": 438833452.0, "step": 11557 }, { "epoch": 1.4702963999491159, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.686108350753784, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8725699782371521, "num_tokens": 438874346.0, "step": 11558 }, { "epoch": 1.4704236102277064, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.152737617492676, "learning_rate": 1e-05, "loss": 0.4975, "mean_token_accuracy": 0.8589724898338318, "num_tokens": 438908888.0, "step": 11559 }, { "epoch": 1.470550820506297, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.754749298095703, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8747397065162659, "num_tokens": 438953127.0, "step": 11560 }, { "epoch": 1.4706780307848875, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9351158142089844, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8684654235839844, "num_tokens": 438986485.0, "step": 11561 }, { "epoch": 1.470805241063478, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6640784740448, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8697482347488403, "num_tokens": 439030958.0, "step": 11562 }, { "epoch": 1.4709324513420685, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.765378713607788, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8644229769706726, "num_tokens": 439069139.0, "step": 11563 }, { "epoch": 1.471059661620659, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5675556659698486, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8778706789016724, "num_tokens": 439106526.0, "step": 11564 }, { "epoch": 1.4711868718992496, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.648447036743164, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8723803758621216, "num_tokens": 439146037.0, "step": 11565 }, { "epoch": 1.47131408217784, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.45849347114563, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8821072578430176, "num_tokens": 439177239.0, "step": 11566 }, { "epoch": 1.4714412924564304, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6074798107147217, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8727635145187378, "num_tokens": 439211009.0, "step": 11567 }, { "epoch": 1.471568502735021, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.361234188079834, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.873310923576355, "num_tokens": 439250656.0, "step": 11568 }, { "epoch": 1.4716957130136115, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.624850034713745, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.85343337059021, "num_tokens": 439285103.0, "step": 11569 }, { "epoch": 1.471822923292202, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.27044677734375, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8616257309913635, "num_tokens": 439325104.0, "step": 11570 }, { "epoch": 1.4719501335707925, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7007956504821777, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.8571212291717529, "num_tokens": 439357760.0, "step": 11571 }, { "epoch": 1.472077343849383, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.192115545272827, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8672629594802856, "num_tokens": 439397823.0, "step": 11572 }, { "epoch": 1.4722045541279736, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.839535713195801, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8626521825790405, "num_tokens": 439435989.0, "step": 11573 }, { "epoch": 1.472331764406564, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.172308921813965, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.879991352558136, "num_tokens": 439474272.0, "step": 11574 }, { "epoch": 1.4724589746851546, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.009891033172607, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8809553384780884, "num_tokens": 439512729.0, "step": 11575 }, { "epoch": 1.472586184963745, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.281991958618164, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8607105016708374, "num_tokens": 439550368.0, "step": 11576 }, { "epoch": 1.4727133952423355, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.101568698883057, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.870781660079956, "num_tokens": 439591244.0, "step": 11577 }, { "epoch": 1.472840605520926, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002460479736328125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4595820903778076, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8669325113296509, "num_tokens": 439632389.0, "step": 11578 }, { "epoch": 1.4729678157995165, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.310948848724365, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.8596316576004028, "num_tokens": 439674543.0, "step": 11579 }, { "epoch": 1.473095026078107, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7220962047576904, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8693569898605347, "num_tokens": 439709009.0, "step": 11580 }, { "epoch": 1.4732222363566976, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7966930866241455, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8742243051528931, "num_tokens": 439754154.0, "step": 11581 }, { "epoch": 1.473349446635288, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.648169994354248, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8902308940887451, "num_tokens": 439794986.0, "step": 11582 }, { "epoch": 1.4734766569138786, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.594496726989746, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8650670051574707, "num_tokens": 439833402.0, "step": 11583 }, { "epoch": 1.4736038671924692, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.620213747024536, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8829199075698853, "num_tokens": 439878670.0, "step": 11584 }, { "epoch": 1.4737310774710597, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4015331268310547, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8791333436965942, "num_tokens": 439916505.0, "step": 11585 }, { "epoch": 1.4738582877496502, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6302428245544434, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8731323480606079, "num_tokens": 439955833.0, "step": 11586 }, { "epoch": 1.4739854980282407, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2801263332366943, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8831044435501099, "num_tokens": 439994105.0, "step": 11587 }, { "epoch": 1.4741127083068313, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6759402751922607, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.876802384853363, "num_tokens": 440030298.0, "step": 11588 }, { "epoch": 1.4742399185854218, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.209641218185425, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8763782978057861, "num_tokens": 440064792.0, "step": 11589 }, { "epoch": 1.4743671288640123, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7800517082214355, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8709990978240967, "num_tokens": 440097886.0, "step": 11590 }, { "epoch": 1.4744943391426026, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.166456937789917, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8760803937911987, "num_tokens": 440133488.0, "step": 11591 }, { "epoch": 1.4746215494211932, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9121248722076416, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8769148588180542, "num_tokens": 440169448.0, "step": 11592 }, { "epoch": 1.4747487596997837, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.190751075744629, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8797481060028076, "num_tokens": 440204755.0, "step": 11593 }, { "epoch": 1.4748759699783742, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0669379234313965, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8679581880569458, "num_tokens": 440239129.0, "step": 11594 }, { "epoch": 1.4750031802569648, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3153200149536133, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8728302717208862, "num_tokens": 440280962.0, "step": 11595 }, { "epoch": 1.4751303905355553, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.108748912811279, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8755759000778198, "num_tokens": 440315900.0, "step": 11596 }, { "epoch": 1.4752576008141458, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4785826206207275, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.880096971988678, "num_tokens": 440356699.0, "step": 11597 }, { "epoch": 1.4753848110927363, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0127272605896, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8775142431259155, "num_tokens": 440398655.0, "step": 11598 }, { "epoch": 1.4755120213713269, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5820224285125732, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8624851703643799, "num_tokens": 440436204.0, "step": 11599 }, { "epoch": 1.4756392316499174, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.82822322845459, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.873844563961029, "num_tokens": 440478363.0, "step": 11600 }, { "epoch": 1.4757664419285077, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.553316116333008, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8831051588058472, "num_tokens": 440515454.0, "step": 11601 }, { "epoch": 1.4758936522070982, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6783244609832764, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.871591329574585, "num_tokens": 440553526.0, "step": 11602 }, { "epoch": 1.4760208624856888, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5040812492370605, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8611670732498169, "num_tokens": 440591446.0, "step": 11603 }, { "epoch": 1.4761480727642793, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.563829183578491, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8800956010818481, "num_tokens": 440630505.0, "step": 11604 }, { "epoch": 1.4762752830428698, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4374642372131348, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8846536874771118, "num_tokens": 440669318.0, "step": 11605 }, { "epoch": 1.4764024933214603, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.499969959259033, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.880579948425293, "num_tokens": 440711218.0, "step": 11606 }, { "epoch": 1.4765297036000509, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3645904064178467, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8710441589355469, "num_tokens": 440754581.0, "step": 11607 }, { "epoch": 1.4766569138786414, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4940381050109863, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8714176416397095, "num_tokens": 440795714.0, "step": 11608 }, { "epoch": 1.476784124157232, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3160922527313232, "learning_rate": 1e-05, "loss": 0.386, "mean_token_accuracy": 0.8861522674560547, "num_tokens": 440827423.0, "step": 11609 }, { "epoch": 1.4769113344358225, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5223984718322754, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.861424446105957, "num_tokens": 440868857.0, "step": 11610 }, { "epoch": 1.477038544714413, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2590537071228027, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8578232526779175, "num_tokens": 440904788.0, "step": 11611 }, { "epoch": 1.4771657549930035, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.572145462036133, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8711283206939697, "num_tokens": 440949013.0, "step": 11612 }, { "epoch": 1.477292965271594, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1970021724700928, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8600783348083496, "num_tokens": 440987392.0, "step": 11613 }, { "epoch": 1.4774201755501846, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6737940311431885, "learning_rate": 1e-05, "loss": 0.3609, "mean_token_accuracy": 0.8976230025291443, "num_tokens": 441024917.0, "step": 11614 }, { "epoch": 1.477547385828775, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1327831745147705, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8781121969223022, "num_tokens": 441066355.0, "step": 11615 }, { "epoch": 1.4776745961073654, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.85251522064209, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8648215532302856, "num_tokens": 441100192.0, "step": 11616 }, { "epoch": 1.477801806385956, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.128805160522461, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.862045168876648, "num_tokens": 441144627.0, "step": 11617 }, { "epoch": 1.4779290166645465, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.048930644989014, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.875505805015564, "num_tokens": 441181656.0, "step": 11618 }, { "epoch": 1.478056226943137, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2663233280181885, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8678349256515503, "num_tokens": 441215476.0, "step": 11619 }, { "epoch": 1.4781834372217275, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.129471778869629, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8654943704605103, "num_tokens": 441262919.0, "step": 11620 }, { "epoch": 1.478310647500318, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.451808452606201, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8690575957298279, "num_tokens": 441302602.0, "step": 11621 }, { "epoch": 1.4784378577789086, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.061325550079346, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8724765181541443, "num_tokens": 441335432.0, "step": 11622 }, { "epoch": 1.478565068057499, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5820443630218506, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8635921478271484, "num_tokens": 441377849.0, "step": 11623 }, { "epoch": 1.4786922783360896, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8682661056518555, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8716061115264893, "num_tokens": 441413225.0, "step": 11624 }, { "epoch": 1.47881948861468, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.589226722717285, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8770792484283447, "num_tokens": 441448041.0, "step": 11625 }, { "epoch": 1.4789466988932705, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6688196659088135, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8836492300033569, "num_tokens": 441483919.0, "step": 11626 }, { "epoch": 1.479073909171861, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5445163249969482, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8684996366500854, "num_tokens": 441518313.0, "step": 11627 }, { "epoch": 1.4792011194504515, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5206820964813232, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8678380250930786, "num_tokens": 441557248.0, "step": 11628 }, { "epoch": 1.479328329729042, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5126185417175293, "learning_rate": 1e-05, "loss": 0.5478, "mean_token_accuracy": 0.8418686985969543, "num_tokens": 441599759.0, "step": 11629 }, { "epoch": 1.4794555400076326, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4225833415985107, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8667699098587036, "num_tokens": 441631643.0, "step": 11630 }, { "epoch": 1.479582750286223, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.479883909225464, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8609790205955505, "num_tokens": 441669804.0, "step": 11631 }, { "epoch": 1.4797099605648136, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.352797746658325, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8603837490081787, "num_tokens": 441708661.0, "step": 11632 }, { "epoch": 1.4798371708434042, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.5768165588378906, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.871150016784668, "num_tokens": 441745510.0, "step": 11633 }, { "epoch": 1.4799643811219947, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2523953914642334, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8714902400970459, "num_tokens": 441781482.0, "step": 11634 }, { "epoch": 1.4800915914005852, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5783913135528564, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8813920617103577, "num_tokens": 441814299.0, "step": 11635 }, { "epoch": 1.4802188016791757, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.163275957107544, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8560546636581421, "num_tokens": 441859742.0, "step": 11636 }, { "epoch": 1.4803460119577663, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6995763778686523, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8817977905273438, "num_tokens": 441896563.0, "step": 11637 }, { "epoch": 1.4804732222363568, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.107285261154175, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8796929121017456, "num_tokens": 441935380.0, "step": 11638 }, { "epoch": 1.4806004325149473, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8759350776672363, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8620420694351196, "num_tokens": 441977077.0, "step": 11639 }, { "epoch": 1.4807276427935376, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1056020259857178, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8736081123352051, "num_tokens": 442020538.0, "step": 11640 }, { "epoch": 1.4808548530721282, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.114373683929443, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8737767934799194, "num_tokens": 442057093.0, "step": 11641 }, { "epoch": 1.4809820633507187, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2569706439971924, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8660640716552734, "num_tokens": 442092364.0, "step": 11642 }, { "epoch": 1.4811092736293092, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2415971755981445, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8652239441871643, "num_tokens": 442134181.0, "step": 11643 }, { "epoch": 1.4812364839078997, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.4996376037597656, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.8567344546318054, "num_tokens": 442174987.0, "step": 11644 }, { "epoch": 1.4813636941864903, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.183956623077393, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8768740892410278, "num_tokens": 442204214.0, "step": 11645 }, { "epoch": 1.4814909044650808, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.661632776260376, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8804972171783447, "num_tokens": 442243227.0, "step": 11646 }, { "epoch": 1.4816181147436713, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9752283096313477, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8717641830444336, "num_tokens": 442282107.0, "step": 11647 }, { "epoch": 1.4817453250222619, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.651144504547119, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8786801099777222, "num_tokens": 442320445.0, "step": 11648 }, { "epoch": 1.4818725353008524, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.793790102005005, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.865985631942749, "num_tokens": 442358626.0, "step": 11649 }, { "epoch": 1.4819997455794427, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.574864149093628, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.870769739151001, "num_tokens": 442405095.0, "step": 11650 }, { "epoch": 1.4821269558580332, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6457772254943848, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8705525398254395, "num_tokens": 442442131.0, "step": 11651 }, { "epoch": 1.4822541661366238, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5048184394836426, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8580124974250793, "num_tokens": 442479660.0, "step": 11652 }, { "epoch": 1.4823813764152143, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.551058769226074, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8666521310806274, "num_tokens": 442523659.0, "step": 11653 }, { "epoch": 1.4825085866938048, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.426076889038086, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.865548312664032, "num_tokens": 442564445.0, "step": 11654 }, { "epoch": 1.4826357969723953, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5277743339538574, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8727311491966248, "num_tokens": 442597662.0, "step": 11655 }, { "epoch": 1.4827630072509859, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.347848415374756, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8654969930648804, "num_tokens": 442633365.0, "step": 11656 }, { "epoch": 1.4828902175295764, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.538771629333496, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8696253299713135, "num_tokens": 442668389.0, "step": 11657 }, { "epoch": 1.483017427808167, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2963149547576904, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8729308843612671, "num_tokens": 442695089.0, "step": 11658 }, { "epoch": 1.4831446380867574, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.561687707901001, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8723943829536438, "num_tokens": 442738602.0, "step": 11659 }, { "epoch": 1.483271848365348, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.206348180770874, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8688115477561951, "num_tokens": 442777400.0, "step": 11660 }, { "epoch": 1.4833990586439385, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.666375160217285, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8745140433311462, "num_tokens": 442815502.0, "step": 11661 }, { "epoch": 1.483526268922529, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1435632705688477, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8615871667861938, "num_tokens": 442858418.0, "step": 11662 }, { "epoch": 1.4836534792011196, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8345162868499756, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8864769339561462, "num_tokens": 442896073.0, "step": 11663 }, { "epoch": 1.48378068947971, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1185004711151123, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.869010329246521, "num_tokens": 442940719.0, "step": 11664 }, { "epoch": 1.4839078997583004, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.057552814483643, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8804101347923279, "num_tokens": 442977724.0, "step": 11665 }, { "epoch": 1.484035110036891, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.233630418777466, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8808210492134094, "num_tokens": 443017754.0, "step": 11666 }, { "epoch": 1.4841623203154815, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.204807281494141, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8728984594345093, "num_tokens": 443048661.0, "step": 11667 }, { "epoch": 1.484289530594072, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4794840812683105, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8695541620254517, "num_tokens": 443094445.0, "step": 11668 }, { "epoch": 1.4844167408726625, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.121801853179932, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8829984664916992, "num_tokens": 443134468.0, "step": 11669 }, { "epoch": 1.484543951151253, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6403188705444336, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8737602233886719, "num_tokens": 443174458.0, "step": 11670 }, { "epoch": 1.4846711614298436, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9026906490325928, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8781592845916748, "num_tokens": 443212421.0, "step": 11671 }, { "epoch": 1.484798371708434, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6425998210906982, "learning_rate": 1e-05, "loss": 0.5302, "mean_token_accuracy": 0.8464181423187256, "num_tokens": 443252083.0, "step": 11672 }, { "epoch": 1.4849255819870246, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.698134660720825, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8629353046417236, "num_tokens": 443286776.0, "step": 11673 }, { "epoch": 1.485052792265615, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.580923080444336, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8783730864524841, "num_tokens": 443325734.0, "step": 11674 }, { "epoch": 1.4851800025442055, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5335919857025146, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8704657554626465, "num_tokens": 443362637.0, "step": 11675 }, { "epoch": 1.485307212822796, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.525613307952881, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8807685971260071, "num_tokens": 443401629.0, "step": 11676 }, { "epoch": 1.4854344231013865, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.422039747238159, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8872302770614624, "num_tokens": 443437619.0, "step": 11677 }, { "epoch": 1.485561633379977, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5105063915252686, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.8572711944580078, "num_tokens": 443472175.0, "step": 11678 }, { "epoch": 1.4856888436585676, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3484973907470703, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8658984899520874, "num_tokens": 443506904.0, "step": 11679 }, { "epoch": 1.485816053937158, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.501018762588501, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.859123170375824, "num_tokens": 443548141.0, "step": 11680 }, { "epoch": 1.4859432642157486, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2777750492095947, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8702930212020874, "num_tokens": 443586609.0, "step": 11681 }, { "epoch": 1.4860704744943392, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5181162357330322, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8744890689849854, "num_tokens": 443628792.0, "step": 11682 }, { "epoch": 1.4861976847729297, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2212612628936768, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.853492796421051, "num_tokens": 443671336.0, "step": 11683 }, { "epoch": 1.4863248950515202, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.586108684539795, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8701043128967285, "num_tokens": 443708987.0, "step": 11684 }, { "epoch": 1.4864521053301107, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.153838872909546, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8732266426086426, "num_tokens": 443749968.0, "step": 11685 }, { "epoch": 1.4865793156087013, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.716507911682129, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8745441436767578, "num_tokens": 443786287.0, "step": 11686 }, { "epoch": 1.4867065258872918, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0943708419799805, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8848458528518677, "num_tokens": 443820729.0, "step": 11687 }, { "epoch": 1.4868337361658823, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.914072036743164, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.8847792744636536, "num_tokens": 443859401.0, "step": 11688 }, { "epoch": 1.4869609464444726, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.0995702743530273, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8877694606781006, "num_tokens": 443895697.0, "step": 11689 }, { "epoch": 1.4870881567230632, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.16339111328125, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8770309686660767, "num_tokens": 443935229.0, "step": 11690 }, { "epoch": 1.4872153670016537, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3020052909851074, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8694298267364502, "num_tokens": 443967544.0, "step": 11691 }, { "epoch": 1.4873425772802442, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2638726234436035, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8679234981536865, "num_tokens": 444006973.0, "step": 11692 }, { "epoch": 1.4874697875588347, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5571746826171875, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8785051107406616, "num_tokens": 444041143.0, "step": 11693 }, { "epoch": 1.4875969978374253, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.147495746612549, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8633130788803101, "num_tokens": 444086279.0, "step": 11694 }, { "epoch": 1.4877242081160158, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.679903745651245, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.871832549571991, "num_tokens": 444123559.0, "step": 11695 }, { "epoch": 1.4878514183946063, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9435887336730957, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8602466583251953, "num_tokens": 444159197.0, "step": 11696 }, { "epoch": 1.4879786286731969, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6600911617279053, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8731226921081543, "num_tokens": 444190217.0, "step": 11697 }, { "epoch": 1.4881058389517874, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7669034004211426, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8866804242134094, "num_tokens": 444225241.0, "step": 11698 }, { "epoch": 1.4882330492303777, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5710649490356445, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8569939136505127, "num_tokens": 444263702.0, "step": 11699 }, { "epoch": 1.4883602595089682, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6095571517944336, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8730197548866272, "num_tokens": 444303911.0, "step": 11700 }, { "epoch": 1.4884874697875587, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.510984182357788, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8710405826568604, "num_tokens": 444344705.0, "step": 11701 }, { "epoch": 1.4886146800661493, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.522620439529419, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8842384815216064, "num_tokens": 444381598.0, "step": 11702 }, { "epoch": 1.4887418903447398, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4245827198028564, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8759632110595703, "num_tokens": 444413077.0, "step": 11703 }, { "epoch": 1.4888691006233303, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.480433702468872, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8746325373649597, "num_tokens": 444453028.0, "step": 11704 }, { "epoch": 1.4889963109019209, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3648974895477295, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8683785200119019, "num_tokens": 444486382.0, "step": 11705 }, { "epoch": 1.4891235211805114, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4855873584747314, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8799659013748169, "num_tokens": 444523299.0, "step": 11706 }, { "epoch": 1.489250731459102, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2929043769836426, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8726972341537476, "num_tokens": 444561772.0, "step": 11707 }, { "epoch": 1.4893779417376924, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.524306058883667, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8585641980171204, "num_tokens": 444603056.0, "step": 11708 }, { "epoch": 1.489505152016283, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2228569984436035, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8848172426223755, "num_tokens": 444640469.0, "step": 11709 }, { "epoch": 1.4896323622948735, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6154651641845703, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.871519923210144, "num_tokens": 444680737.0, "step": 11710 }, { "epoch": 1.489759572573464, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1576032638549805, "learning_rate": 1e-05, "loss": 0.3977, "mean_token_accuracy": 0.8844975233078003, "num_tokens": 444713828.0, "step": 11711 }, { "epoch": 1.4898867828520546, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.765159845352173, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8635843992233276, "num_tokens": 444755823.0, "step": 11712 }, { "epoch": 1.490013993130645, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.0985653400421143, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8758100271224976, "num_tokens": 444791558.0, "step": 11713 }, { "epoch": 1.4901412034092354, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9841866493225098, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8764234185218811, "num_tokens": 444832037.0, "step": 11714 }, { "epoch": 1.490268413687826, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.1634604930877686, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8591890335083008, "num_tokens": 444870092.0, "step": 11715 }, { "epoch": 1.4903956239664164, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.186523914337158, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8707098960876465, "num_tokens": 444909267.0, "step": 11716 }, { "epoch": 1.490522834245007, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000247955322265625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3698408603668213, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8682714700698853, "num_tokens": 444946851.0, "step": 11717 }, { "epoch": 1.4906500445235975, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.222200870513916, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8876883387565613, "num_tokens": 444986324.0, "step": 11718 }, { "epoch": 1.490777254802188, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6061699390411377, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8800137639045715, "num_tokens": 445020443.0, "step": 11719 }, { "epoch": 1.4909044650807786, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.058213710784912, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8558584451675415, "num_tokens": 445057592.0, "step": 11720 }, { "epoch": 1.491031675359369, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6663639545440674, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8855771422386169, "num_tokens": 445091513.0, "step": 11721 }, { "epoch": 1.4911588856379596, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8440325260162354, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8581995368003845, "num_tokens": 445132579.0, "step": 11722 }, { "epoch": 1.49128609591655, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6274986267089844, "learning_rate": 1e-05, "loss": 0.5192, "mean_token_accuracy": 0.8504805564880371, "num_tokens": 445167336.0, "step": 11723 }, { "epoch": 1.4914133061951405, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.662588119506836, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8623934388160706, "num_tokens": 445208873.0, "step": 11724 }, { "epoch": 1.491540516473731, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5471160411834717, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8695911765098572, "num_tokens": 445245649.0, "step": 11725 }, { "epoch": 1.4916677267523215, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5313074588775635, "learning_rate": 1e-05, "loss": 0.4038, "mean_token_accuracy": 0.8854982852935791, "num_tokens": 445285890.0, "step": 11726 }, { "epoch": 1.491794937030912, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4818053245544434, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8632702827453613, "num_tokens": 445323408.0, "step": 11727 }, { "epoch": 1.4919221473095026, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4611387252807617, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8805956244468689, "num_tokens": 445361086.0, "step": 11728 }, { "epoch": 1.492049357588093, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.410429000854492, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8747289776802063, "num_tokens": 445404726.0, "step": 11729 }, { "epoch": 1.4921765678666836, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4394237995147705, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8728849291801453, "num_tokens": 445444190.0, "step": 11730 }, { "epoch": 1.4923037781452742, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3577330112457275, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8632636070251465, "num_tokens": 445484395.0, "step": 11731 }, { "epoch": 1.4924309884238647, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.449019432067871, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8680583238601685, "num_tokens": 445518104.0, "step": 11732 }, { "epoch": 1.4925581987024552, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.307459592819214, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8631436824798584, "num_tokens": 445554761.0, "step": 11733 }, { "epoch": 1.4926854089810457, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.462608814239502, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8666746020317078, "num_tokens": 445595331.0, "step": 11734 }, { "epoch": 1.4928126192596363, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.235661506652832, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8772728443145752, "num_tokens": 445639280.0, "step": 11735 }, { "epoch": 1.4929398295382268, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5480575561523438, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8696836233139038, "num_tokens": 445675437.0, "step": 11736 }, { "epoch": 1.4930670398168173, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.173091411590576, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.876701831817627, "num_tokens": 445712419.0, "step": 11737 }, { "epoch": 1.4931942500954076, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6716811656951904, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8537994027137756, "num_tokens": 445751465.0, "step": 11738 }, { "epoch": 1.4933214603739982, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.0952537059783936, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8831754922866821, "num_tokens": 445788145.0, "step": 11739 }, { "epoch": 1.4934486706525887, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.871584177017212, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8772375583648682, "num_tokens": 445825954.0, "step": 11740 }, { "epoch": 1.4935758809311792, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0734784603118896, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8578026294708252, "num_tokens": 445866207.0, "step": 11741 }, { "epoch": 1.4937030912097697, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.149754524230957, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8608905076980591, "num_tokens": 445904679.0, "step": 11742 }, { "epoch": 1.4938303014883603, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.242464780807495, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8579113483428955, "num_tokens": 445936001.0, "step": 11743 }, { "epoch": 1.4939575117669508, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.323140621185303, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8650072813034058, "num_tokens": 445969212.0, "step": 11744 }, { "epoch": 1.4940847220455413, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.540813446044922, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8690264821052551, "num_tokens": 446013455.0, "step": 11745 }, { "epoch": 1.4942119323241319, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.260926723480225, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.867039680480957, "num_tokens": 446044152.0, "step": 11746 }, { "epoch": 1.4943391426027224, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7396912574768066, "learning_rate": 1e-05, "loss": 0.5119, "mean_token_accuracy": 0.8570284843444824, "num_tokens": 446087047.0, "step": 11747 }, { "epoch": 1.4944663528813127, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.025862216949463, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8782014846801758, "num_tokens": 446120095.0, "step": 11748 }, { "epoch": 1.4945935631599032, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.710094451904297, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8766601085662842, "num_tokens": 446156827.0, "step": 11749 }, { "epoch": 1.4947207734384937, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8380985260009766, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.869827926158905, "num_tokens": 446197101.0, "step": 11750 }, { "epoch": 1.4948479837170843, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.613922595977783, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8687136173248291, "num_tokens": 446230442.0, "step": 11751 }, { "epoch": 1.4949751939956748, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6719882488250732, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8807854652404785, "num_tokens": 446268395.0, "step": 11752 }, { "epoch": 1.4951024042742653, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5050203800201416, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8751075267791748, "num_tokens": 446307855.0, "step": 11753 }, { "epoch": 1.4952296145528559, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5908873081207275, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8702747225761414, "num_tokens": 446348808.0, "step": 11754 }, { "epoch": 1.4953568248314464, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4200735092163086, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.877430260181427, "num_tokens": 446380580.0, "step": 11755 }, { "epoch": 1.495484035110037, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.570497989654541, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8772396445274353, "num_tokens": 446414420.0, "step": 11756 }, { "epoch": 1.4956112453886274, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3157143592834473, "learning_rate": 1e-05, "loss": 0.389, "mean_token_accuracy": 0.888415515422821, "num_tokens": 446450220.0, "step": 11757 }, { "epoch": 1.495738455667218, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.610809803009033, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.88512122631073, "num_tokens": 446483823.0, "step": 11758 }, { "epoch": 1.4958656659458085, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2401466369628906, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8727813959121704, "num_tokens": 446522317.0, "step": 11759 }, { "epoch": 1.495992876224399, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.70563006401062, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8647201061248779, "num_tokens": 446564491.0, "step": 11760 }, { "epoch": 1.4961200865029896, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1674561500549316, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8739122152328491, "num_tokens": 446605490.0, "step": 11761 }, { "epoch": 1.49624729678158, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.875798225402832, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8606290817260742, "num_tokens": 446647385.0, "step": 11762 }, { "epoch": 1.4963745070601704, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.169781446456909, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8668146133422852, "num_tokens": 446682895.0, "step": 11763 }, { "epoch": 1.496501717338761, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.06826639175415, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8614176511764526, "num_tokens": 446724724.0, "step": 11764 }, { "epoch": 1.4966289276173514, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2957072257995605, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8717129230499268, "num_tokens": 446763433.0, "step": 11765 }, { "epoch": 1.496756137895942, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.156418323516846, "learning_rate": 1e-05, "loss": 0.5342, "mean_token_accuracy": 0.8459311127662659, "num_tokens": 446809275.0, "step": 11766 }, { "epoch": 1.4968833481745325, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5157835483551025, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.8839352130889893, "num_tokens": 446840619.0, "step": 11767 }, { "epoch": 1.497010558453123, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.037911415100098, "learning_rate": 1e-05, "loss": 0.4966, "mean_token_accuracy": 0.8581234216690063, "num_tokens": 446879979.0, "step": 11768 }, { "epoch": 1.4971377687317136, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.623401403427124, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8817043304443359, "num_tokens": 446913729.0, "step": 11769 }, { "epoch": 1.497264979010304, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8239564895629883, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8576004505157471, "num_tokens": 446951843.0, "step": 11770 }, { "epoch": 1.4973921892888946, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6236612796783447, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8671942353248596, "num_tokens": 446988091.0, "step": 11771 }, { "epoch": 1.497519399567485, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6155242919921875, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8617939949035645, "num_tokens": 447023833.0, "step": 11772 }, { "epoch": 1.4976466098460754, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.575577735900879, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8768231868743896, "num_tokens": 447056232.0, "step": 11773 }, { "epoch": 1.497773820124666, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4573612213134766, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.881233811378479, "num_tokens": 447094524.0, "step": 11774 }, { "epoch": 1.4979010304032565, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5496132373809814, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8660237193107605, "num_tokens": 447129076.0, "step": 11775 }, { "epoch": 1.498028240681847, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.353163003921509, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8699511289596558, "num_tokens": 447168651.0, "step": 11776 }, { "epoch": 1.4981554509604376, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.544278860092163, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8705571293830872, "num_tokens": 447207491.0, "step": 11777 }, { "epoch": 1.498282661239028, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.278348207473755, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8782180547714233, "num_tokens": 447242955.0, "step": 11778 }, { "epoch": 1.4984098715176186, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5976884365081787, "learning_rate": 1e-05, "loss": 0.3941, "mean_token_accuracy": 0.8900112509727478, "num_tokens": 447277634.0, "step": 11779 }, { "epoch": 1.4985370817962091, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2041170597076416, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8657821416854858, "num_tokens": 447315255.0, "step": 11780 }, { "epoch": 1.4986642920747997, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6695666313171387, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8827735781669617, "num_tokens": 447353801.0, "step": 11781 }, { "epoch": 1.4987915023533902, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.145895481109619, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8671027421951294, "num_tokens": 447388098.0, "step": 11782 }, { "epoch": 1.4989187126319807, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8202662467956543, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8702181577682495, "num_tokens": 447424860.0, "step": 11783 }, { "epoch": 1.4990459229105713, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.125101327896118, "learning_rate": 1e-05, "loss": 0.5067, "mean_token_accuracy": 0.8522698283195496, "num_tokens": 447465179.0, "step": 11784 }, { "epoch": 1.4991731331891618, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.010898113250732, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8585269451141357, "num_tokens": 447505960.0, "step": 11785 }, { "epoch": 1.4993003434677523, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.195866823196411, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8625952005386353, "num_tokens": 447545410.0, "step": 11786 }, { "epoch": 1.4994275537463426, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.167491912841797, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8667494058609009, "num_tokens": 447586294.0, "step": 11787 }, { "epoch": 1.4995547640249332, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.397883176803589, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8733056783676147, "num_tokens": 447628841.0, "step": 11788 }, { "epoch": 1.4996819743035237, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.173067092895508, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8704429268836975, "num_tokens": 447669456.0, "step": 11789 }, { "epoch": 1.4998091845821142, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.598388433456421, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8781103491783142, "num_tokens": 447703297.0, "step": 11790 }, { "epoch": 1.4999363948607047, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.014503002166748, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.8561021089553833, "num_tokens": 447742497.0, "step": 11791 }, { "epoch": 1.5000636051392953, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6366021633148193, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8718347549438477, "num_tokens": 447783956.0, "step": 11792 }, { "epoch": 1.5001908154178858, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8042640686035156, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8698394894599915, "num_tokens": 447824527.0, "step": 11793 }, { "epoch": 1.5003180256964763, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5997049808502197, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8673532009124756, "num_tokens": 447863676.0, "step": 11794 }, { "epoch": 1.5004452359750666, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.63830828666687, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8575102686882019, "num_tokens": 447906157.0, "step": 11795 }, { "epoch": 1.5005724462536572, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.515716075897217, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8844130039215088, "num_tokens": 447944016.0, "step": 11796 }, { "epoch": 1.5006996565322477, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5329105854034424, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8800994753837585, "num_tokens": 447979500.0, "step": 11797 }, { "epoch": 1.5008268668108382, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4672300815582275, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8576650619506836, "num_tokens": 448012770.0, "step": 11798 }, { "epoch": 1.5009540770894287, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4761478900909424, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8722879886627197, "num_tokens": 448048227.0, "step": 11799 }, { "epoch": 1.5010812873680193, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.405444383621216, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8613033294677734, "num_tokens": 448081672.0, "step": 11800 }, { "epoch": 1.5012084976466098, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.446140766143799, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8690133094787598, "num_tokens": 448122693.0, "step": 11801 }, { "epoch": 1.5013357079252003, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3463082313537598, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.8603890538215637, "num_tokens": 448161246.0, "step": 11802 }, { "epoch": 1.5014629182037909, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4477996826171875, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8679955005645752, "num_tokens": 448205233.0, "step": 11803 }, { "epoch": 1.5015901284823814, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3027093410491943, "learning_rate": 1e-05, "loss": 0.5146, "mean_token_accuracy": 0.8478525876998901, "num_tokens": 448247404.0, "step": 11804 }, { "epoch": 1.501717338760972, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4795215129852295, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8735373020172119, "num_tokens": 448286874.0, "step": 11805 }, { "epoch": 1.5018445490395624, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.249467611312866, "learning_rate": 1e-05, "loss": 0.512, "mean_token_accuracy": 0.8509276509284973, "num_tokens": 448325461.0, "step": 11806 }, { "epoch": 1.501971759318153, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5267903804779053, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8774658441543579, "num_tokens": 448366871.0, "step": 11807 }, { "epoch": 1.5020989695967435, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1998543739318848, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8768730163574219, "num_tokens": 448396798.0, "step": 11808 }, { "epoch": 1.502226179875334, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6268320083618164, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8724797964096069, "num_tokens": 448437782.0, "step": 11809 }, { "epoch": 1.5023533901539246, "ewc_loss": 0.06884765625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1236770153045654, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8732505440711975, "num_tokens": 448469892.0, "step": 11810 }, { "epoch": 1.502480600432515, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7878899574279785, "learning_rate": 1e-05, "loss": 0.4726, "mean_token_accuracy": 0.8662105202674866, "num_tokens": 448508233.0, "step": 11811 }, { "epoch": 1.5026078107111056, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.0821659564971924, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8587424159049988, "num_tokens": 448541071.0, "step": 11812 }, { "epoch": 1.502735020989696, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.034965991973877, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8624746203422546, "num_tokens": 448571348.0, "step": 11813 }, { "epoch": 1.5028622312682864, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.146813154220581, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8703827857971191, "num_tokens": 448611889.0, "step": 11814 }, { "epoch": 1.502989441546877, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.260293960571289, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8764562010765076, "num_tokens": 448646683.0, "step": 11815 }, { "epoch": 1.5031166518254675, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.409710645675659, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8723976612091064, "num_tokens": 448688154.0, "step": 11816 }, { "epoch": 1.503243862104058, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.297394752502441, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8668965697288513, "num_tokens": 448726343.0, "step": 11817 }, { "epoch": 1.5033710723826486, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.673351287841797, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8875740766525269, "num_tokens": 448770376.0, "step": 11818 }, { "epoch": 1.503498282661239, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.111009120941162, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8801164627075195, "num_tokens": 448799247.0, "step": 11819 }, { "epoch": 1.5036254929398294, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7287347316741943, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.8609054684638977, "num_tokens": 448839379.0, "step": 11820 }, { "epoch": 1.50375270321842, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.909960985183716, "learning_rate": 1e-05, "loss": 0.5213, "mean_token_accuracy": 0.8517037630081177, "num_tokens": 448873924.0, "step": 11821 }, { "epoch": 1.5038799134970104, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6341209411621094, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8674833178520203, "num_tokens": 448915765.0, "step": 11822 }, { "epoch": 1.504007123775601, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7360494136810303, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8772896528244019, "num_tokens": 448953116.0, "step": 11823 }, { "epoch": 1.5041343340541915, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5525004863739014, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.864788293838501, "num_tokens": 448989065.0, "step": 11824 }, { "epoch": 1.504261544332782, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6262564659118652, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8644244074821472, "num_tokens": 449029608.0, "step": 11825 }, { "epoch": 1.5043887546113726, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.461207151412964, "learning_rate": 1e-05, "loss": 0.5034, "mean_token_accuracy": 0.8608678579330444, "num_tokens": 449068288.0, "step": 11826 }, { "epoch": 1.504515964889963, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.564016342163086, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8679516911506653, "num_tokens": 449107754.0, "step": 11827 }, { "epoch": 1.5046431751685536, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.353965997695923, "learning_rate": 1e-05, "loss": 0.3758, "mean_token_accuracy": 0.8932532668113708, "num_tokens": 449148680.0, "step": 11828 }, { "epoch": 1.5047703854471441, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.582174301147461, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8682886362075806, "num_tokens": 449191616.0, "step": 11829 }, { "epoch": 1.5048975957257347, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2751104831695557, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8797135949134827, "num_tokens": 449227149.0, "step": 11830 }, { "epoch": 1.5050248060043252, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6682419776916504, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8679693937301636, "num_tokens": 449259186.0, "step": 11831 }, { "epoch": 1.5051520162829157, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2059826850891113, "learning_rate": 1e-05, "loss": 0.4167, "mean_token_accuracy": 0.8801480531692505, "num_tokens": 449294403.0, "step": 11832 }, { "epoch": 1.5052792265615063, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.773986339569092, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8891075849533081, "num_tokens": 449334205.0, "step": 11833 }, { "epoch": 1.5054064368400968, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002498626708984375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1661953926086426, "learning_rate": 1e-05, "loss": 0.5197, "mean_token_accuracy": 0.8529275059700012, "num_tokens": 449372501.0, "step": 11834 }, { "epoch": 1.5055336471186873, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9622802734375, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8594226837158203, "num_tokens": 449410394.0, "step": 11835 }, { "epoch": 1.5056608573972778, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2093076705932617, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.873293936252594, "num_tokens": 449449494.0, "step": 11836 }, { "epoch": 1.5057880676758684, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.115893363952637, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8663702607154846, "num_tokens": 449485086.0, "step": 11837 }, { "epoch": 1.5059152779544587, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.4055213928222656, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8696063756942749, "num_tokens": 449518681.0, "step": 11838 }, { "epoch": 1.5060424882330492, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.080589771270752, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8761575222015381, "num_tokens": 449557439.0, "step": 11839 }, { "epoch": 1.5061696985116397, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5466268062591553, "learning_rate": 1e-05, "loss": 0.5071, "mean_token_accuracy": 0.8548265695571899, "num_tokens": 449600571.0, "step": 11840 }, { "epoch": 1.5062969087902303, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9199540615081787, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8592279553413391, "num_tokens": 449639323.0, "step": 11841 }, { "epoch": 1.5064241190688208, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.586484670639038, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.867203950881958, "num_tokens": 449680082.0, "step": 11842 }, { "epoch": 1.5065513293474113, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7223961353302, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8565695285797119, "num_tokens": 449717655.0, "step": 11843 }, { "epoch": 1.5066785396260016, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5729944705963135, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.881511390209198, "num_tokens": 449756833.0, "step": 11844 }, { "epoch": 1.5068057499045922, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.530782699584961, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8712457418441772, "num_tokens": 449796525.0, "step": 11845 }, { "epoch": 1.5069329601831827, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5366976261138916, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8684555292129517, "num_tokens": 449837731.0, "step": 11846 }, { "epoch": 1.5070601704617732, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4040451049804688, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8767900466918945, "num_tokens": 449876235.0, "step": 11847 }, { "epoch": 1.5071873807403637, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5271589756011963, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8708903789520264, "num_tokens": 449916153.0, "step": 11848 }, { "epoch": 1.5073145910189543, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.320802927017212, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8559006452560425, "num_tokens": 449957969.0, "step": 11849 }, { "epoch": 1.5074418012975448, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5350229740142822, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8734095692634583, "num_tokens": 449997464.0, "step": 11850 }, { "epoch": 1.5075690115761353, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2551517486572266, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8612550497055054, "num_tokens": 450038102.0, "step": 11851 }, { "epoch": 1.5076962218547258, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5796573162078857, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8784060478210449, "num_tokens": 450073832.0, "step": 11852 }, { "epoch": 1.5078234321333164, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1902573108673096, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8797948956489563, "num_tokens": 450108662.0, "step": 11853 }, { "epoch": 1.507950642411907, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6757469177246094, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8859367966651917, "num_tokens": 450144601.0, "step": 11854 }, { "epoch": 1.5080778526904974, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.122927665710449, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8649015426635742, "num_tokens": 450190112.0, "step": 11855 }, { "epoch": 1.508205062969088, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8415257930755615, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8816877603530884, "num_tokens": 450227660.0, "step": 11856 }, { "epoch": 1.5083322732476785, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.101907730102539, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.877743124961853, "num_tokens": 450265243.0, "step": 11857 }, { "epoch": 1.508459483526269, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.071328639984131, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8639634847640991, "num_tokens": 450303382.0, "step": 11858 }, { "epoch": 1.5085866938048595, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.229640007019043, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8731077909469604, "num_tokens": 450335458.0, "step": 11859 }, { "epoch": 1.50871390408345, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.223177909851074, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8681236505508423, "num_tokens": 450373375.0, "step": 11860 }, { "epoch": 1.5088411143620406, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.4634785652160645, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.880182147026062, "num_tokens": 450406258.0, "step": 11861 }, { "epoch": 1.508968324640631, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1712446212768555, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.876987099647522, "num_tokens": 450441558.0, "step": 11862 }, { "epoch": 1.5090955349192214, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.623112440109253, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8757523894309998, "num_tokens": 450477685.0, "step": 11863 }, { "epoch": 1.509222745197812, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9916391372680664, "learning_rate": 1e-05, "loss": 0.5034, "mean_token_accuracy": 0.8551962375640869, "num_tokens": 450516661.0, "step": 11864 }, { "epoch": 1.5093499554764025, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.637101888656616, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8714089393615723, "num_tokens": 450555091.0, "step": 11865 }, { "epoch": 1.509477165754993, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7944376468658447, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.868902325630188, "num_tokens": 450596172.0, "step": 11866 }, { "epoch": 1.5096043760335836, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5806639194488525, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8655798435211182, "num_tokens": 450636902.0, "step": 11867 }, { "epoch": 1.509731586312174, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6460628509521484, "learning_rate": 1e-05, "loss": 0.5244, "mean_token_accuracy": 0.8563341498374939, "num_tokens": 450682826.0, "step": 11868 }, { "epoch": 1.5098587965907644, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.516517162322998, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8605509996414185, "num_tokens": 450722995.0, "step": 11869 }, { "epoch": 1.509986006869355, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5381147861480713, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8770418167114258, "num_tokens": 450762109.0, "step": 11870 }, { "epoch": 1.5101132171479454, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.441462993621826, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8730847835540771, "num_tokens": 450799543.0, "step": 11871 }, { "epoch": 1.510240427426536, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4900481700897217, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.867557168006897, "num_tokens": 450840424.0, "step": 11872 }, { "epoch": 1.5103676377051265, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3701157569885254, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8759311437606812, "num_tokens": 450877652.0, "step": 11873 }, { "epoch": 1.510494847983717, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4902331829071045, "learning_rate": 1e-05, "loss": 0.5087, "mean_token_accuracy": 0.8508948683738708, "num_tokens": 450916171.0, "step": 11874 }, { "epoch": 1.5106220582623076, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3144452571868896, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.875678300857544, "num_tokens": 450950455.0, "step": 11875 }, { "epoch": 1.510749268540898, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5122451782226562, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8693311214447021, "num_tokens": 450990747.0, "step": 11876 }, { "epoch": 1.5108764788194886, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2541677951812744, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8752540946006775, "num_tokens": 451029396.0, "step": 11877 }, { "epoch": 1.5110036890980791, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5875284671783447, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8692549467086792, "num_tokens": 451065071.0, "step": 11878 }, { "epoch": 1.5111308993766697, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.186387777328491, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.874207079410553, "num_tokens": 451105487.0, "step": 11879 }, { "epoch": 1.5112581096552602, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7166590690612793, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.860881507396698, "num_tokens": 451141538.0, "step": 11880 }, { "epoch": 1.5113853199338507, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.115341901779175, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.884221613407135, "num_tokens": 451175080.0, "step": 11881 }, { "epoch": 1.5115125302124413, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8936221599578857, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.869349479675293, "num_tokens": 451216117.0, "step": 11882 }, { "epoch": 1.5116397404910318, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1204116344451904, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8743634223937988, "num_tokens": 451251585.0, "step": 11883 }, { "epoch": 1.5117669507696223, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.124222755432129, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8586449027061462, "num_tokens": 451290401.0, "step": 11884 }, { "epoch": 1.5118941610482128, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2821476459503174, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8625090718269348, "num_tokens": 451331896.0, "step": 11885 }, { "epoch": 1.5120213713268034, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2545247077941895, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8739246129989624, "num_tokens": 451368802.0, "step": 11886 }, { "epoch": 1.5121485816053937, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5777971744537354, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8832619190216064, "num_tokens": 451400101.0, "step": 11887 }, { "epoch": 1.5122757918839842, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.096072196960449, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8695011138916016, "num_tokens": 451444732.0, "step": 11888 }, { "epoch": 1.5124030021625747, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6526176929473877, "learning_rate": 1e-05, "loss": 0.3977, "mean_token_accuracy": 0.8916296362876892, "num_tokens": 451481467.0, "step": 11889 }, { "epoch": 1.5125302124411653, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.903555154800415, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8702516555786133, "num_tokens": 451520048.0, "step": 11890 }, { "epoch": 1.5126574227197558, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.641827344894409, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.871925950050354, "num_tokens": 451555121.0, "step": 11891 }, { "epoch": 1.5127846329983463, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7094459533691406, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8682117462158203, "num_tokens": 451594894.0, "step": 11892 }, { "epoch": 1.5129118432769366, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5770535469055176, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8760907649993896, "num_tokens": 451627175.0, "step": 11893 }, { "epoch": 1.5130390535555271, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5606346130371094, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.877178430557251, "num_tokens": 451664177.0, "step": 11894 }, { "epoch": 1.5131662638341177, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.510430335998535, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8777574896812439, "num_tokens": 451696964.0, "step": 11895 }, { "epoch": 1.5132934741127082, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4737632274627686, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.885740339756012, "num_tokens": 451735349.0, "step": 11896 }, { "epoch": 1.5134206843912987, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4450831413269043, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8805450201034546, "num_tokens": 451770410.0, "step": 11897 }, { "epoch": 1.5135478946698893, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4158010482788086, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8756747245788574, "num_tokens": 451812170.0, "step": 11898 }, { "epoch": 1.5136751049484798, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3875844478607178, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8723980188369751, "num_tokens": 451858079.0, "step": 11899 }, { "epoch": 1.5138023152270703, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4033567905426025, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8630902767181396, "num_tokens": 451901120.0, "step": 11900 }, { "epoch": 1.5139295255056608, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.356661558151245, "learning_rate": 1e-05, "loss": 0.5045, "mean_token_accuracy": 0.8571759462356567, "num_tokens": 451944990.0, "step": 11901 }, { "epoch": 1.5140567357842514, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3993942737579346, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8784230351448059, "num_tokens": 451983261.0, "step": 11902 }, { "epoch": 1.514183946062842, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3287062644958496, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8665874004364014, "num_tokens": 452019438.0, "step": 11903 }, { "epoch": 1.5143111563414324, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4066977500915527, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8557794094085693, "num_tokens": 452055941.0, "step": 11904 }, { "epoch": 1.514438366620023, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.292097330093384, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8804285526275635, "num_tokens": 452092625.0, "step": 11905 }, { "epoch": 1.5145655768986135, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.433725357055664, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.866402268409729, "num_tokens": 452126279.0, "step": 11906 }, { "epoch": 1.514692787177204, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.245887279510498, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8759223818778992, "num_tokens": 452163506.0, "step": 11907 }, { "epoch": 1.5148199974557945, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.512571096420288, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8614935278892517, "num_tokens": 452200898.0, "step": 11908 }, { "epoch": 1.514947207734385, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.176150321960449, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.859290361404419, "num_tokens": 452237772.0, "step": 11909 }, { "epoch": 1.5150744180129756, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.644423246383667, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8686501979827881, "num_tokens": 452272741.0, "step": 11910 }, { "epoch": 1.515201628291566, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.093158006668091, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8676165342330933, "num_tokens": 452307674.0, "step": 11911 }, { "epoch": 1.5153288385701564, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8342790603637695, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8610718250274658, "num_tokens": 452352186.0, "step": 11912 }, { "epoch": 1.515456048848747, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0341458320617676, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8777161836624146, "num_tokens": 452390371.0, "step": 11913 }, { "epoch": 1.5155832591273375, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.145688533782959, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8828122615814209, "num_tokens": 452424545.0, "step": 11914 }, { "epoch": 1.515710469405928, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.170530080795288, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8661092519760132, "num_tokens": 452456543.0, "step": 11915 }, { "epoch": 1.5158376796845185, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.437834739685059, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8599454164505005, "num_tokens": 452497884.0, "step": 11916 }, { "epoch": 1.515964889963109, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.54992938041687, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8637614846229553, "num_tokens": 452542596.0, "step": 11917 }, { "epoch": 1.5160921002416994, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.388688087463379, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8792499303817749, "num_tokens": 452581385.0, "step": 11918 }, { "epoch": 1.51621931052029, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.830109119415283, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8678906559944153, "num_tokens": 452621399.0, "step": 11919 }, { "epoch": 1.5163465207988804, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.138063430786133, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8645241856575012, "num_tokens": 452657782.0, "step": 11920 }, { "epoch": 1.516473731077471, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00025177001953125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.8081631660461426, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8811721801757812, "num_tokens": 452696239.0, "step": 11921 }, { "epoch": 1.5166009413560615, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9233076572418213, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.862972617149353, "num_tokens": 452734235.0, "step": 11922 }, { "epoch": 1.516728151634652, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6696205139160156, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8768298029899597, "num_tokens": 452769151.0, "step": 11923 }, { "epoch": 1.5168553619132426, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7687060832977295, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8757976293563843, "num_tokens": 452808944.0, "step": 11924 }, { "epoch": 1.516982572191833, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.548152446746826, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8690183758735657, "num_tokens": 452845815.0, "step": 11925 }, { "epoch": 1.5171097824704236, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6903507709503174, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8721197247505188, "num_tokens": 452888010.0, "step": 11926 }, { "epoch": 1.5172369927490141, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4230875968933105, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.880254864692688, "num_tokens": 452924149.0, "step": 11927 }, { "epoch": 1.5173642030276047, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6822476387023926, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8759777545928955, "num_tokens": 452964436.0, "step": 11928 }, { "epoch": 1.5174914133061952, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3232808113098145, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8639260530471802, "num_tokens": 453000455.0, "step": 11929 }, { "epoch": 1.5176186235847857, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.728158473968506, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.884925365447998, "num_tokens": 453040419.0, "step": 11930 }, { "epoch": 1.5177458338633762, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2387309074401855, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8692171573638916, "num_tokens": 453077668.0, "step": 11931 }, { "epoch": 1.5178730441419668, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.86445689201355, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.87721848487854, "num_tokens": 453115423.0, "step": 11932 }, { "epoch": 1.5180002544205573, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2139785289764404, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8687254190444946, "num_tokens": 453150529.0, "step": 11933 }, { "epoch": 1.5181274646991478, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.037716388702393, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8710638284683228, "num_tokens": 453189123.0, "step": 11934 }, { "epoch": 1.5182546749777384, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.300182580947876, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8835703134536743, "num_tokens": 453231381.0, "step": 11935 }, { "epoch": 1.5183818852563287, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.146231174468994, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8576052784919739, "num_tokens": 453272134.0, "step": 11936 }, { "epoch": 1.5185090955349192, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.518432378768921, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.86580491065979, "num_tokens": 453306950.0, "step": 11937 }, { "epoch": 1.5186363058135097, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.032851696014404, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.8590850830078125, "num_tokens": 453354206.0, "step": 11938 }, { "epoch": 1.5187635160921003, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6444649696350098, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8667941093444824, "num_tokens": 453388033.0, "step": 11939 }, { "epoch": 1.5188907263706908, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8176581859588623, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8650494813919067, "num_tokens": 453421636.0, "step": 11940 }, { "epoch": 1.5190179366492813, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6201565265655518, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8731006383895874, "num_tokens": 453467994.0, "step": 11941 }, { "epoch": 1.5191451469278716, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.594388961791992, "learning_rate": 1e-05, "loss": 0.492, "mean_token_accuracy": 0.8567184209823608, "num_tokens": 453508627.0, "step": 11942 }, { "epoch": 1.5192723572064621, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5967140197753906, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8717250823974609, "num_tokens": 453547286.0, "step": 11943 }, { "epoch": 1.5193995674850527, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.428165912628174, "learning_rate": 1e-05, "loss": 0.5156, "mean_token_accuracy": 0.8479911684989929, "num_tokens": 453588485.0, "step": 11944 }, { "epoch": 1.5195267777636432, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5767695903778076, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.866897463798523, "num_tokens": 453633103.0, "step": 11945 }, { "epoch": 1.5196539880422337, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.318974018096924, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.874887228012085, "num_tokens": 453671817.0, "step": 11946 }, { "epoch": 1.5197811983208243, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.597095012664795, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8711657524108887, "num_tokens": 453704704.0, "step": 11947 }, { "epoch": 1.5199084085994148, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2356340885162354, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8646115064620972, "num_tokens": 453743899.0, "step": 11948 }, { "epoch": 1.5200356188780053, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.66041898727417, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8759431838989258, "num_tokens": 453783052.0, "step": 11949 }, { "epoch": 1.5201628291565958, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.162651300430298, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8630654811859131, "num_tokens": 453825522.0, "step": 11950 }, { "epoch": 1.5202900394351864, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7943575382232666, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8763406276702881, "num_tokens": 453859353.0, "step": 11951 }, { "epoch": 1.520417249713777, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1203808784484863, "learning_rate": 1e-05, "loss": 0.3827, "mean_token_accuracy": 0.8880404829978943, "num_tokens": 453896792.0, "step": 11952 }, { "epoch": 1.5205444599923674, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9858169555664062, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.872816264629364, "num_tokens": 453939679.0, "step": 11953 }, { "epoch": 1.520671670270958, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2014567852020264, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.8554717898368835, "num_tokens": 453970171.0, "step": 11954 }, { "epoch": 1.5207988805495485, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1473069190979, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.860588550567627, "num_tokens": 454008470.0, "step": 11955 }, { "epoch": 1.520926090828139, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.379760265350342, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8618854284286499, "num_tokens": 454047711.0, "step": 11956 }, { "epoch": 1.5210533011067295, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.161535739898682, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8674378395080566, "num_tokens": 454083494.0, "step": 11957 }, { "epoch": 1.52118051138532, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.579014539718628, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.874488353729248, "num_tokens": 454115220.0, "step": 11958 }, { "epoch": 1.5213077216639106, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9906797409057617, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8753793239593506, "num_tokens": 454152754.0, "step": 11959 }, { "epoch": 1.521434931942501, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6267271041870117, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8716339468955994, "num_tokens": 454185357.0, "step": 11960 }, { "epoch": 1.5215621422210914, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7986929416656494, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.879308819770813, "num_tokens": 454217277.0, "step": 11961 }, { "epoch": 1.521689352499682, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5910191535949707, "learning_rate": 1e-05, "loss": 0.5088, "mean_token_accuracy": 0.8538604378700256, "num_tokens": 454253252.0, "step": 11962 }, { "epoch": 1.5218165627782725, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6304845809936523, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8794889450073242, "num_tokens": 454292779.0, "step": 11963 }, { "epoch": 1.521943773056863, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.522040605545044, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8706284761428833, "num_tokens": 454336077.0, "step": 11964 }, { "epoch": 1.5220709833354535, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.515799045562744, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.881883978843689, "num_tokens": 454369088.0, "step": 11965 }, { "epoch": 1.5221981936140438, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4537079334259033, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8612685203552246, "num_tokens": 454413057.0, "step": 11966 }, { "epoch": 1.5223254038926344, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.46286678314209, "learning_rate": 1e-05, "loss": 0.4083, "mean_token_accuracy": 0.8837560415267944, "num_tokens": 454447458.0, "step": 11967 }, { "epoch": 1.522452614171225, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.41163969039917, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8825274109840393, "num_tokens": 454479057.0, "step": 11968 }, { "epoch": 1.5225798244498154, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4438722133636475, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8589040637016296, "num_tokens": 454520676.0, "step": 11969 }, { "epoch": 1.522707034728406, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3907485008239746, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.864429235458374, "num_tokens": 454558021.0, "step": 11970 }, { "epoch": 1.5228342450069965, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.413999557495117, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8767473697662354, "num_tokens": 454594701.0, "step": 11971 }, { "epoch": 1.522961455285587, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3336691856384277, "learning_rate": 1e-05, "loss": 0.5176, "mean_token_accuracy": 0.846653938293457, "num_tokens": 454638746.0, "step": 11972 }, { "epoch": 1.5230886655641775, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4248862266540527, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8741977214813232, "num_tokens": 454676040.0, "step": 11973 }, { "epoch": 1.523215875842768, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.300448417663574, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8685868382453918, "num_tokens": 454713262.0, "step": 11974 }, { "epoch": 1.5233430861213586, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4645259380340576, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8658239245414734, "num_tokens": 454748613.0, "step": 11975 }, { "epoch": 1.5234702963999491, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2387278079986572, "learning_rate": 1e-05, "loss": 0.3888, "mean_token_accuracy": 0.8888784646987915, "num_tokens": 454782040.0, "step": 11976 }, { "epoch": 1.5235975066785397, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.529582977294922, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8600413203239441, "num_tokens": 454828394.0, "step": 11977 }, { "epoch": 1.5237247169571302, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.18139386177063, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8616407513618469, "num_tokens": 454866324.0, "step": 11978 }, { "epoch": 1.5238519272357207, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6421921253204346, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8748271465301514, "num_tokens": 454907458.0, "step": 11979 }, { "epoch": 1.5239791375143112, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.095106840133667, "learning_rate": 1e-05, "loss": 0.3968, "mean_token_accuracy": 0.8854968547821045, "num_tokens": 454948159.0, "step": 11980 }, { "epoch": 1.5241063477929018, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8504652976989746, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8779109716415405, "num_tokens": 454984760.0, "step": 11981 }, { "epoch": 1.5242335580714923, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0515904426574707, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.870267927646637, "num_tokens": 455026670.0, "step": 11982 }, { "epoch": 1.5243607683500828, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.145743370056152, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8689455986022949, "num_tokens": 455063470.0, "step": 11983 }, { "epoch": 1.5244879786286734, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.187401056289673, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8836085200309753, "num_tokens": 455097787.0, "step": 11984 }, { "epoch": 1.5246151889072637, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.380601406097412, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8846572041511536, "num_tokens": 455137987.0, "step": 11985 }, { "epoch": 1.5247423991858542, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.5467703342437744, "learning_rate": 1e-05, "loss": 0.5039, "mean_token_accuracy": 0.8549978733062744, "num_tokens": 455184048.0, "step": 11986 }, { "epoch": 1.5248696094644447, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.330247402191162, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8738541603088379, "num_tokens": 455219876.0, "step": 11987 }, { "epoch": 1.5249968197430352, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.793743848800659, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8730692863464355, "num_tokens": 455253491.0, "step": 11988 }, { "epoch": 1.5251240300216258, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.086807727813721, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.866500198841095, "num_tokens": 455293183.0, "step": 11989 }, { "epoch": 1.5252512403002163, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.748807191848755, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8861393928527832, "num_tokens": 455333970.0, "step": 11990 }, { "epoch": 1.5253784505788066, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.89583158493042, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8686056137084961, "num_tokens": 455371547.0, "step": 11991 }, { "epoch": 1.5255056608573971, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6477155685424805, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8692503571510315, "num_tokens": 455403427.0, "step": 11992 }, { "epoch": 1.5256328711359877, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.759976625442505, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8936219811439514, "num_tokens": 455434707.0, "step": 11993 }, { "epoch": 1.5257600814145782, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.51459002494812, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8758395910263062, "num_tokens": 455473255.0, "step": 11994 }, { "epoch": 1.5258872916931687, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6969799995422363, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.873083770275116, "num_tokens": 455504631.0, "step": 11995 }, { "epoch": 1.5260145019717593, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3933920860290527, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8822476267814636, "num_tokens": 455542652.0, "step": 11996 }, { "epoch": 1.5261417122503498, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.7067298889160156, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.875312328338623, "num_tokens": 455577912.0, "step": 11997 }, { "epoch": 1.5262689225289403, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3005716800689697, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8768183588981628, "num_tokens": 455617900.0, "step": 11998 }, { "epoch": 1.5263961328075308, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.784946918487549, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8779007196426392, "num_tokens": 455656779.0, "step": 11999 }, { "epoch": 1.5265233430861214, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.230876922607422, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.875930666923523, "num_tokens": 455692620.0, "step": 12000 }, { "epoch": 1.526650553364712, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.926119089126587, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8755314350128174, "num_tokens": 455729805.0, "step": 12001 }, { "epoch": 1.5267777636433024, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.247408151626587, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8764808773994446, "num_tokens": 455767529.0, "step": 12002 }, { "epoch": 1.526904973921893, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.065704345703125, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.870948076248169, "num_tokens": 455807898.0, "step": 12003 }, { "epoch": 1.5270321842004835, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3831474781036377, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8707556128501892, "num_tokens": 455848726.0, "step": 12004 }, { "epoch": 1.527159394479074, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.075877666473389, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.869612455368042, "num_tokens": 455892363.0, "step": 12005 }, { "epoch": 1.5272866047576645, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5501697063446045, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.860387921333313, "num_tokens": 455930104.0, "step": 12006 }, { "epoch": 1.527413815036255, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9274046421051025, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8652414083480835, "num_tokens": 455964272.0, "step": 12007 }, { "epoch": 1.5275410253148456, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.609889507293701, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8675978183746338, "num_tokens": 456003020.0, "step": 12008 }, { "epoch": 1.527668235593436, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.701111316680908, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8711792230606079, "num_tokens": 456040305.0, "step": 12009 }, { "epoch": 1.5277954458720264, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5970067977905273, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8649387955665588, "num_tokens": 456080477.0, "step": 12010 }, { "epoch": 1.527922656150617, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.5001087188720703, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8793366551399231, "num_tokens": 456116941.0, "step": 12011 }, { "epoch": 1.5280498664292075, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5795538425445557, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8659536242485046, "num_tokens": 456156184.0, "step": 12012 }, { "epoch": 1.528177076707798, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3502731323242188, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8577650785446167, "num_tokens": 456200920.0, "step": 12013 }, { "epoch": 1.5283042869863885, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6009411811828613, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8788257837295532, "num_tokens": 456235632.0, "step": 12014 }, { "epoch": 1.5284314972649788, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.259030818939209, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.8724970817565918, "num_tokens": 456264898.0, "step": 12015 }, { "epoch": 1.5285587075435694, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.652947425842285, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8673080205917358, "num_tokens": 456304972.0, "step": 12016 }, { "epoch": 1.52868591782216, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1706812381744385, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8699811697006226, "num_tokens": 456344496.0, "step": 12017 }, { "epoch": 1.5288131281007504, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7788045406341553, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8772799968719482, "num_tokens": 456382408.0, "step": 12018 }, { "epoch": 1.528940338379341, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1279947757720947, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8651066422462463, "num_tokens": 456424946.0, "step": 12019 }, { "epoch": 1.5290675486579315, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9501514434814453, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8807671666145325, "num_tokens": 456467432.0, "step": 12020 }, { "epoch": 1.529194758936522, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.1741690635681152, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8759549856185913, "num_tokens": 456501557.0, "step": 12021 }, { "epoch": 1.5293219692151125, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.132831573486328, "learning_rate": 1e-05, "loss": 0.5515, "mean_token_accuracy": 0.8455650806427002, "num_tokens": 456540500.0, "step": 12022 }, { "epoch": 1.529449179493703, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3511717319488525, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8773038387298584, "num_tokens": 456574403.0, "step": 12023 }, { "epoch": 1.5295763897722936, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.171575546264648, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8764471411705017, "num_tokens": 456610044.0, "step": 12024 }, { "epoch": 1.5297036000508841, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.547305107116699, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8737994432449341, "num_tokens": 456650289.0, "step": 12025 }, { "epoch": 1.5298308103294747, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.041396141052246, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8674120903015137, "num_tokens": 456687193.0, "step": 12026 }, { "epoch": 1.5299580206080652, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6405575275421143, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8709535598754883, "num_tokens": 456720553.0, "step": 12027 }, { "epoch": 1.5300852308866557, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.838249921798706, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8658998012542725, "num_tokens": 456754736.0, "step": 12028 }, { "epoch": 1.5302124411652462, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6011836528778076, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8852506875991821, "num_tokens": 456791268.0, "step": 12029 }, { "epoch": 1.5303396514438368, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6497511863708496, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8803296089172363, "num_tokens": 456831728.0, "step": 12030 }, { "epoch": 1.5304668617224273, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.536166191101074, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8774538636207581, "num_tokens": 456869375.0, "step": 12031 }, { "epoch": 1.5305940720010178, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.534926414489746, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8625454902648926, "num_tokens": 456909376.0, "step": 12032 }, { "epoch": 1.5307212822796084, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4699277877807617, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.8835576772689819, "num_tokens": 456945851.0, "step": 12033 }, { "epoch": 1.5308484925581987, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4889657497406006, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8698669075965881, "num_tokens": 456979833.0, "step": 12034 }, { "epoch": 1.5309757028367892, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3983635902404785, "learning_rate": 1e-05, "loss": 0.3918, "mean_token_accuracy": 0.8883366584777832, "num_tokens": 457016032.0, "step": 12035 }, { "epoch": 1.5311029131153797, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4742743968963623, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8864741921424866, "num_tokens": 457050015.0, "step": 12036 }, { "epoch": 1.5312301233939702, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3320116996765137, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8735710978507996, "num_tokens": 457087287.0, "step": 12037 }, { "epoch": 1.5313573336725608, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4929473400115967, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8696883916854858, "num_tokens": 457123237.0, "step": 12038 }, { "epoch": 1.5314845439511513, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2612850666046143, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8684380650520325, "num_tokens": 457161834.0, "step": 12039 }, { "epoch": 1.5316117542297416, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5541489124298096, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8623949289321899, "num_tokens": 457203672.0, "step": 12040 }, { "epoch": 1.5317389645083321, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.196254253387451, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8823836445808411, "num_tokens": 457238486.0, "step": 12041 }, { "epoch": 1.5318661747869227, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.671396017074585, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8574466705322266, "num_tokens": 457270817.0, "step": 12042 }, { "epoch": 1.5319933850655132, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.1273245811462402, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8715261220932007, "num_tokens": 457306812.0, "step": 12043 }, { "epoch": 1.5321205953441037, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8414433002471924, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8802384734153748, "num_tokens": 457340893.0, "step": 12044 }, { "epoch": 1.5322478056226942, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.096057176589966, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8733004331588745, "num_tokens": 457381668.0, "step": 12045 }, { "epoch": 1.5323750159012848, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.098605632781982, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8601428270339966, "num_tokens": 457417498.0, "step": 12046 }, { "epoch": 1.5325022261798753, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2225143909454346, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8711460828781128, "num_tokens": 457460300.0, "step": 12047 }, { "epoch": 1.5326294364584658, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.267112731933594, "learning_rate": 1e-05, "loss": 0.5045, "mean_token_accuracy": 0.8617900609970093, "num_tokens": 457502030.0, "step": 12048 }, { "epoch": 1.5327566467370564, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5223159790039062, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8797478675842285, "num_tokens": 457540170.0, "step": 12049 }, { "epoch": 1.532883857015647, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.169272422790527, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.877182126045227, "num_tokens": 457585834.0, "step": 12050 }, { "epoch": 1.5330110672942374, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.677137851715088, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8812533020973206, "num_tokens": 457622027.0, "step": 12051 }, { "epoch": 1.533138277572828, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9621410369873047, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8642585277557373, "num_tokens": 457659129.0, "step": 12052 }, { "epoch": 1.5332654878514185, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.657508373260498, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.887566089630127, "num_tokens": 457695046.0, "step": 12053 }, { "epoch": 1.533392698130009, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.76436185836792, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8776170611381531, "num_tokens": 457735650.0, "step": 12054 }, { "epoch": 1.5335199084085995, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.592395782470703, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8792915344238281, "num_tokens": 457767395.0, "step": 12055 }, { "epoch": 1.53364711868719, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6238181591033936, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.8850470781326294, "num_tokens": 457800100.0, "step": 12056 }, { "epoch": 1.5337743289657806, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5039570331573486, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8755607604980469, "num_tokens": 457837497.0, "step": 12057 }, { "epoch": 1.533901539244371, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.534759044647217, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8737305402755737, "num_tokens": 457875711.0, "step": 12058 }, { "epoch": 1.5340287495229614, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4252054691314697, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8603249788284302, "num_tokens": 457914888.0, "step": 12059 }, { "epoch": 1.534155959801552, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.508004903793335, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.8563986420631409, "num_tokens": 457953315.0, "step": 12060 }, { "epoch": 1.5342831700801425, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3668551445007324, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8582778573036194, "num_tokens": 457990857.0, "step": 12061 }, { "epoch": 1.534410380358733, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4909756183624268, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.865473747253418, "num_tokens": 458034060.0, "step": 12062 }, { "epoch": 1.5345375906373235, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.299558162689209, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8646445870399475, "num_tokens": 458074093.0, "step": 12063 }, { "epoch": 1.5346648009159138, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5425260066986084, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8792299032211304, "num_tokens": 458107374.0, "step": 12064 }, { "epoch": 1.5347920111945044, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.225964307785034, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8731603026390076, "num_tokens": 458141873.0, "step": 12065 }, { "epoch": 1.534919221473095, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6204917430877686, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8728294372558594, "num_tokens": 458181483.0, "step": 12066 }, { "epoch": 1.5350464317516854, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.167766571044922, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8641012907028198, "num_tokens": 458223644.0, "step": 12067 }, { "epoch": 1.535173642030276, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.764336585998535, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8726544380187988, "num_tokens": 458263518.0, "step": 12068 }, { "epoch": 1.5353008523088665, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.116407632827759, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8779327273368835, "num_tokens": 458296621.0, "step": 12069 }, { "epoch": 1.535428062587457, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9780702590942383, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8650392293930054, "num_tokens": 458335161.0, "step": 12070 }, { "epoch": 1.5355552728660475, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1595048904418945, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8604995012283325, "num_tokens": 458373758.0, "step": 12071 }, { "epoch": 1.535682483144638, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.189545631408691, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8751018643379211, "num_tokens": 458412662.0, "step": 12072 }, { "epoch": 1.5358096934232286, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.36838698387146, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8623545169830322, "num_tokens": 458455314.0, "step": 12073 }, { "epoch": 1.5359369037018191, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.233135223388672, "learning_rate": 1e-05, "loss": 0.534, "mean_token_accuracy": 0.851496160030365, "num_tokens": 458498633.0, "step": 12074 }, { "epoch": 1.5360641139804097, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.607105255126953, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8683595657348633, "num_tokens": 458538920.0, "step": 12075 }, { "epoch": 1.5361913242590002, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.061552047729492, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8699479699134827, "num_tokens": 458574333.0, "step": 12076 }, { "epoch": 1.5363185345375907, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6758852005004883, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8623424172401428, "num_tokens": 458612131.0, "step": 12077 }, { "epoch": 1.5364457448161812, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.854430675506592, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8674908876419067, "num_tokens": 458652169.0, "step": 12078 }, { "epoch": 1.5365729550947718, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6301398277282715, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.8655706644058228, "num_tokens": 458687816.0, "step": 12079 }, { "epoch": 1.5367001653733623, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.667703866958618, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8774186372756958, "num_tokens": 458723600.0, "step": 12080 }, { "epoch": 1.5368273756519528, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.592373847961426, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8567466139793396, "num_tokens": 458761863.0, "step": 12081 }, { "epoch": 1.5369545859305433, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.546283483505249, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8588740825653076, "num_tokens": 458794986.0, "step": 12082 }, { "epoch": 1.5370817962091337, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4854936599731445, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.862077534198761, "num_tokens": 458833939.0, "step": 12083 }, { "epoch": 1.5372090064877242, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4665112495422363, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8592858910560608, "num_tokens": 458874522.0, "step": 12084 }, { "epoch": 1.5373362167663147, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4329099655151367, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8501776456832886, "num_tokens": 458911794.0, "step": 12085 }, { "epoch": 1.5374634270449052, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.43074107170105, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8696394562721252, "num_tokens": 458951979.0, "step": 12086 }, { "epoch": 1.5375906373234958, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.379080057144165, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8614633083343506, "num_tokens": 458993657.0, "step": 12087 }, { "epoch": 1.5377178476020863, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.42738676071167, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8837239742279053, "num_tokens": 459030216.0, "step": 12088 }, { "epoch": 1.5378450578806766, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3405537605285645, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8591277599334717, "num_tokens": 459066451.0, "step": 12089 }, { "epoch": 1.5379722681592671, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.434187173843384, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.880675733089447, "num_tokens": 459106928.0, "step": 12090 }, { "epoch": 1.5380994784378577, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3033652305603027, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8606163263320923, "num_tokens": 459142769.0, "step": 12091 }, { "epoch": 1.5382266887164482, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4692108631134033, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8751413822174072, "num_tokens": 459178895.0, "step": 12092 }, { "epoch": 1.5383538989950387, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.239333391189575, "learning_rate": 1e-05, "loss": 0.3969, "mean_token_accuracy": 0.884938657283783, "num_tokens": 459214973.0, "step": 12093 }, { "epoch": 1.5384811092736292, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5445520877838135, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8679617643356323, "num_tokens": 459251279.0, "step": 12094 }, { "epoch": 1.5386083195522198, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1736338138580322, "learning_rate": 1e-05, "loss": 0.4973, "mean_token_accuracy": 0.8520063757896423, "num_tokens": 459285587.0, "step": 12095 }, { "epoch": 1.5387355298308103, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.67179012298584, "learning_rate": 1e-05, "loss": 0.503, "mean_token_accuracy": 0.85858553647995, "num_tokens": 459326103.0, "step": 12096 }, { "epoch": 1.5388627401094008, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.0934319496154785, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8687537908554077, "num_tokens": 459367075.0, "step": 12097 }, { "epoch": 1.5389899503879914, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8797225952148438, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8627206087112427, "num_tokens": 459408164.0, "step": 12098 }, { "epoch": 1.5391171606665819, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0741169452667236, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8609929084777832, "num_tokens": 459443081.0, "step": 12099 }, { "epoch": 1.5392443709451724, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1682305335998535, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.865624189376831, "num_tokens": 459481469.0, "step": 12100 }, { "epoch": 1.539371581223763, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2090861797332764, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8677846193313599, "num_tokens": 459519319.0, "step": 12101 }, { "epoch": 1.5394987915023535, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.38554573059082, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8711288571357727, "num_tokens": 459559998.0, "step": 12102 }, { "epoch": 1.539626001780944, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.566368818283081, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8837680816650391, "num_tokens": 459596765.0, "step": 12103 }, { "epoch": 1.5397532120595345, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002536773681640625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.319704055786133, "learning_rate": 1e-05, "loss": 0.5315, "mean_token_accuracy": 0.8525563478469849, "num_tokens": 459634023.0, "step": 12104 }, { "epoch": 1.539880422338125, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.784191131591797, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8854286670684814, "num_tokens": 459675355.0, "step": 12105 }, { "epoch": 1.5400076326167156, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.085829734802246, "learning_rate": 1e-05, "loss": 0.5084, "mean_token_accuracy": 0.858893632888794, "num_tokens": 459716788.0, "step": 12106 }, { "epoch": 1.5401348428953059, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7602462768554688, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8755285739898682, "num_tokens": 459752580.0, "step": 12107 }, { "epoch": 1.5402620531738964, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.884807586669922, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8753046989440918, "num_tokens": 459788594.0, "step": 12108 }, { "epoch": 1.540389263452487, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7148032188415527, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8824597001075745, "num_tokens": 459826806.0, "step": 12109 }, { "epoch": 1.5405164737310775, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6972601413726807, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8755934238433838, "num_tokens": 459870157.0, "step": 12110 }, { "epoch": 1.540643684009668, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5364623069763184, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8758125305175781, "num_tokens": 459909081.0, "step": 12111 }, { "epoch": 1.5407708942882585, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6212053298950195, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8778554201126099, "num_tokens": 459942886.0, "step": 12112 }, { "epoch": 1.5408981045668488, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4296200275421143, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.870464563369751, "num_tokens": 459979351.0, "step": 12113 }, { "epoch": 1.5410253148454394, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5957858562469482, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8679168224334717, "num_tokens": 460026058.0, "step": 12114 }, { "epoch": 1.54115252512403, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.32271671295166, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8667876720428467, "num_tokens": 460066414.0, "step": 12115 }, { "epoch": 1.5412797354026204, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6373422145843506, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8664900064468384, "num_tokens": 460107747.0, "step": 12116 }, { "epoch": 1.541406945681211, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2403247356414795, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8737379312515259, "num_tokens": 460144418.0, "step": 12117 }, { "epoch": 1.5415341559598015, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.745859384536743, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8602855205535889, "num_tokens": 460186061.0, "step": 12118 }, { "epoch": 1.541661366238392, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.174353837966919, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8649835586547852, "num_tokens": 460222672.0, "step": 12119 }, { "epoch": 1.5417885765169825, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9100053310394287, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8794036507606506, "num_tokens": 460265992.0, "step": 12120 }, { "epoch": 1.541915786795573, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.170199394226074, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8729456663131714, "num_tokens": 460305498.0, "step": 12121 }, { "epoch": 1.5420429970741636, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.139760971069336, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8662674427032471, "num_tokens": 460336013.0, "step": 12122 }, { "epoch": 1.5421702073527541, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.352057695388794, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8707172870635986, "num_tokens": 460373452.0, "step": 12123 }, { "epoch": 1.5422974176313446, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.193155765533447, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8720197677612305, "num_tokens": 460414059.0, "step": 12124 }, { "epoch": 1.5424246279099352, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5872206687927246, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8836326599121094, "num_tokens": 460446098.0, "step": 12125 }, { "epoch": 1.5425518381885257, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.016805171966553, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.87131267786026, "num_tokens": 460485384.0, "step": 12126 }, { "epoch": 1.5426790484671162, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.683682441711426, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8637773990631104, "num_tokens": 460517200.0, "step": 12127 }, { "epoch": 1.5428062587457068, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.778451681137085, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8789288997650146, "num_tokens": 460547557.0, "step": 12128 }, { "epoch": 1.5429334690242973, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.638195514678955, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8787833452224731, "num_tokens": 460583189.0, "step": 12129 }, { "epoch": 1.5430606793028878, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5642693042755127, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8744523525238037, "num_tokens": 460623619.0, "step": 12130 }, { "epoch": 1.5431878895814783, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5919954776763916, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8662988543510437, "num_tokens": 460666667.0, "step": 12131 }, { "epoch": 1.5433150998600687, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4230260848999023, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8700318932533264, "num_tokens": 460704292.0, "step": 12132 }, { "epoch": 1.5434423101386592, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5661752223968506, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8741548657417297, "num_tokens": 460748153.0, "step": 12133 }, { "epoch": 1.5435695204172497, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3260035514831543, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8750813007354736, "num_tokens": 460787663.0, "step": 12134 }, { "epoch": 1.5436967306958402, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5779428482055664, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.879088819026947, "num_tokens": 460821643.0, "step": 12135 }, { "epoch": 1.5438239409744308, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.269854784011841, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.862095057964325, "num_tokens": 460857180.0, "step": 12136 }, { "epoch": 1.5439511512530213, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.625258207321167, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8673291206359863, "num_tokens": 460891573.0, "step": 12137 }, { "epoch": 1.5440783615316116, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1870217323303223, "learning_rate": 1e-05, "loss": 0.4001, "mean_token_accuracy": 0.8837361335754395, "num_tokens": 460927727.0, "step": 12138 }, { "epoch": 1.5442055718102021, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.733355760574341, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8619521856307983, "num_tokens": 460965531.0, "step": 12139 }, { "epoch": 1.5443327820887927, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1342015266418457, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8714462518692017, "num_tokens": 461002900.0, "step": 12140 }, { "epoch": 1.5444599923673832, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8945794105529785, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8716354370117188, "num_tokens": 461041569.0, "step": 12141 }, { "epoch": 1.5445872026459737, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.141864538192749, "learning_rate": 1e-05, "loss": 0.4751, "mean_token_accuracy": 0.8613766431808472, "num_tokens": 461080453.0, "step": 12142 }, { "epoch": 1.5447144129245642, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.0944695472717285, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8670549988746643, "num_tokens": 461120963.0, "step": 12143 }, { "epoch": 1.5448416232031548, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.285621166229248, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8748401403427124, "num_tokens": 461156441.0, "step": 12144 }, { "epoch": 1.5449688334817453, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.200247764587402, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8813313245773315, "num_tokens": 461189058.0, "step": 12145 }, { "epoch": 1.5450960437603358, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5022332668304443, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8784341812133789, "num_tokens": 461223962.0, "step": 12146 }, { "epoch": 1.5452232540389264, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.109148979187012, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8786811232566833, "num_tokens": 461262114.0, "step": 12147 }, { "epoch": 1.5453504643175169, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6340718269348145, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8721121549606323, "num_tokens": 461294633.0, "step": 12148 }, { "epoch": 1.5454776745961074, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9229180812835693, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8660003542900085, "num_tokens": 461338549.0, "step": 12149 }, { "epoch": 1.545604884874698, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6329104900360107, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8665279746055603, "num_tokens": 461372612.0, "step": 12150 }, { "epoch": 1.5457320951532885, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7440671920776367, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8607528805732727, "num_tokens": 461404060.0, "step": 12151 }, { "epoch": 1.545859305431879, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.574887752532959, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8686687350273132, "num_tokens": 461436728.0, "step": 12152 }, { "epoch": 1.5459865157104695, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.578111410140991, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8809552192687988, "num_tokens": 461475418.0, "step": 12153 }, { "epoch": 1.54611372598906, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.503612756729126, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8620777130126953, "num_tokens": 461512094.0, "step": 12154 }, { "epoch": 1.5462409362676506, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.50101637840271, "learning_rate": 1e-05, "loss": 0.5085, "mean_token_accuracy": 0.8524013161659241, "num_tokens": 461553454.0, "step": 12155 }, { "epoch": 1.5463681465462409, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4479103088378906, "learning_rate": 1e-05, "loss": 0.5343, "mean_token_accuracy": 0.8504468202590942, "num_tokens": 461590818.0, "step": 12156 }, { "epoch": 1.5464953568248314, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4393579959869385, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8835128545761108, "num_tokens": 461638692.0, "step": 12157 }, { "epoch": 1.546622567103422, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.394683837890625, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8713535070419312, "num_tokens": 461681040.0, "step": 12158 }, { "epoch": 1.5467497773820125, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4292352199554443, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.871120035648346, "num_tokens": 461716999.0, "step": 12159 }, { "epoch": 1.546876987660603, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3519601821899414, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8625435829162598, "num_tokens": 461755053.0, "step": 12160 }, { "epoch": 1.5470041979391935, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.432401180267334, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8759879469871521, "num_tokens": 461795384.0, "step": 12161 }, { "epoch": 1.5471314082177838, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2941720485687256, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.872282087802887, "num_tokens": 461843951.0, "step": 12162 }, { "epoch": 1.5472586184963744, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.47243595123291, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8547915816307068, "num_tokens": 461884172.0, "step": 12163 }, { "epoch": 1.5473858287749649, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2417333126068115, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8672292232513428, "num_tokens": 461928493.0, "step": 12164 }, { "epoch": 1.5475130390535554, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.542619228363037, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8747727870941162, "num_tokens": 461964462.0, "step": 12165 }, { "epoch": 1.547640249332146, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1694278717041016, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8777408599853516, "num_tokens": 462004123.0, "step": 12166 }, { "epoch": 1.5477674596107365, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.691225290298462, "learning_rate": 1e-05, "loss": 0.5111, "mean_token_accuracy": 0.8519534468650818, "num_tokens": 462042906.0, "step": 12167 }, { "epoch": 1.547894669889327, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0812559127807617, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8818058371543884, "num_tokens": 462083802.0, "step": 12168 }, { "epoch": 1.5480218801679175, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.927795648574829, "learning_rate": 1e-05, "loss": 0.4726, "mean_token_accuracy": 0.8655531406402588, "num_tokens": 462119861.0, "step": 12169 }, { "epoch": 1.548149090446508, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0600078105926514, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8635576963424683, "num_tokens": 462162445.0, "step": 12170 }, { "epoch": 1.5482763007250986, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.234039783477783, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8686230182647705, "num_tokens": 462195176.0, "step": 12171 }, { "epoch": 1.5484035110036891, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.280768632888794, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8644759058952332, "num_tokens": 462236155.0, "step": 12172 }, { "epoch": 1.5485307212822796, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.420316219329834, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8706759810447693, "num_tokens": 462267099.0, "step": 12173 }, { "epoch": 1.5486579315608702, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.665693521499634, "learning_rate": 1e-05, "loss": 0.5084, "mean_token_accuracy": 0.8582040667533875, "num_tokens": 462311485.0, "step": 12174 }, { "epoch": 1.5487851418394607, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2464680671691895, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8708764314651489, "num_tokens": 462349078.0, "step": 12175 }, { "epoch": 1.5489123521180512, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.80796480178833, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8694899082183838, "num_tokens": 462381477.0, "step": 12176 }, { "epoch": 1.5490395623966418, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 4.00792932510376, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8701366186141968, "num_tokens": 462415996.0, "step": 12177 }, { "epoch": 1.5491667726752323, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.726358413696289, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8754703998565674, "num_tokens": 462446737.0, "step": 12178 }, { "epoch": 1.5492939829538228, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8170361518859863, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8638375401496887, "num_tokens": 462487742.0, "step": 12179 }, { "epoch": 1.5494211932324133, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5985934734344482, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8740500211715698, "num_tokens": 462529231.0, "step": 12180 }, { "epoch": 1.5495484035110036, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7059414386749268, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8729156255722046, "num_tokens": 462565608.0, "step": 12181 }, { "epoch": 1.5496756137895942, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4905643463134766, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8732379674911499, "num_tokens": 462600919.0, "step": 12182 }, { "epoch": 1.5498028240681847, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6500704288482666, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8665030002593994, "num_tokens": 462635594.0, "step": 12183 }, { "epoch": 1.5499300343467752, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.380697011947632, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8573176860809326, "num_tokens": 462672349.0, "step": 12184 }, { "epoch": 1.5500572446253658, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.665426015853882, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8739924430847168, "num_tokens": 462710286.0, "step": 12185 }, { "epoch": 1.5501844549039563, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2923953533172607, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8682835698127747, "num_tokens": 462744721.0, "step": 12186 }, { "epoch": 1.5503116651825466, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7288427352905273, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8826924562454224, "num_tokens": 462782812.0, "step": 12187 }, { "epoch": 1.5504388754611371, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2060303688049316, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.878726065158844, "num_tokens": 462827919.0, "step": 12188 }, { "epoch": 1.5505660857397277, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8849945068359375, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8661572933197021, "num_tokens": 462864878.0, "step": 12189 }, { "epoch": 1.5506932960183182, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.199065923690796, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8781315684318542, "num_tokens": 462906213.0, "step": 12190 }, { "epoch": 1.5508205062969087, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.084561347961426, "learning_rate": 1e-05, "loss": 0.5378, "mean_token_accuracy": 0.8506357669830322, "num_tokens": 462943078.0, "step": 12191 }, { "epoch": 1.5509477165754992, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.336320638656616, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8793368339538574, "num_tokens": 462982444.0, "step": 12192 }, { "epoch": 1.5510749268540898, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.139636993408203, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8798466324806213, "num_tokens": 463024335.0, "step": 12193 }, { "epoch": 1.5512021371326803, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.54282546043396, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.87254399061203, "num_tokens": 463062278.0, "step": 12194 }, { "epoch": 1.5513293474112708, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.998112440109253, "learning_rate": 1e-05, "loss": 0.5252, "mean_token_accuracy": 0.849536120891571, "num_tokens": 463100257.0, "step": 12195 }, { "epoch": 1.5514565576898613, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6427159309387207, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8736754655838013, "num_tokens": 463139468.0, "step": 12196 }, { "epoch": 1.5515837679684519, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7637624740600586, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8655388355255127, "num_tokens": 463176145.0, "step": 12197 }, { "epoch": 1.5517109782470424, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6361241340637207, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8790215253829956, "num_tokens": 463208899.0, "step": 12198 }, { "epoch": 1.551838188525633, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5302839279174805, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8742809295654297, "num_tokens": 463250163.0, "step": 12199 }, { "epoch": 1.5519653988042235, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.626997232437134, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8666577339172363, "num_tokens": 463285447.0, "step": 12200 }, { "epoch": 1.552092609082814, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.359497547149658, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.885989785194397, "num_tokens": 463322601.0, "step": 12201 }, { "epoch": 1.5522198193614045, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.646782875061035, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8799194693565369, "num_tokens": 463360624.0, "step": 12202 }, { "epoch": 1.552347029639995, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.244631052017212, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.872111439704895, "num_tokens": 463401662.0, "step": 12203 }, { "epoch": 1.5524742399185856, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7286791801452637, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8682572841644287, "num_tokens": 463433999.0, "step": 12204 }, { "epoch": 1.5526014501971759, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1861324310302734, "learning_rate": 1e-05, "loss": 0.5119, "mean_token_accuracy": 0.8535304069519043, "num_tokens": 463473673.0, "step": 12205 }, { "epoch": 1.5527286604757664, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.857926368713379, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8587691783905029, "num_tokens": 463507197.0, "step": 12206 }, { "epoch": 1.552855870754357, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1752281188964844, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8712316155433655, "num_tokens": 463540777.0, "step": 12207 }, { "epoch": 1.5529830810329475, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 4.004171371459961, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8883926868438721, "num_tokens": 463573335.0, "step": 12208 }, { "epoch": 1.553110291311538, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2450830936431885, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.873928427696228, "num_tokens": 463613898.0, "step": 12209 }, { "epoch": 1.5532375015901285, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.116730690002441, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.883547842502594, "num_tokens": 463648948.0, "step": 12210 }, { "epoch": 1.5533647118687188, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.425617218017578, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8787776231765747, "num_tokens": 463683417.0, "step": 12211 }, { "epoch": 1.5534919221473094, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.100550174713135, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8616417646408081, "num_tokens": 463719533.0, "step": 12212 }, { "epoch": 1.5536191324258999, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5524251461029053, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8804453015327454, "num_tokens": 463757740.0, "step": 12213 }, { "epoch": 1.5537463427044904, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.952704906463623, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8705885410308838, "num_tokens": 463795619.0, "step": 12214 }, { "epoch": 1.553873552983081, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.586953639984131, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8727771043777466, "num_tokens": 463829934.0, "step": 12215 }, { "epoch": 1.5540007632616715, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7953810691833496, "learning_rate": 1e-05, "loss": 0.5163, "mean_token_accuracy": 0.848222017288208, "num_tokens": 463869584.0, "step": 12216 }, { "epoch": 1.554127973540262, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.543097734451294, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8557429909706116, "num_tokens": 463909220.0, "step": 12217 }, { "epoch": 1.5542551838188525, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.634827136993408, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8772351741790771, "num_tokens": 463949367.0, "step": 12218 }, { "epoch": 1.554382394097443, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.493396759033203, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.862328052520752, "num_tokens": 463986037.0, "step": 12219 }, { "epoch": 1.5545096043760336, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.542508840560913, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8794646859169006, "num_tokens": 464021903.0, "step": 12220 }, { "epoch": 1.554636814654624, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4225447177886963, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8756885528564453, "num_tokens": 464058738.0, "step": 12221 }, { "epoch": 1.5547640249332146, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.516204595565796, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.8525235652923584, "num_tokens": 464090791.0, "step": 12222 }, { "epoch": 1.5548912352118052, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.352748155593872, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8691316843032837, "num_tokens": 464133897.0, "step": 12223 }, { "epoch": 1.5550184454903957, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.509983777999878, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8782532811164856, "num_tokens": 464173808.0, "step": 12224 }, { "epoch": 1.5551456557689862, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.296842098236084, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8739084005355835, "num_tokens": 464209456.0, "step": 12225 }, { "epoch": 1.5552728660475768, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5537073612213135, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.861674427986145, "num_tokens": 464250192.0, "step": 12226 }, { "epoch": 1.5554000763261673, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2327287197113037, "learning_rate": 1e-05, "loss": 0.5103, "mean_token_accuracy": 0.8517159223556519, "num_tokens": 464291485.0, "step": 12227 }, { "epoch": 1.5555272866047578, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.649535655975342, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8622195720672607, "num_tokens": 464329963.0, "step": 12228 }, { "epoch": 1.5556544968833483, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1698296070098877, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8731601238250732, "num_tokens": 464366944.0, "step": 12229 }, { "epoch": 1.5557817071619386, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7720582485198975, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8634074926376343, "num_tokens": 464409070.0, "step": 12230 }, { "epoch": 1.5559089174405292, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1410348415374756, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8773975372314453, "num_tokens": 464448305.0, "step": 12231 }, { "epoch": 1.5560361277191197, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.964106798171997, "learning_rate": 1e-05, "loss": 0.5215, "mean_token_accuracy": 0.8533646464347839, "num_tokens": 464482453.0, "step": 12232 }, { "epoch": 1.5561633379977102, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.205979108810425, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8594964742660522, "num_tokens": 464518692.0, "step": 12233 }, { "epoch": 1.5562905482763008, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.098401069641113, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8647868633270264, "num_tokens": 464554502.0, "step": 12234 }, { "epoch": 1.5564177585548913, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.362288236618042, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8780356645584106, "num_tokens": 464591414.0, "step": 12235 }, { "epoch": 1.5565449688334816, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.113255023956299, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8648054599761963, "num_tokens": 464627116.0, "step": 12236 }, { "epoch": 1.5566721791120721, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.553297758102417, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8700711131095886, "num_tokens": 464661783.0, "step": 12237 }, { "epoch": 1.5567993893906626, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.942915439605713, "learning_rate": 1e-05, "loss": 0.5051, "mean_token_accuracy": 0.8595184087753296, "num_tokens": 464704851.0, "step": 12238 }, { "epoch": 1.5569265996692532, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.592866897583008, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.861100971698761, "num_tokens": 464746583.0, "step": 12239 }, { "epoch": 1.5570538099478437, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7401039600372314, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8653801083564758, "num_tokens": 464783275.0, "step": 12240 }, { "epoch": 1.5571810202264342, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.573287010192871, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8759443759918213, "num_tokens": 464818436.0, "step": 12241 }, { "epoch": 1.5573082305050248, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5701003074645996, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8682702779769897, "num_tokens": 464859708.0, "step": 12242 }, { "epoch": 1.5574354407836153, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5443685054779053, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8744041919708252, "num_tokens": 464894339.0, "step": 12243 }, { "epoch": 1.5575626510622058, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4412403106689453, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8776889443397522, "num_tokens": 464931517.0, "step": 12244 }, { "epoch": 1.5576898613407963, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5168845653533936, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8672624826431274, "num_tokens": 464971616.0, "step": 12245 }, { "epoch": 1.5578170716193869, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.352756977081299, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8784774541854858, "num_tokens": 465012551.0, "step": 12246 }, { "epoch": 1.5579442818979774, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.507361650466919, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8803894519805908, "num_tokens": 465054266.0, "step": 12247 }, { "epoch": 1.558071492176568, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.28834867477417, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8764963150024414, "num_tokens": 465090057.0, "step": 12248 }, { "epoch": 1.5581987024551585, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5271475315093994, "learning_rate": 1e-05, "loss": 0.3959, "mean_token_accuracy": 0.8867238759994507, "num_tokens": 465132169.0, "step": 12249 }, { "epoch": 1.558325912733749, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2471303939819336, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8679454326629639, "num_tokens": 465169922.0, "step": 12250 }, { "epoch": 1.5584531230123395, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5831098556518555, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8844553232192993, "num_tokens": 465207792.0, "step": 12251 }, { "epoch": 1.55858033329093, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1813766956329346, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8614044189453125, "num_tokens": 465250498.0, "step": 12252 }, { "epoch": 1.5587075435695206, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.706373929977417, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8657729625701904, "num_tokens": 465288321.0, "step": 12253 }, { "epoch": 1.5588347538481109, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1135451793670654, "learning_rate": 1e-05, "loss": 0.3572, "mean_token_accuracy": 0.8952562808990479, "num_tokens": 465322241.0, "step": 12254 }, { "epoch": 1.5589619641267014, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8868613243103027, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8849670886993408, "num_tokens": 465359680.0, "step": 12255 }, { "epoch": 1.559089174405292, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1047239303588867, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8768489956855774, "num_tokens": 465396951.0, "step": 12256 }, { "epoch": 1.5592163846838825, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.130624771118164, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8676834106445312, "num_tokens": 465433087.0, "step": 12257 }, { "epoch": 1.559343594962473, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2491793632507324, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8767210841178894, "num_tokens": 465472810.0, "step": 12258 }, { "epoch": 1.5594708052410635, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.291326999664307, "learning_rate": 1e-05, "loss": 0.4997, "mean_token_accuracy": 0.8663437366485596, "num_tokens": 465512714.0, "step": 12259 }, { "epoch": 1.5595980155196538, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5414607524871826, "learning_rate": 1e-05, "loss": 0.503, "mean_token_accuracy": 0.8553171753883362, "num_tokens": 465547629.0, "step": 12260 }, { "epoch": 1.5597252257982444, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.202401638031006, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8649626970291138, "num_tokens": 465584226.0, "step": 12261 }, { "epoch": 1.5598524360768349, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.686206102371216, "learning_rate": 1e-05, "loss": 0.5083, "mean_token_accuracy": 0.8584668040275574, "num_tokens": 465625854.0, "step": 12262 }, { "epoch": 1.5599796463554254, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.015388488769531, "learning_rate": 1e-05, "loss": 0.5101, "mean_token_accuracy": 0.8519983887672424, "num_tokens": 465660670.0, "step": 12263 }, { "epoch": 1.560106856634016, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6735100746154785, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8686937689781189, "num_tokens": 465700359.0, "step": 12264 }, { "epoch": 1.5602340669126065, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.825629711151123, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8728957176208496, "num_tokens": 465733648.0, "step": 12265 }, { "epoch": 1.560361277191197, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6061041355133057, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.862800121307373, "num_tokens": 465769814.0, "step": 12266 }, { "epoch": 1.5604884874697875, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6741490364074707, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8764561414718628, "num_tokens": 465804637.0, "step": 12267 }, { "epoch": 1.560615697748378, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5216548442840576, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8832280039787292, "num_tokens": 465838649.0, "step": 12268 }, { "epoch": 1.5607429080269686, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5787010192871094, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8826795220375061, "num_tokens": 465870628.0, "step": 12269 }, { "epoch": 1.560870118305559, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.444362163543701, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8658918142318726, "num_tokens": 465905612.0, "step": 12270 }, { "epoch": 1.5609973285841496, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.526719093322754, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8693691492080688, "num_tokens": 465945964.0, "step": 12271 }, { "epoch": 1.5611245388627402, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3715014457702637, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8842331171035767, "num_tokens": 465979600.0, "step": 12272 }, { "epoch": 1.5612517491413307, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.534550666809082, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8650599718093872, "num_tokens": 466017406.0, "step": 12273 }, { "epoch": 1.5613789594199212, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.289710283279419, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8657569885253906, "num_tokens": 466059178.0, "step": 12274 }, { "epoch": 1.5615061696985117, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.593553066253662, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.867109477519989, "num_tokens": 466096385.0, "step": 12275 }, { "epoch": 1.5616333799771023, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2222740650177, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8706806898117065, "num_tokens": 466131160.0, "step": 12276 }, { "epoch": 1.5617605902556928, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.689207077026367, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8725674152374268, "num_tokens": 466174842.0, "step": 12277 }, { "epoch": 1.5618878005342833, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1548476219177246, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8770511150360107, "num_tokens": 466212943.0, "step": 12278 }, { "epoch": 1.5620150108128736, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.858318328857422, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8857626914978027, "num_tokens": 466251560.0, "step": 12279 }, { "epoch": 1.5621422210914642, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.14922833442688, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8593540191650391, "num_tokens": 466287150.0, "step": 12280 }, { "epoch": 1.5622694313700547, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.058883190155029, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8689802885055542, "num_tokens": 466328508.0, "step": 12281 }, { "epoch": 1.5623966416486452, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.266167163848877, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8655263781547546, "num_tokens": 466362552.0, "step": 12282 }, { "epoch": 1.5625238519272358, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1632890701293945, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8671118021011353, "num_tokens": 466404253.0, "step": 12283 }, { "epoch": 1.5626510622058263, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4766297340393066, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8615059852600098, "num_tokens": 466443258.0, "step": 12284 }, { "epoch": 1.5627782724844166, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.099124908447266, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8692019581794739, "num_tokens": 466477575.0, "step": 12285 }, { "epoch": 1.5629054827630071, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6146771907806396, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8675663471221924, "num_tokens": 466516301.0, "step": 12286 }, { "epoch": 1.5630326930415976, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8981668949127197, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8605548143386841, "num_tokens": 466553904.0, "step": 12287 }, { "epoch": 1.5631599033201882, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.637195587158203, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8647496700286865, "num_tokens": 466591668.0, "step": 12288 }, { "epoch": 1.5632871135987787, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6841115951538086, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8689631223678589, "num_tokens": 466626484.0, "step": 12289 }, { "epoch": 1.5634143238773692, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5796661376953125, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8802589178085327, "num_tokens": 466663590.0, "step": 12290 }, { "epoch": 1.5635415341559598, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.523686408996582, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8832540512084961, "num_tokens": 466699856.0, "step": 12291 }, { "epoch": 1.5636687444345503, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5362863540649414, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8845945596694946, "num_tokens": 466738703.0, "step": 12292 }, { "epoch": 1.5637959547131408, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4299113750457764, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8755289316177368, "num_tokens": 466770948.0, "step": 12293 }, { "epoch": 1.5639231649917313, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.51169753074646, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8720263838768005, "num_tokens": 466804655.0, "step": 12294 }, { "epoch": 1.5640503752703219, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3535335063934326, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8767500519752502, "num_tokens": 466847592.0, "step": 12295 }, { "epoch": 1.5641775855489124, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4947431087493896, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8780251741409302, "num_tokens": 466887855.0, "step": 12296 }, { "epoch": 1.564304795827503, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3029961585998535, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8669082522392273, "num_tokens": 466925749.0, "step": 12297 }, { "epoch": 1.5644320061060935, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.508662462234497, "learning_rate": 1e-05, "loss": 0.3795, "mean_token_accuracy": 0.893650233745575, "num_tokens": 466966487.0, "step": 12298 }, { "epoch": 1.564559216384684, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.257150650024414, "learning_rate": 1e-05, "loss": 0.5255, "mean_token_accuracy": 0.8473782539367676, "num_tokens": 467008673.0, "step": 12299 }, { "epoch": 1.5646864266632745, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.552030086517334, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8784317970275879, "num_tokens": 467048389.0, "step": 12300 }, { "epoch": 1.564813636941865, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1925272941589355, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8753259181976318, "num_tokens": 467087059.0, "step": 12301 }, { "epoch": 1.5649408472204556, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6572911739349365, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.88813316822052, "num_tokens": 467123003.0, "step": 12302 }, { "epoch": 1.5650680574990459, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.135437488555908, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8597694635391235, "num_tokens": 467160526.0, "step": 12303 }, { "epoch": 1.5651952677776364, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8170266151428223, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8690593242645264, "num_tokens": 467195846.0, "step": 12304 }, { "epoch": 1.565322478056227, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.100001096725464, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8597946166992188, "num_tokens": 467232412.0, "step": 12305 }, { "epoch": 1.5654496883348175, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.049686908721924, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8653919696807861, "num_tokens": 467269695.0, "step": 12306 }, { "epoch": 1.565576898613408, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.176535129547119, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8561768531799316, "num_tokens": 467304395.0, "step": 12307 }, { "epoch": 1.5657041088919985, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.253417491912842, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.870180606842041, "num_tokens": 467348606.0, "step": 12308 }, { "epoch": 1.5658313191705888, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.41768741607666, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8693199157714844, "num_tokens": 467382673.0, "step": 12309 }, { "epoch": 1.5659585294491793, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.283227443695068, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8780665397644043, "num_tokens": 467426291.0, "step": 12310 }, { "epoch": 1.5660857397277699, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.651270866394043, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8683174848556519, "num_tokens": 467469299.0, "step": 12311 }, { "epoch": 1.5662129500063604, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.122166156768799, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8654372692108154, "num_tokens": 467510876.0, "step": 12312 }, { "epoch": 1.566340160284951, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7197179794311523, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8710978627204895, "num_tokens": 467546112.0, "step": 12313 }, { "epoch": 1.5664673705635415, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9159016609191895, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8732887506484985, "num_tokens": 467580009.0, "step": 12314 }, { "epoch": 1.566594580842132, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6476082801818848, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8630083203315735, "num_tokens": 467621009.0, "step": 12315 }, { "epoch": 1.5667217911207225, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7319746017456055, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8820891380310059, "num_tokens": 467662304.0, "step": 12316 }, { "epoch": 1.566849001399313, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6138389110565186, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8553763031959534, "num_tokens": 467697608.0, "step": 12317 }, { "epoch": 1.5669762116779036, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5944459438323975, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8724822402000427, "num_tokens": 467739978.0, "step": 12318 }, { "epoch": 1.567103421956494, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4883666038513184, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.860762357711792, "num_tokens": 467775055.0, "step": 12319 }, { "epoch": 1.5672306322350846, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.531536102294922, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8751178979873657, "num_tokens": 467812385.0, "step": 12320 }, { "epoch": 1.5673578425136752, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4174365997314453, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.854806661605835, "num_tokens": 467846784.0, "step": 12321 }, { "epoch": 1.5674850527922657, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.509962558746338, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8674801588058472, "num_tokens": 467880259.0, "step": 12322 }, { "epoch": 1.5676122630708562, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.366342544555664, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8713790774345398, "num_tokens": 467914613.0, "step": 12323 }, { "epoch": 1.5677394733494467, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5316691398620605, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8489464521408081, "num_tokens": 467953070.0, "step": 12324 }, { "epoch": 1.5678666836280373, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2774345874786377, "learning_rate": 1e-05, "loss": 0.4878, "mean_token_accuracy": 0.8579668402671814, "num_tokens": 467989404.0, "step": 12325 }, { "epoch": 1.5679938939066278, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5694382190704346, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.86872798204422, "num_tokens": 468030670.0, "step": 12326 }, { "epoch": 1.5681211041852183, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.208878993988037, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8686227798461914, "num_tokens": 468064130.0, "step": 12327 }, { "epoch": 1.5682483144638086, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6878926753997803, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8701225519180298, "num_tokens": 468098393.0, "step": 12328 }, { "epoch": 1.5683755247423992, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1373629570007324, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8811019659042358, "num_tokens": 468134961.0, "step": 12329 }, { "epoch": 1.5685027350209897, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.863978862762451, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8807592391967773, "num_tokens": 468169389.0, "step": 12330 }, { "epoch": 1.5686299452995802, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.113945245742798, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8603227734565735, "num_tokens": 468211504.0, "step": 12331 }, { "epoch": 1.5687571555781707, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.111143112182617, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8591182231903076, "num_tokens": 468245204.0, "step": 12332 }, { "epoch": 1.5688843658567613, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2689549922943115, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8717421889305115, "num_tokens": 468277054.0, "step": 12333 }, { "epoch": 1.5690115761353516, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.228839874267578, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8840316534042358, "num_tokens": 468316399.0, "step": 12334 }, { "epoch": 1.569138786413942, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.509495735168457, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8773495554924011, "num_tokens": 468351943.0, "step": 12335 }, { "epoch": 1.5692659966925326, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.158581733703613, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8687120676040649, "num_tokens": 468387443.0, "step": 12336 }, { "epoch": 1.5693932069711232, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6718015670776367, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8731567859649658, "num_tokens": 468420503.0, "step": 12337 }, { "epoch": 1.5695204172497137, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9708406925201416, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8644206523895264, "num_tokens": 468456171.0, "step": 12338 }, { "epoch": 1.5696476275283042, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.651188850402832, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8827536106109619, "num_tokens": 468494121.0, "step": 12339 }, { "epoch": 1.5697748378068948, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7592127323150635, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8731440305709839, "num_tokens": 468540960.0, "step": 12340 }, { "epoch": 1.5699020480854853, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5918753147125244, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.868465781211853, "num_tokens": 468577851.0, "step": 12341 }, { "epoch": 1.5700292583640758, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6269965171813965, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8683868646621704, "num_tokens": 468613210.0, "step": 12342 }, { "epoch": 1.5701564686426663, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.493527889251709, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8673317432403564, "num_tokens": 468655871.0, "step": 12343 }, { "epoch": 1.5702836789212569, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5463099479675293, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.878940224647522, "num_tokens": 468692757.0, "step": 12344 }, { "epoch": 1.5704108891998474, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.42663836479187, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8785002827644348, "num_tokens": 468730163.0, "step": 12345 }, { "epoch": 1.570538099478438, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5063371658325195, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8725833892822266, "num_tokens": 468768578.0, "step": 12346 }, { "epoch": 1.5706653097570284, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.362342357635498, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8834481239318848, "num_tokens": 468802027.0, "step": 12347 }, { "epoch": 1.570792520035619, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5106406211853027, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8595207929611206, "num_tokens": 468845266.0, "step": 12348 }, { "epoch": 1.5709197303142095, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.308612108230591, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8746980428695679, "num_tokens": 468879780.0, "step": 12349 }, { "epoch": 1.5710469405928, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.542355537414551, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8735203742980957, "num_tokens": 468916846.0, "step": 12350 }, { "epoch": 1.5711741508713906, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2460176944732666, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8832780122756958, "num_tokens": 468947382.0, "step": 12351 }, { "epoch": 1.5713013611499809, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.609661102294922, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8840548396110535, "num_tokens": 468987258.0, "step": 12352 }, { "epoch": 1.5714285714285714, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.171731472015381, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8762050271034241, "num_tokens": 469024878.0, "step": 12353 }, { "epoch": 1.571555781707162, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7522103786468506, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8733952045440674, "num_tokens": 469065636.0, "step": 12354 }, { "epoch": 1.5716829919857525, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.119494915008545, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8697452545166016, "num_tokens": 469099817.0, "step": 12355 }, { "epoch": 1.571810202264343, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.959782361984253, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8671576976776123, "num_tokens": 469135185.0, "step": 12356 }, { "epoch": 1.5719374125429335, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1513848304748535, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.8871767520904541, "num_tokens": 469173980.0, "step": 12357 }, { "epoch": 1.5720646228215238, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.172255516052246, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8749713897705078, "num_tokens": 469206939.0, "step": 12358 }, { "epoch": 1.5721918331001143, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3583192825317383, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8782227039337158, "num_tokens": 469245261.0, "step": 12359 }, { "epoch": 1.5723190433787049, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2203288078308105, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8689603805541992, "num_tokens": 469285569.0, "step": 12360 }, { "epoch": 1.5724462536572954, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5870227813720703, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8676657676696777, "num_tokens": 469325249.0, "step": 12361 }, { "epoch": 1.572573463935886, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.075194358825684, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8627081513404846, "num_tokens": 469360187.0, "step": 12362 }, { "epoch": 1.5727006742144765, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6636135578155518, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8666443824768066, "num_tokens": 469399363.0, "step": 12363 }, { "epoch": 1.572827884493067, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8581080436706543, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.868123471736908, "num_tokens": 469441523.0, "step": 12364 }, { "epoch": 1.5729550947716575, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6259124279022217, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8794008493423462, "num_tokens": 469475169.0, "step": 12365 }, { "epoch": 1.573082305050248, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000255584716796875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.689393997192383, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8664525151252747, "num_tokens": 469509794.0, "step": 12366 }, { "epoch": 1.5732095153288386, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5583624839782715, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8661595582962036, "num_tokens": 469549007.0, "step": 12367 }, { "epoch": 1.573336725607429, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.548426866531372, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8751479983329773, "num_tokens": 469586959.0, "step": 12368 }, { "epoch": 1.5734639358860196, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4955337047576904, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.877722442150116, "num_tokens": 469617767.0, "step": 12369 }, { "epoch": 1.5735911461646102, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4680416584014893, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8670458793640137, "num_tokens": 469655300.0, "step": 12370 }, { "epoch": 1.5737183564432007, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.44160795211792, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8735100626945496, "num_tokens": 469693666.0, "step": 12371 }, { "epoch": 1.5738455667217912, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4335734844207764, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.87312912940979, "num_tokens": 469731081.0, "step": 12372 }, { "epoch": 1.5739727770003817, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3907742500305176, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8761059045791626, "num_tokens": 469768081.0, "step": 12373 }, { "epoch": 1.5740999872789723, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4180116653442383, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.870404839515686, "num_tokens": 469806992.0, "step": 12374 }, { "epoch": 1.5742271975575628, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3594372272491455, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8664628863334656, "num_tokens": 469844969.0, "step": 12375 }, { "epoch": 1.5743544078361533, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.418612480163574, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8780184984207153, "num_tokens": 469881558.0, "step": 12376 }, { "epoch": 1.5744816181147436, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.340702772140503, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8699256181716919, "num_tokens": 469914659.0, "step": 12377 }, { "epoch": 1.5746088283933342, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.4213204383850098, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8653175830841064, "num_tokens": 469956012.0, "step": 12378 }, { "epoch": 1.5747360386719247, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2963194847106934, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8654820322990417, "num_tokens": 469993840.0, "step": 12379 }, { "epoch": 1.5748632489505152, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4549102783203125, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8609227538108826, "num_tokens": 470032741.0, "step": 12380 }, { "epoch": 1.5749904592291057, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2466557025909424, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.8825847506523132, "num_tokens": 470068969.0, "step": 12381 }, { "epoch": 1.5751176695076963, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 4.882691860198975, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8769845962524414, "num_tokens": 470108972.0, "step": 12382 }, { "epoch": 1.5752448797862866, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3735876083374023, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8861984014511108, "num_tokens": 470146655.0, "step": 12383 }, { "epoch": 1.575372090064877, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.208058834075928, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8700523972511292, "num_tokens": 470184068.0, "step": 12384 }, { "epoch": 1.5754993003434676, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3164255619049072, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8691286444664001, "num_tokens": 470221328.0, "step": 12385 }, { "epoch": 1.5756265106220582, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0002574920654296875, "ewc_loss_parallel": 7.390975952148438e-05, "grad_norm": 4.575303554534912, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8692564368247986, "num_tokens": 470261393.0, "step": 12386 }, { "epoch": 1.5757537209006487, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.748896598815918, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8689206838607788, "num_tokens": 470293705.0, "step": 12387 }, { "epoch": 1.5758809311792392, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.470136642456055, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8693287372589111, "num_tokens": 470329891.0, "step": 12388 }, { "epoch": 1.5760081414578297, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.962528705596924, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8722045421600342, "num_tokens": 470368763.0, "step": 12389 }, { "epoch": 1.5761353517364203, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.206682205200195, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8715910911560059, "num_tokens": 470407267.0, "step": 12390 }, { "epoch": 1.5762625620150108, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.879906415939331, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8695789575576782, "num_tokens": 470448872.0, "step": 12391 }, { "epoch": 1.5763897722936013, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.021510124206543, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8772466778755188, "num_tokens": 470486460.0, "step": 12392 }, { "epoch": 1.5765169825721919, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.721367359161377, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8822473883628845, "num_tokens": 470528800.0, "step": 12393 }, { "epoch": 1.5766441928507824, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9195029735565186, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.87367844581604, "num_tokens": 470568185.0, "step": 12394 }, { "epoch": 1.576771403129373, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.583364486694336, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8699488639831543, "num_tokens": 470607785.0, "step": 12395 }, { "epoch": 1.5768986134079634, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8814644813537598, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8661612868309021, "num_tokens": 470648536.0, "step": 12396 }, { "epoch": 1.577025823686554, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.444814920425415, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8791496753692627, "num_tokens": 470686071.0, "step": 12397 }, { "epoch": 1.5771530339651445, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.956819772720337, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.880436897277832, "num_tokens": 470715165.0, "step": 12398 }, { "epoch": 1.577280244243735, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3776586055755615, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8763879537582397, "num_tokens": 470748950.0, "step": 12399 }, { "epoch": 1.5774074545223256, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.067770004272461, "learning_rate": 1e-05, "loss": 0.3941, "mean_token_accuracy": 0.889346718788147, "num_tokens": 470783856.0, "step": 12400 }, { "epoch": 1.5775346648009159, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4202864170074463, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8605226278305054, "num_tokens": 470823103.0, "step": 12401 }, { "epoch": 1.5776618750795064, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1553730964660645, "learning_rate": 1e-05, "loss": 0.5222, "mean_token_accuracy": 0.8528334498405457, "num_tokens": 470865582.0, "step": 12402 }, { "epoch": 1.577789085358097, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5686280727386475, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8612089157104492, "num_tokens": 470901710.0, "step": 12403 }, { "epoch": 1.5779162956366874, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.069511413574219, "learning_rate": 1e-05, "loss": 0.514, "mean_token_accuracy": 0.8529053926467896, "num_tokens": 470944850.0, "step": 12404 }, { "epoch": 1.578043505915278, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6725351810455322, "learning_rate": 1e-05, "loss": 0.5028, "mean_token_accuracy": 0.861695408821106, "num_tokens": 470985092.0, "step": 12405 }, { "epoch": 1.5781707161938685, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.857598304748535, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8779129385948181, "num_tokens": 471026018.0, "step": 12406 }, { "epoch": 1.5782979264724588, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6648805141448975, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8670028448104858, "num_tokens": 471061365.0, "step": 12407 }, { "epoch": 1.5784251367510493, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.643322706222534, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8749818801879883, "num_tokens": 471098142.0, "step": 12408 }, { "epoch": 1.5785523470296399, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6119143962860107, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.8556668758392334, "num_tokens": 471138001.0, "step": 12409 }, { "epoch": 1.5786795573082304, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.494987726211548, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8759143948554993, "num_tokens": 471170392.0, "step": 12410 }, { "epoch": 1.578806767586821, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.547950506210327, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8633968234062195, "num_tokens": 471215704.0, "step": 12411 }, { "epoch": 1.5789339778654115, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3878750801086426, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8854629993438721, "num_tokens": 471252084.0, "step": 12412 }, { "epoch": 1.579061188144002, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.534207344055176, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8643184304237366, "num_tokens": 471287096.0, "step": 12413 }, { "epoch": 1.5791883984225925, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.308119773864746, "learning_rate": 1e-05, "loss": 0.3926, "mean_token_accuracy": 0.8862852454185486, "num_tokens": 471321732.0, "step": 12414 }, { "epoch": 1.579315608701183, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.56571102142334, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.870032548904419, "num_tokens": 471353680.0, "step": 12415 }, { "epoch": 1.5794428189797736, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.2300219535827637, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8568651080131531, "num_tokens": 471395286.0, "step": 12416 }, { "epoch": 1.579570029258364, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6347901821136475, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8817476034164429, "num_tokens": 471433708.0, "step": 12417 }, { "epoch": 1.5796972395369546, "ewc_loss": 0.06982421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.1513562202453613, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8707846403121948, "num_tokens": 471472655.0, "step": 12418 }, { "epoch": 1.5798244498155452, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7945404052734375, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8694975972175598, "num_tokens": 471507013.0, "step": 12419 }, { "epoch": 1.5799516600941357, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1068809032440186, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8731920123100281, "num_tokens": 471539951.0, "step": 12420 }, { "epoch": 1.5800788703727262, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 4.026157379150391, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8670157194137573, "num_tokens": 471580505.0, "step": 12421 }, { "epoch": 1.5802060806513167, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.175182819366455, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8809986114501953, "num_tokens": 471610408.0, "step": 12422 }, { "epoch": 1.5803332909299073, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.250647068023682, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8673635721206665, "num_tokens": 471648529.0, "step": 12423 }, { "epoch": 1.5804605012084978, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.431129217147827, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8719213008880615, "num_tokens": 471682182.0, "step": 12424 }, { "epoch": 1.5805877114870883, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.254559516906738, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8741033673286438, "num_tokens": 471719180.0, "step": 12425 }, { "epoch": 1.5807149217656786, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.66194224357605, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8672441244125366, "num_tokens": 471759965.0, "step": 12426 }, { "epoch": 1.5808421320442692, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.058586120605469, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8820055723190308, "num_tokens": 471795405.0, "step": 12427 }, { "epoch": 1.5809693423228597, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.711766242980957, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8771102428436279, "num_tokens": 471837957.0, "step": 12428 }, { "epoch": 1.5810965526014502, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.83418869972229, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8700944185256958, "num_tokens": 471875422.0, "step": 12429 }, { "epoch": 1.5812237628800407, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6464180946350098, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8771800994873047, "num_tokens": 471913752.0, "step": 12430 }, { "epoch": 1.5813509731586313, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6616156101226807, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.863900899887085, "num_tokens": 471952362.0, "step": 12431 }, { "epoch": 1.5814781834372216, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.563530206680298, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8790542483329773, "num_tokens": 471990798.0, "step": 12432 }, { "epoch": 1.581605393715812, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.546645164489746, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8703604936599731, "num_tokens": 472027398.0, "step": 12433 }, { "epoch": 1.5817326039944026, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.485018730163574, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8751972317695618, "num_tokens": 472071725.0, "step": 12434 }, { "epoch": 1.5818598142729932, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4868340492248535, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8818811774253845, "num_tokens": 472106913.0, "step": 12435 }, { "epoch": 1.5819870245515837, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.442570209503174, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.870293378829956, "num_tokens": 472138136.0, "step": 12436 }, { "epoch": 1.5821142348301742, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.453428268432617, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.866572380065918, "num_tokens": 472174884.0, "step": 12437 }, { "epoch": 1.5822414451087647, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.371692419052124, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8673454523086548, "num_tokens": 472215862.0, "step": 12438 }, { "epoch": 1.5823686553873553, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4414796829223633, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8779966831207275, "num_tokens": 472255770.0, "step": 12439 }, { "epoch": 1.5824958656659458, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.3262553215026855, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.862725555896759, "num_tokens": 472290815.0, "step": 12440 }, { "epoch": 1.5826230759445363, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.477435350418091, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8620702028274536, "num_tokens": 472328180.0, "step": 12441 }, { "epoch": 1.5827502862231269, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.265392303466797, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8618029952049255, "num_tokens": 472366519.0, "step": 12442 }, { "epoch": 1.5828774965017174, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5317816734313965, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8554064035415649, "num_tokens": 472405270.0, "step": 12443 }, { "epoch": 1.583004706780308, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.193324327468872, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8766137361526489, "num_tokens": 472447572.0, "step": 12444 }, { "epoch": 1.5831319170588984, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.64536190032959, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.874229907989502, "num_tokens": 472487869.0, "step": 12445 }, { "epoch": 1.583259127337489, "ewc_loss": 0.0693359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.363059997558594e-05, "grad_norm": 3.1150999069213867, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8730614185333252, "num_tokens": 472521856.0, "step": 12446 }, { "epoch": 1.5833863376160795, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8419697284698486, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8737162947654724, "num_tokens": 472566550.0, "step": 12447 }, { "epoch": 1.58351354789467, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.0725860595703125, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.868800163269043, "num_tokens": 472605054.0, "step": 12448 }, { "epoch": 1.5836407581732606, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.1037726402282715, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8725132346153259, "num_tokens": 472649938.0, "step": 12449 }, { "epoch": 1.5837679684518509, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.182529926300049, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8744714260101318, "num_tokens": 472688399.0, "step": 12450 }, { "epoch": 1.5838951787304414, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.346505165100098, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8651742935180664, "num_tokens": 472726383.0, "step": 12451 }, { "epoch": 1.584022389009032, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.5311357975006104, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8660019040107727, "num_tokens": 472759692.0, "step": 12452 }, { "epoch": 1.5841495992876224, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.279496669769287, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8771758079528809, "num_tokens": 472799679.0, "step": 12453 }, { "epoch": 1.584276809566213, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.735992193222046, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.858249306678772, "num_tokens": 472842043.0, "step": 12454 }, { "epoch": 1.5844040198448035, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.067785263061523, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8660078048706055, "num_tokens": 472877594.0, "step": 12455 }, { "epoch": 1.5845312301233938, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.726043939590454, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.873438835144043, "num_tokens": 472914052.0, "step": 12456 }, { "epoch": 1.5846584404019843, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.860713243484497, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8727133274078369, "num_tokens": 472953359.0, "step": 12457 }, { "epoch": 1.5847856506805749, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.622696876525879, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8672785758972168, "num_tokens": 472993898.0, "step": 12458 }, { "epoch": 1.5849128609591654, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.721621036529541, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8744466304779053, "num_tokens": 473032026.0, "step": 12459 }, { "epoch": 1.585040071237756, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5077288150787354, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8776170015335083, "num_tokens": 473075282.0, "step": 12460 }, { "epoch": 1.5851672815163464, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6500866413116455, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8755336999893188, "num_tokens": 473112834.0, "step": 12461 }, { "epoch": 1.585294491794937, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.420102834701538, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8691744804382324, "num_tokens": 473150331.0, "step": 12462 }, { "epoch": 1.5854217020735275, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.636735439300537, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8775135278701782, "num_tokens": 473185175.0, "step": 12463 }, { "epoch": 1.585548912352118, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3284525871276855, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8626285791397095, "num_tokens": 473223546.0, "step": 12464 }, { "epoch": 1.5856761226307086, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.687974214553833, "learning_rate": 1e-05, "loss": 0.513, "mean_token_accuracy": 0.8568781614303589, "num_tokens": 473260098.0, "step": 12465 }, { "epoch": 1.585803332909299, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.251502752304077, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.87339186668396, "num_tokens": 473299283.0, "step": 12466 }, { "epoch": 1.5859305431878896, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.796994924545288, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8663182854652405, "num_tokens": 473332511.0, "step": 12467 }, { "epoch": 1.5860577534664801, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3532683849334717, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8758758306503296, "num_tokens": 473367259.0, "step": 12468 }, { "epoch": 1.5861849637450707, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7901718616485596, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8697444200515747, "num_tokens": 473412421.0, "step": 12469 }, { "epoch": 1.5863121740236612, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.225174903869629, "learning_rate": 1e-05, "loss": 0.5173, "mean_token_accuracy": 0.8500069975852966, "num_tokens": 473450444.0, "step": 12470 }, { "epoch": 1.5864393843022517, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.883713483810425, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8695681095123291, "num_tokens": 473485093.0, "step": 12471 }, { "epoch": 1.5865665945808423, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.256157159805298, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.883263111114502, "num_tokens": 473522461.0, "step": 12472 }, { "epoch": 1.5866938048594328, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.977327585220337, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8661404252052307, "num_tokens": 473559860.0, "step": 12473 }, { "epoch": 1.5868210151380233, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3633360862731934, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.879794716835022, "num_tokens": 473590144.0, "step": 12474 }, { "epoch": 1.5869482254166136, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9617886543273926, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8759127855300903, "num_tokens": 473631384.0, "step": 12475 }, { "epoch": 1.5870754356952042, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.4732506275177, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8699643611907959, "num_tokens": 473667396.0, "step": 12476 }, { "epoch": 1.5872026459737947, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.842057466506958, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8734732866287231, "num_tokens": 473699856.0, "step": 12477 }, { "epoch": 1.5873298562523852, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5406980514526367, "learning_rate": 1e-05, "loss": 0.5133, "mean_token_accuracy": 0.8512628674507141, "num_tokens": 473731164.0, "step": 12478 }, { "epoch": 1.5874570665309757, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.654616355895996, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8683731555938721, "num_tokens": 473767005.0, "step": 12479 }, { "epoch": 1.5875842768095663, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5279324054718018, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8692442774772644, "num_tokens": 473805753.0, "step": 12480 }, { "epoch": 1.5877114870881566, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4788782596588135, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8701893091201782, "num_tokens": 473845375.0, "step": 12481 }, { "epoch": 1.587838697366747, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5278663635253906, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.86069655418396, "num_tokens": 473889184.0, "step": 12482 }, { "epoch": 1.5879659076453376, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3509652614593506, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.879016637802124, "num_tokens": 473927917.0, "step": 12483 }, { "epoch": 1.5880931179239282, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.553072214126587, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8828975558280945, "num_tokens": 473964670.0, "step": 12484 }, { "epoch": 1.5882203282025187, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2586395740509033, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8862755298614502, "num_tokens": 473998455.0, "step": 12485 }, { "epoch": 1.5883475384811092, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6043171882629395, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8707530498504639, "num_tokens": 474041623.0, "step": 12486 }, { "epoch": 1.5884747487596997, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1829819679260254, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8691805601119995, "num_tokens": 474083294.0, "step": 12487 }, { "epoch": 1.5886019590382903, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7197251319885254, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8710888624191284, "num_tokens": 474117737.0, "step": 12488 }, { "epoch": 1.5887291693168808, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.128096342086792, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8686373233795166, "num_tokens": 474154467.0, "step": 12489 }, { "epoch": 1.5888563795954713, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8988194465637207, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8677086234092712, "num_tokens": 474191845.0, "step": 12490 }, { "epoch": 1.5889835898740619, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1301236152648926, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8755357265472412, "num_tokens": 474227787.0, "step": 12491 }, { "epoch": 1.5891108001526524, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.100651741027832, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8675123453140259, "num_tokens": 474271890.0, "step": 12492 }, { "epoch": 1.589238010431243, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2629780769348145, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.857433557510376, "num_tokens": 474314493.0, "step": 12493 }, { "epoch": 1.5893652207098334, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2346625328063965, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8800802230834961, "num_tokens": 474349348.0, "step": 12494 }, { "epoch": 1.589492430988424, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5226101875305176, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.864097535610199, "num_tokens": 474384145.0, "step": 12495 }, { "epoch": 1.5896196412670145, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.138251304626465, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8809759616851807, "num_tokens": 474419064.0, "step": 12496 }, { "epoch": 1.589746851545605, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6393308639526367, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.868075430393219, "num_tokens": 474458716.0, "step": 12497 }, { "epoch": 1.5898740618241956, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9558157920837402, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8760457038879395, "num_tokens": 474497672.0, "step": 12498 }, { "epoch": 1.5900012721027859, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6413872241973877, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8752930164337158, "num_tokens": 474530442.0, "step": 12499 }, { "epoch": 1.5901284823813764, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7671902179718018, "learning_rate": 1e-05, "loss": 0.5109, "mean_token_accuracy": 0.8537550568580627, "num_tokens": 474573103.0, "step": 12500 }, { "epoch": 1.590255692659967, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5684497356414795, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8782703876495361, "num_tokens": 474610687.0, "step": 12501 }, { "epoch": 1.5903829029385574, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6280243396759033, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8645995259284973, "num_tokens": 474650893.0, "step": 12502 }, { "epoch": 1.590510113217148, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4973714351654053, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8700029253959656, "num_tokens": 474690808.0, "step": 12503 }, { "epoch": 1.5906373234957385, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.544318437576294, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8808293342590332, "num_tokens": 474726814.0, "step": 12504 }, { "epoch": 1.5907645337743288, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.430220127105713, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8680538535118103, "num_tokens": 474765170.0, "step": 12505 }, { "epoch": 1.5908917440529193, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5205190181732178, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.868049144744873, "num_tokens": 474808466.0, "step": 12506 }, { "epoch": 1.5910189543315099, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.356873035430908, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8772158622741699, "num_tokens": 474847334.0, "step": 12507 }, { "epoch": 1.5911461646101004, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5270750522613525, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8708006739616394, "num_tokens": 474885937.0, "step": 12508 }, { "epoch": 1.591273374888691, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.292635202407837, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8746800422668457, "num_tokens": 474921089.0, "step": 12509 }, { "epoch": 1.5914005851672814, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5783164501190186, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8770301342010498, "num_tokens": 474961660.0, "step": 12510 }, { "epoch": 1.591527795445872, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.22257924079895, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.87374347448349, "num_tokens": 474997815.0, "step": 12511 }, { "epoch": 1.5916550057244625, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.685621976852417, "learning_rate": 1e-05, "loss": 0.5241, "mean_token_accuracy": 0.8445397615432739, "num_tokens": 475038334.0, "step": 12512 }, { "epoch": 1.591782216003053, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1568284034729004, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8730857372283936, "num_tokens": 475073522.0, "step": 12513 }, { "epoch": 1.5919094262816436, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8449783325195312, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.877348780632019, "num_tokens": 475111811.0, "step": 12514 }, { "epoch": 1.592036636560234, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1345062255859375, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8875322341918945, "num_tokens": 475146496.0, "step": 12515 }, { "epoch": 1.5921638468388246, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.047436237335205, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8705955743789673, "num_tokens": 475187947.0, "step": 12516 }, { "epoch": 1.5922910571174151, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.243715763092041, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8723059892654419, "num_tokens": 475225420.0, "step": 12517 }, { "epoch": 1.5924182673960057, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.185225486755371, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8638013005256653, "num_tokens": 475263913.0, "step": 12518 }, { "epoch": 1.5925454776745962, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.475215196609497, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8825104236602783, "num_tokens": 475300446.0, "step": 12519 }, { "epoch": 1.5926726879531867, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.119090557098389, "learning_rate": 1e-05, "loss": 0.5173, "mean_token_accuracy": 0.8599219918251038, "num_tokens": 475338728.0, "step": 12520 }, { "epoch": 1.5927998982317773, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.634739398956299, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8707181215286255, "num_tokens": 475369453.0, "step": 12521 }, { "epoch": 1.5929271085103678, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.897801160812378, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8837082386016846, "num_tokens": 475407994.0, "step": 12522 }, { "epoch": 1.5930543187889583, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.625153064727783, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.866435706615448, "num_tokens": 475445480.0, "step": 12523 }, { "epoch": 1.5931815290675486, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7058639526367188, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8582327365875244, "num_tokens": 475483971.0, "step": 12524 }, { "epoch": 1.5933087393461391, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.577565908432007, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8740183115005493, "num_tokens": 475521332.0, "step": 12525 }, { "epoch": 1.5934359496247297, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5407893657684326, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8661437034606934, "num_tokens": 475556085.0, "step": 12526 }, { "epoch": 1.5935631599033202, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.540584087371826, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8667435050010681, "num_tokens": 475590966.0, "step": 12527 }, { "epoch": 1.5936903701819107, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4365451335906982, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8637211322784424, "num_tokens": 475627430.0, "step": 12528 }, { "epoch": 1.5938175804605013, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4987895488739014, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8795250654220581, "num_tokens": 475662346.0, "step": 12529 }, { "epoch": 1.5939447907390916, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3676719665527344, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8621740937232971, "num_tokens": 475702312.0, "step": 12530 }, { "epoch": 1.594072001017682, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.479743719100952, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8707383871078491, "num_tokens": 475743997.0, "step": 12531 }, { "epoch": 1.5941992112962726, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3174285888671875, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8593446016311646, "num_tokens": 475783369.0, "step": 12532 }, { "epoch": 1.5943264215748632, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.497393846511841, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8850306272506714, "num_tokens": 475815225.0, "step": 12533 }, { "epoch": 1.5944536318534537, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.273953676223755, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8654308915138245, "num_tokens": 475847917.0, "step": 12534 }, { "epoch": 1.5945808421320442, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5328879356384277, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8758877515792847, "num_tokens": 475889308.0, "step": 12535 }, { "epoch": 1.5947080524106347, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2117202281951904, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8799567222595215, "num_tokens": 475930603.0, "step": 12536 }, { "epoch": 1.5948352626892253, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6191647052764893, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8783047199249268, "num_tokens": 475968288.0, "step": 12537 }, { "epoch": 1.5949624729678158, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.15274715423584, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8758630752563477, "num_tokens": 476002967.0, "step": 12538 }, { "epoch": 1.5950896832464063, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.761242628097534, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8725156784057617, "num_tokens": 476037958.0, "step": 12539 }, { "epoch": 1.5952168935249968, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1037473678588867, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.85123610496521, "num_tokens": 476072036.0, "step": 12540 }, { "epoch": 1.5953441038035874, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.968623638153076, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8753675222396851, "num_tokens": 476115665.0, "step": 12541 }, { "epoch": 1.595471314082178, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1069774627685547, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8659806847572327, "num_tokens": 476155380.0, "step": 12542 }, { "epoch": 1.5955985243607684, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.227148056030273, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8655997514724731, "num_tokens": 476194833.0, "step": 12543 }, { "epoch": 1.595725734639359, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3330795764923096, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8744956254959106, "num_tokens": 476233617.0, "step": 12544 }, { "epoch": 1.5958529449179495, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.346386432647705, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8672397136688232, "num_tokens": 476270312.0, "step": 12545 }, { "epoch": 1.59598015519654, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6379852294921875, "learning_rate": 1e-05, "loss": 0.3809, "mean_token_accuracy": 0.896275520324707, "num_tokens": 476304466.0, "step": 12546 }, { "epoch": 1.5961073654751305, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.184914588928223, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8737572431564331, "num_tokens": 476345136.0, "step": 12547 }, { "epoch": 1.5962345757537209, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7507526874542236, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8762657046318054, "num_tokens": 476381546.0, "step": 12548 }, { "epoch": 1.5963617860323114, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9717020988464355, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8719915747642517, "num_tokens": 476413990.0, "step": 12549 }, { "epoch": 1.596488996310902, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6843411922454834, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8833814859390259, "num_tokens": 476451487.0, "step": 12550 }, { "epoch": 1.5966162065894924, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.789473295211792, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8580007553100586, "num_tokens": 476493071.0, "step": 12551 }, { "epoch": 1.596743416868083, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5882210731506348, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8783358335494995, "num_tokens": 476531390.0, "step": 12552 }, { "epoch": 1.5968706271466735, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.664085626602173, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8682283163070679, "num_tokens": 476572491.0, "step": 12553 }, { "epoch": 1.5969978374252638, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.490365505218506, "learning_rate": 1e-05, "loss": 0.4794, "mean_token_accuracy": 0.8636694550514221, "num_tokens": 476613897.0, "step": 12554 }, { "epoch": 1.5971250477038543, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6034793853759766, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8646576404571533, "num_tokens": 476650773.0, "step": 12555 }, { "epoch": 1.5972522579824449, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.410552740097046, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8720948696136475, "num_tokens": 476685546.0, "step": 12556 }, { "epoch": 1.5973794682610354, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5988330841064453, "learning_rate": 1e-05, "loss": 0.5125, "mean_token_accuracy": 0.8545140027999878, "num_tokens": 476720569.0, "step": 12557 }, { "epoch": 1.597506678539626, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.315716505050659, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8746331930160522, "num_tokens": 476760162.0, "step": 12558 }, { "epoch": 1.5976338888182164, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6417629718780518, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8700345754623413, "num_tokens": 476793575.0, "step": 12559 }, { "epoch": 1.597761099096807, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2468810081481934, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8739069104194641, "num_tokens": 476826714.0, "step": 12560 }, { "epoch": 1.5978883093753975, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7327523231506348, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.8854553699493408, "num_tokens": 476866715.0, "step": 12561 }, { "epoch": 1.598015519653988, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1783952713012695, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8662467002868652, "num_tokens": 476906036.0, "step": 12562 }, { "epoch": 1.5981427299325786, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9008891582489014, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8703641295433044, "num_tokens": 476945005.0, "step": 12563 }, { "epoch": 1.598269940211169, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1884210109710693, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8763174414634705, "num_tokens": 476989491.0, "step": 12564 }, { "epoch": 1.5983971504897596, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.078148365020752, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8707019090652466, "num_tokens": 477026062.0, "step": 12565 }, { "epoch": 1.5985243607683501, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.334470510482788, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8682971000671387, "num_tokens": 477060045.0, "step": 12566 }, { "epoch": 1.5986515710469407, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.12116813659668, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8838446140289307, "num_tokens": 477102343.0, "step": 12567 }, { "epoch": 1.5987787813255312, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5251803398132324, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8789056539535522, "num_tokens": 477136239.0, "step": 12568 }, { "epoch": 1.5989059916041217, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9878995418548584, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8692094087600708, "num_tokens": 477178291.0, "step": 12569 }, { "epoch": 1.5990332018827123, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6213059425354004, "learning_rate": 1e-05, "loss": 0.5079, "mean_token_accuracy": 0.8533280491828918, "num_tokens": 477213993.0, "step": 12570 }, { "epoch": 1.5991604121613028, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7692673206329346, "learning_rate": 1e-05, "loss": 0.4924, "mean_token_accuracy": 0.8607473373413086, "num_tokens": 477258384.0, "step": 12571 }, { "epoch": 1.5992876224398933, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6003122329711914, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8709040880203247, "num_tokens": 477297386.0, "step": 12572 }, { "epoch": 1.5994148327184836, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.561852216720581, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8804069757461548, "num_tokens": 477334214.0, "step": 12573 }, { "epoch": 1.5995420429970741, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5876476764678955, "learning_rate": 1e-05, "loss": 0.4915, "mean_token_accuracy": 0.8589795827865601, "num_tokens": 477374777.0, "step": 12574 }, { "epoch": 1.5996692532756647, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4085206985473633, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.870061993598938, "num_tokens": 477415206.0, "step": 12575 }, { "epoch": 1.5997964635542552, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.8668603897094727, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8691091537475586, "num_tokens": 477453864.0, "step": 12576 }, { "epoch": 1.5999236738328457, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2452614307403564, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8552234768867493, "num_tokens": 477489236.0, "step": 12577 }, { "epoch": 1.6000508841114363, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8083131313323975, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8827968835830688, "num_tokens": 477524459.0, "step": 12578 }, { "epoch": 1.6001780943900266, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1570796966552734, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8671553730964661, "num_tokens": 477560045.0, "step": 12579 }, { "epoch": 1.600305304668617, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9880893230438232, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8584004640579224, "num_tokens": 477596438.0, "step": 12580 }, { "epoch": 1.6004325149472076, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.214700698852539, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8715278506278992, "num_tokens": 477634370.0, "step": 12581 }, { "epoch": 1.6005597252257981, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.136987686157227, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8642363548278809, "num_tokens": 477669798.0, "step": 12582 }, { "epoch": 1.6006869355043887, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3747167587280273, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8691024780273438, "num_tokens": 477714127.0, "step": 12583 }, { "epoch": 1.6008141457829792, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.164063930511475, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.875922441482544, "num_tokens": 477753176.0, "step": 12584 }, { "epoch": 1.6009413560615697, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.562222957611084, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8736830949783325, "num_tokens": 477789372.0, "step": 12585 }, { "epoch": 1.6010685663401603, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.043918132781982, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.865532636642456, "num_tokens": 477820891.0, "step": 12586 }, { "epoch": 1.6011957766187508, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.639885663986206, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8692847490310669, "num_tokens": 477853952.0, "step": 12587 }, { "epoch": 1.6013229868973413, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.850116491317749, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8857260942459106, "num_tokens": 477888692.0, "step": 12588 }, { "epoch": 1.6014501971759318, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5738887786865234, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8831474781036377, "num_tokens": 477923769.0, "step": 12589 }, { "epoch": 1.6015774074545224, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7171764373779297, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8873953819274902, "num_tokens": 477959997.0, "step": 12590 }, { "epoch": 1.601704617733113, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5238051414489746, "learning_rate": 1e-05, "loss": 0.4939, "mean_token_accuracy": 0.854928195476532, "num_tokens": 477996211.0, "step": 12591 }, { "epoch": 1.6018318280117034, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.634005069732666, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8727960586547852, "num_tokens": 478029817.0, "step": 12592 }, { "epoch": 1.601959038290294, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.427283525466919, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8800034523010254, "num_tokens": 478071216.0, "step": 12593 }, { "epoch": 1.6020862485688845, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5965774059295654, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.876918613910675, "num_tokens": 478112581.0, "step": 12594 }, { "epoch": 1.602213458847475, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.354849100112915, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8588049411773682, "num_tokens": 478148887.0, "step": 12595 }, { "epoch": 1.6023406691260655, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6268973350524902, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8602318167686462, "num_tokens": 478187580.0, "step": 12596 }, { "epoch": 1.6024678794046558, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.280174732208252, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8741476535797119, "num_tokens": 478222679.0, "step": 12597 }, { "epoch": 1.6025950896832464, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6931569576263428, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8773441314697266, "num_tokens": 478259464.0, "step": 12598 }, { "epoch": 1.602722299961837, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.222196578979492, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.884922444820404, "num_tokens": 478293790.0, "step": 12599 }, { "epoch": 1.6028495102404274, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8146817684173584, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8733776807785034, "num_tokens": 478329311.0, "step": 12600 }, { "epoch": 1.602976720519018, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.219982862472534, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8725348711013794, "num_tokens": 478362818.0, "step": 12601 }, { "epoch": 1.6031039307976085, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.927661180496216, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8679207563400269, "num_tokens": 478402783.0, "step": 12602 }, { "epoch": 1.6032311410761988, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.27718186378479, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.859207034111023, "num_tokens": 478438141.0, "step": 12603 }, { "epoch": 1.6033583513547893, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9997715950012207, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8782833814620972, "num_tokens": 478473176.0, "step": 12604 }, { "epoch": 1.6034855616333799, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3925940990448, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8752744197845459, "num_tokens": 478511899.0, "step": 12605 }, { "epoch": 1.6036127719119704, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.951732873916626, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.885151207447052, "num_tokens": 478545018.0, "step": 12606 }, { "epoch": 1.603739982190561, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4979467391967773, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.864884614944458, "num_tokens": 478582384.0, "step": 12607 }, { "epoch": 1.6038671924691514, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.804529905319214, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8701225519180298, "num_tokens": 478621491.0, "step": 12608 }, { "epoch": 1.603994402747742, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.522745132446289, "learning_rate": 1e-05, "loss": 0.3915, "mean_token_accuracy": 0.8925449848175049, "num_tokens": 478662018.0, "step": 12609 }, { "epoch": 1.6041216130263325, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.621170997619629, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.869726300239563, "num_tokens": 478702730.0, "step": 12610 }, { "epoch": 1.604248823304923, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.54213547706604, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8685865998268127, "num_tokens": 478736696.0, "step": 12611 }, { "epoch": 1.6043760335835135, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4638383388519287, "learning_rate": 1e-05, "loss": 0.5123, "mean_token_accuracy": 0.8513203859329224, "num_tokens": 478775873.0, "step": 12612 }, { "epoch": 1.604503243862104, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5456013679504395, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.878206729888916, "num_tokens": 478810571.0, "step": 12613 }, { "epoch": 1.6046304541406946, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3392128944396973, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8583597540855408, "num_tokens": 478845696.0, "step": 12614 }, { "epoch": 1.6047576644192851, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.56933331489563, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8661773204803467, "num_tokens": 478884561.0, "step": 12615 }, { "epoch": 1.6048848746978757, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.259962558746338, "learning_rate": 1e-05, "loss": 0.5276, "mean_token_accuracy": 0.848436176776886, "num_tokens": 478921989.0, "step": 12616 }, { "epoch": 1.6050120849764662, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.628274440765381, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8696804046630859, "num_tokens": 478958346.0, "step": 12617 }, { "epoch": 1.6051392952550567, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.18406343460083, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8727827072143555, "num_tokens": 479000141.0, "step": 12618 }, { "epoch": 1.6052665055336472, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.734668254852295, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8781758546829224, "num_tokens": 479038785.0, "step": 12619 }, { "epoch": 1.6053937158122378, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1440558433532715, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8623628616333008, "num_tokens": 479074679.0, "step": 12620 }, { "epoch": 1.6055209260908283, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9023375511169434, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8756058812141418, "num_tokens": 479106745.0, "step": 12621 }, { "epoch": 1.6056481363694186, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.139716148376465, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8727208375930786, "num_tokens": 479145159.0, "step": 12622 }, { "epoch": 1.6057753466480091, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.100985527038574, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8838711977005005, "num_tokens": 479181033.0, "step": 12623 }, { "epoch": 1.6059025569265997, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2731728553771973, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8658112287521362, "num_tokens": 479223954.0, "step": 12624 }, { "epoch": 1.6060297672051902, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.210021018981934, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8570064902305603, "num_tokens": 479268676.0, "step": 12625 }, { "epoch": 1.6061569774837807, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.501112222671509, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8800156116485596, "num_tokens": 479305203.0, "step": 12626 }, { "epoch": 1.6062841877623713, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.14354133605957, "learning_rate": 1e-05, "loss": 0.5455, "mean_token_accuracy": 0.8453432321548462, "num_tokens": 479347358.0, "step": 12627 }, { "epoch": 1.6064113980409616, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6320722103118896, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8619063496589661, "num_tokens": 479392973.0, "step": 12628 }, { "epoch": 1.606538608319552, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.967665672302246, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8576277494430542, "num_tokens": 479430832.0, "step": 12629 }, { "epoch": 1.6066658185981426, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.634234666824341, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8587815761566162, "num_tokens": 479471363.0, "step": 12630 }, { "epoch": 1.6067930288767331, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.771380662918091, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8760858774185181, "num_tokens": 479506041.0, "step": 12631 }, { "epoch": 1.6069202391553237, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5900871753692627, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8718863129615784, "num_tokens": 479537290.0, "step": 12632 }, { "epoch": 1.6070474494339142, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6232688426971436, "learning_rate": 1e-05, "loss": 0.5039, "mean_token_accuracy": 0.854836106300354, "num_tokens": 479573191.0, "step": 12633 }, { "epoch": 1.6071746597125047, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5198557376861572, "learning_rate": 1e-05, "loss": 0.4028, "mean_token_accuracy": 0.883837878704071, "num_tokens": 479606210.0, "step": 12634 }, { "epoch": 1.6073018699910953, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5078630447387695, "learning_rate": 1e-05, "loss": 0.4032, "mean_token_accuracy": 0.8854054808616638, "num_tokens": 479645198.0, "step": 12635 }, { "epoch": 1.6074290802696858, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4548685550689697, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8669382333755493, "num_tokens": 479686319.0, "step": 12636 }, { "epoch": 1.6075562905482763, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.461015224456787, "learning_rate": 1e-05, "loss": 0.3917, "mean_token_accuracy": 0.8923484086990356, "num_tokens": 479720170.0, "step": 12637 }, { "epoch": 1.6076835008268668, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4084839820861816, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8762410879135132, "num_tokens": 479757263.0, "step": 12638 }, { "epoch": 1.6078107111054574, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4451417922973633, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8685560822486877, "num_tokens": 479793050.0, "step": 12639 }, { "epoch": 1.607937921384048, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3641557693481445, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8680251836776733, "num_tokens": 479831705.0, "step": 12640 }, { "epoch": 1.6080651316626384, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.487142562866211, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8840258121490479, "num_tokens": 479863888.0, "step": 12641 }, { "epoch": 1.608192341941229, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2869327068328857, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8796928524971008, "num_tokens": 479902280.0, "step": 12642 }, { "epoch": 1.6083195522198195, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5259127616882324, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8739708662033081, "num_tokens": 479935649.0, "step": 12643 }, { "epoch": 1.60844676249841, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.221392869949341, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8681718111038208, "num_tokens": 479971834.0, "step": 12644 }, { "epoch": 1.6085739727770005, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.617971181869507, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8568598031997681, "num_tokens": 480011985.0, "step": 12645 }, { "epoch": 1.6087011830555908, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.153751850128174, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8582936525344849, "num_tokens": 480047833.0, "step": 12646 }, { "epoch": 1.6088283933341814, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.786985158920288, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8634595274925232, "num_tokens": 480080774.0, "step": 12647 }, { "epoch": 1.608955603612772, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.0840704441070557, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8725758790969849, "num_tokens": 480119114.0, "step": 12648 }, { "epoch": 1.6090828138913624, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.032222747802734, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8649190664291382, "num_tokens": 480156624.0, "step": 12649 }, { "epoch": 1.609210024169953, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.1488988399505615, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8721742630004883, "num_tokens": 480192573.0, "step": 12650 }, { "epoch": 1.6093372344485435, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.288306713104248, "learning_rate": 1e-05, "loss": 0.5151, "mean_token_accuracy": 0.8566291928291321, "num_tokens": 480232139.0, "step": 12651 }, { "epoch": 1.6094644447271338, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.4208009243011475, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8629046678543091, "num_tokens": 480269811.0, "step": 12652 }, { "epoch": 1.6095916550057243, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.319314956665039, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.86420077085495, "num_tokens": 480307785.0, "step": 12653 }, { "epoch": 1.6097188652843148, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6922812461853027, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.8687877655029297, "num_tokens": 480347028.0, "step": 12654 }, { "epoch": 1.6098460755629054, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.126079559326172, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8773301839828491, "num_tokens": 480383915.0, "step": 12655 }, { "epoch": 1.609973285841496, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.740424633026123, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8731048703193665, "num_tokens": 480424226.0, "step": 12656 }, { "epoch": 1.6101004961200864, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9137370586395264, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8812065720558167, "num_tokens": 480461920.0, "step": 12657 }, { "epoch": 1.610227706398677, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.664846658706665, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8785106539726257, "num_tokens": 480498509.0, "step": 12658 }, { "epoch": 1.6103549166772675, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7428882122039795, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8619692921638489, "num_tokens": 480541912.0, "step": 12659 }, { "epoch": 1.610482126955858, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5644047260284424, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8747727274894714, "num_tokens": 480576034.0, "step": 12660 }, { "epoch": 1.6106093372344485, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6619763374328613, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8633519411087036, "num_tokens": 480612265.0, "step": 12661 }, { "epoch": 1.610736547513039, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.472656488418579, "learning_rate": 1e-05, "loss": 0.5122, "mean_token_accuracy": 0.8515928387641907, "num_tokens": 480650667.0, "step": 12662 }, { "epoch": 1.6108637577916296, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5931997299194336, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8693596124649048, "num_tokens": 480688931.0, "step": 12663 }, { "epoch": 1.6109909680702201, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3879175186157227, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8632786273956299, "num_tokens": 480719525.0, "step": 12664 }, { "epoch": 1.6111181783488107, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.586597204208374, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8815361261367798, "num_tokens": 480755872.0, "step": 12665 }, { "epoch": 1.6112453886274012, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.297806739807129, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.874499499797821, "num_tokens": 480794994.0, "step": 12666 }, { "epoch": 1.6113725989059917, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.648455858230591, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.877403736114502, "num_tokens": 480828421.0, "step": 12667 }, { "epoch": 1.6114998091845822, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.219332456588745, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.882858157157898, "num_tokens": 480868571.0, "step": 12668 }, { "epoch": 1.6116270194631728, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.763857364654541, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8825973272323608, "num_tokens": 480910897.0, "step": 12669 }, { "epoch": 1.6117542297417633, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.18438458442688, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8704650402069092, "num_tokens": 480942243.0, "step": 12670 }, { "epoch": 1.6118814400203536, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.932206869125366, "learning_rate": 1e-05, "loss": 0.509, "mean_token_accuracy": 0.8592313528060913, "num_tokens": 480980510.0, "step": 12671 }, { "epoch": 1.6120086502989441, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2151737213134766, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.876164436340332, "num_tokens": 481016837.0, "step": 12672 }, { "epoch": 1.6121358605775347, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.078474521636963, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8696355819702148, "num_tokens": 481055862.0, "step": 12673 }, { "epoch": 1.6122630708561252, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3552374839782715, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8631344437599182, "num_tokens": 481094632.0, "step": 12674 }, { "epoch": 1.6123902811347157, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.102105617523193, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8705163598060608, "num_tokens": 481138468.0, "step": 12675 }, { "epoch": 1.6125174914133062, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5171289443969727, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.8663458824157715, "num_tokens": 481182983.0, "step": 12676 }, { "epoch": 1.6126447016918966, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9723730087280273, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8713382482528687, "num_tokens": 481225805.0, "step": 12677 }, { "epoch": 1.612771911970487, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6541974544525146, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.874640703201294, "num_tokens": 481260532.0, "step": 12678 }, { "epoch": 1.6128991222490776, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.742868185043335, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8665813207626343, "num_tokens": 481300494.0, "step": 12679 }, { "epoch": 1.6130263325276681, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.593496084213257, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.884663999080658, "num_tokens": 481335244.0, "step": 12680 }, { "epoch": 1.6131535428062587, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5489604473114014, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8762322664260864, "num_tokens": 481370894.0, "step": 12681 }, { "epoch": 1.6132807530848492, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.569580078125, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8664945363998413, "num_tokens": 481413570.0, "step": 12682 }, { "epoch": 1.6134079633634397, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.400738000869751, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.872981607913971, "num_tokens": 481461380.0, "step": 12683 }, { "epoch": 1.6135351736420303, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.57977557182312, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8696660995483398, "num_tokens": 481497738.0, "step": 12684 }, { "epoch": 1.6136623839206208, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2985892295837402, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8823239207267761, "num_tokens": 481541000.0, "step": 12685 }, { "epoch": 1.6137895941992113, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6027262210845947, "learning_rate": 1e-05, "loss": 0.3944, "mean_token_accuracy": 0.8876987099647522, "num_tokens": 481578012.0, "step": 12686 }, { "epoch": 1.6139168044778018, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2508866786956787, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8642710447311401, "num_tokens": 481612996.0, "step": 12687 }, { "epoch": 1.6140440147563924, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.68255352973938, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8730454444885254, "num_tokens": 481645570.0, "step": 12688 }, { "epoch": 1.614171225034983, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.164607524871826, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8786630034446716, "num_tokens": 481687873.0, "step": 12689 }, { "epoch": 1.6142984353135734, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.816256284713745, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8619052767753601, "num_tokens": 481728105.0, "step": 12690 }, { "epoch": 1.614425645592164, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1433539390563965, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.880073070526123, "num_tokens": 481763558.0, "step": 12691 }, { "epoch": 1.6145528558707545, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.0002055168151855, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8574272394180298, "num_tokens": 481801153.0, "step": 12692 }, { "epoch": 1.614680066149345, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.1992530822753906, "learning_rate": 1e-05, "loss": 0.3882, "mean_token_accuracy": 0.8890583515167236, "num_tokens": 481830523.0, "step": 12693 }, { "epoch": 1.6148072764279355, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.156958103179932, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8692989349365234, "num_tokens": 481870814.0, "step": 12694 }, { "epoch": 1.6149344867065258, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.379513740539551, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.873338520526886, "num_tokens": 481912279.0, "step": 12695 }, { "epoch": 1.6150616969851164, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.179409503936768, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8685706257820129, "num_tokens": 481957506.0, "step": 12696 }, { "epoch": 1.615188907263707, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.568385124206543, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.879594087600708, "num_tokens": 481996718.0, "step": 12697 }, { "epoch": 1.6153161175422974, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0484938621521, "learning_rate": 1e-05, "loss": 0.4726, "mean_token_accuracy": 0.8700030446052551, "num_tokens": 482035891.0, "step": 12698 }, { "epoch": 1.615443327820888, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6444616317749023, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8733619451522827, "num_tokens": 482070381.0, "step": 12699 }, { "epoch": 1.6155705380994785, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8454396724700928, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8751538395881653, "num_tokens": 482114484.0, "step": 12700 }, { "epoch": 1.6156977483780688, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6127495765686035, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8718076944351196, "num_tokens": 482147978.0, "step": 12701 }, { "epoch": 1.6158249586566593, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6738569736480713, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8773415088653564, "num_tokens": 482193017.0, "step": 12702 }, { "epoch": 1.6159521689352498, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.555637836456299, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8572624921798706, "num_tokens": 482230856.0, "step": 12703 }, { "epoch": 1.6160793792138404, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5460879802703857, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8820548057556152, "num_tokens": 482267540.0, "step": 12704 }, { "epoch": 1.616206589492431, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.492342710494995, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8819583058357239, "num_tokens": 482302256.0, "step": 12705 }, { "epoch": 1.6163337997710214, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4734559059143066, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8774052262306213, "num_tokens": 482340706.0, "step": 12706 }, { "epoch": 1.616461010049612, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.436738967895508, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8794033527374268, "num_tokens": 482377313.0, "step": 12707 }, { "epoch": 1.6165882203282025, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4357388019561768, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8841898441314697, "num_tokens": 482415510.0, "step": 12708 }, { "epoch": 1.616715430606793, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.387033700942993, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8710809946060181, "num_tokens": 482453568.0, "step": 12709 }, { "epoch": 1.6168426408853835, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4394376277923584, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8697891235351562, "num_tokens": 482487024.0, "step": 12710 }, { "epoch": 1.616969851163974, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3443098068237305, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.881309986114502, "num_tokens": 482524328.0, "step": 12711 }, { "epoch": 1.6170970614425646, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4641268253326416, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8805411458015442, "num_tokens": 482557042.0, "step": 12712 }, { "epoch": 1.6172242717211551, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3031787872314453, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8509031534194946, "num_tokens": 482594486.0, "step": 12713 }, { "epoch": 1.6173514819997457, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.500922441482544, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8716816902160645, "num_tokens": 482635335.0, "step": 12714 }, { "epoch": 1.6174786922783362, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.223602533340454, "learning_rate": 1e-05, "loss": 0.4083, "mean_token_accuracy": 0.8825688362121582, "num_tokens": 482678945.0, "step": 12715 }, { "epoch": 1.6176059025569267, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5897152423858643, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8728152513504028, "num_tokens": 482717864.0, "step": 12716 }, { "epoch": 1.6177331128355172, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.153038501739502, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8689004778862, "num_tokens": 482756361.0, "step": 12717 }, { "epoch": 1.6178603231141078, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7564938068389893, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8762670755386353, "num_tokens": 482794658.0, "step": 12718 }, { "epoch": 1.6179875333926983, "ewc_loss": 0.0703125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.095400333404541, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8683807849884033, "num_tokens": 482829426.0, "step": 12719 }, { "epoch": 1.6181147436712886, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9895925521850586, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8774410486221313, "num_tokens": 482867891.0, "step": 12720 }, { "epoch": 1.6182419539498791, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002593994140625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1059510707855225, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8782805800437927, "num_tokens": 482902307.0, "step": 12721 }, { "epoch": 1.6183691642284697, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.277480602264404, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8692177534103394, "num_tokens": 482939751.0, "step": 12722 }, { "epoch": 1.6184963745070602, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3560829162597656, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.864814281463623, "num_tokens": 482978899.0, "step": 12723 }, { "epoch": 1.6186235847856507, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.385764122009277, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.881126880645752, "num_tokens": 483011086.0, "step": 12724 }, { "epoch": 1.6187507950642412, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6770341396331787, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8868895769119263, "num_tokens": 483054235.0, "step": 12725 }, { "epoch": 1.6188780053428315, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.2207722663879395, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8756842613220215, "num_tokens": 483086843.0, "step": 12726 }, { "epoch": 1.619005215621422, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.804788589477539, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8750022053718567, "num_tokens": 483119198.0, "step": 12727 }, { "epoch": 1.6191324259000126, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.966050386428833, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8818717002868652, "num_tokens": 483160835.0, "step": 12728 }, { "epoch": 1.6192596361786031, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7154853343963623, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8654126524925232, "num_tokens": 483194906.0, "step": 12729 }, { "epoch": 1.6193868464571937, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.798556089401245, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8658768534660339, "num_tokens": 483231731.0, "step": 12730 }, { "epoch": 1.6195140567357842, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5849993228912354, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.87479567527771, "num_tokens": 483270058.0, "step": 12731 }, { "epoch": 1.6196412670143747, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6854774951934814, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8855456709861755, "num_tokens": 483303600.0, "step": 12732 }, { "epoch": 1.6197684772929652, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4832253456115723, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.875231921672821, "num_tokens": 483339933.0, "step": 12733 }, { "epoch": 1.6198956875715558, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.642910957336426, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.878623366355896, "num_tokens": 483377153.0, "step": 12734 }, { "epoch": 1.6200228978501463, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3771398067474365, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8847219347953796, "num_tokens": 483415089.0, "step": 12735 }, { "epoch": 1.6201501081287368, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.652194023132324, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8715119957923889, "num_tokens": 483454023.0, "step": 12736 }, { "epoch": 1.6202773184073274, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.280604362487793, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.872560441493988, "num_tokens": 483493572.0, "step": 12737 }, { "epoch": 1.6204045286859179, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7420003414154053, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8670676350593567, "num_tokens": 483529817.0, "step": 12738 }, { "epoch": 1.6205317389645084, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2064616680145264, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8668876886367798, "num_tokens": 483570905.0, "step": 12739 }, { "epoch": 1.620658949243099, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.879948377609253, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8682670593261719, "num_tokens": 483618177.0, "step": 12740 }, { "epoch": 1.6207861595216895, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2089929580688477, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8744269609451294, "num_tokens": 483651227.0, "step": 12741 }, { "epoch": 1.62091336980028, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.052281379699707, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8773583173751831, "num_tokens": 483685577.0, "step": 12742 }, { "epoch": 1.6210405800788705, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.329516887664795, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8595772981643677, "num_tokens": 483721842.0, "step": 12743 }, { "epoch": 1.6211677903574608, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.116450786590576, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8650413751602173, "num_tokens": 483758646.0, "step": 12744 }, { "epoch": 1.6212950006360514, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5189740657806396, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.8592075705528259, "num_tokens": 483789216.0, "step": 12745 }, { "epoch": 1.621422210914642, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9827401638031006, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8821396827697754, "num_tokens": 483832901.0, "step": 12746 }, { "epoch": 1.6215494211932324, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5922799110412598, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.8658437728881836, "num_tokens": 483874331.0, "step": 12747 }, { "epoch": 1.621676631471823, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7877819538116455, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8762708902359009, "num_tokens": 483911608.0, "step": 12748 }, { "epoch": 1.6218038417504135, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6043388843536377, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8724045753479004, "num_tokens": 483946419.0, "step": 12749 }, { "epoch": 1.6219310520290038, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5779268741607666, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8620954751968384, "num_tokens": 483989740.0, "step": 12750 }, { "epoch": 1.6220582623075943, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.587306261062622, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8651822805404663, "num_tokens": 484025672.0, "step": 12751 }, { "epoch": 1.6221854725861848, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4170820713043213, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8748723268508911, "num_tokens": 484060421.0, "step": 12752 }, { "epoch": 1.6223126828647754, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5753252506256104, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8739287853240967, "num_tokens": 484102238.0, "step": 12753 }, { "epoch": 1.622439893143366, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3165032863616943, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8768808841705322, "num_tokens": 484138314.0, "step": 12754 }, { "epoch": 1.6225671034219564, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6094729900360107, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8528555035591125, "num_tokens": 484175541.0, "step": 12755 }, { "epoch": 1.622694313700547, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2405154705047607, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8843610882759094, "num_tokens": 484211276.0, "step": 12756 }, { "epoch": 1.6228215239791375, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.693171977996826, "learning_rate": 1e-05, "loss": 0.52, "mean_token_accuracy": 0.8525862693786621, "num_tokens": 484250192.0, "step": 12757 }, { "epoch": 1.622948734257728, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.164808750152588, "learning_rate": 1e-05, "loss": 0.4125, "mean_token_accuracy": 0.8798266649246216, "num_tokens": 484287330.0, "step": 12758 }, { "epoch": 1.6230759445363185, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8407678604125977, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8677603602409363, "num_tokens": 484319937.0, "step": 12759 }, { "epoch": 1.623203154814909, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1450350284576416, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8668395280838013, "num_tokens": 484353305.0, "step": 12760 }, { "epoch": 1.6233303650934996, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.031924724578857, "learning_rate": 1e-05, "loss": 0.5058, "mean_token_accuracy": 0.8581943511962891, "num_tokens": 484391970.0, "step": 12761 }, { "epoch": 1.6234575753720901, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2077927589416504, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8786112666130066, "num_tokens": 484433448.0, "step": 12762 }, { "epoch": 1.6235847856506807, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.209113597869873, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8775007724761963, "num_tokens": 484469729.0, "step": 12763 }, { "epoch": 1.6237119959292712, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.432612419128418, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8588230609893799, "num_tokens": 484509474.0, "step": 12764 }, { "epoch": 1.6238392062078617, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.192682266235352, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8895670175552368, "num_tokens": 484548999.0, "step": 12765 }, { "epoch": 1.6239664164864522, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.638852119445801, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.8557464480400085, "num_tokens": 484580534.0, "step": 12766 }, { "epoch": 1.6240936267650428, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.019726753234863, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8864322900772095, "num_tokens": 484616183.0, "step": 12767 }, { "epoch": 1.6242208370436333, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6432723999023438, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8704766035079956, "num_tokens": 484658290.0, "step": 12768 }, { "epoch": 1.6243480473222236, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8349297046661377, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.882719874382019, "num_tokens": 484690226.0, "step": 12769 }, { "epoch": 1.6244752576008141, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5912842750549316, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8673883676528931, "num_tokens": 484731911.0, "step": 12770 }, { "epoch": 1.6246024678794047, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6852736473083496, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8733878135681152, "num_tokens": 484768648.0, "step": 12771 }, { "epoch": 1.6247296781579952, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.526521682739258, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.8632076382637024, "num_tokens": 484807266.0, "step": 12772 }, { "epoch": 1.6248568884365857, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.582432508468628, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8604148030281067, "num_tokens": 484850877.0, "step": 12773 }, { "epoch": 1.6249840987151762, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4416069984436035, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8760601282119751, "num_tokens": 484891223.0, "step": 12774 }, { "epoch": 1.6251113089937665, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.550285577774048, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8766570091247559, "num_tokens": 484922081.0, "step": 12775 }, { "epoch": 1.625238519272357, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.381641387939453, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8600108623504639, "num_tokens": 484959016.0, "step": 12776 }, { "epoch": 1.6253657295509476, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5437026023864746, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8711397647857666, "num_tokens": 484998493.0, "step": 12777 }, { "epoch": 1.6254929398295381, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.317157745361328, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8731859922409058, "num_tokens": 485035164.0, "step": 12778 }, { "epoch": 1.6256201501081287, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5714311599731445, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8836475014686584, "num_tokens": 485076347.0, "step": 12779 }, { "epoch": 1.6257473603867192, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2588627338409424, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.873957097530365, "num_tokens": 485107940.0, "step": 12780 }, { "epoch": 1.6258745706653097, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.648489475250244, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8694209456443787, "num_tokens": 485145440.0, "step": 12781 }, { "epoch": 1.6260017809439002, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.189700126647949, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8763273358345032, "num_tokens": 485182948.0, "step": 12782 }, { "epoch": 1.6261289912224908, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.78190541267395, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8811408877372742, "num_tokens": 485220397.0, "step": 12783 }, { "epoch": 1.6262562015010813, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.145280361175537, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8730239868164062, "num_tokens": 485265382.0, "step": 12784 }, { "epoch": 1.6263834117796718, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.970722198486328, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8709782361984253, "num_tokens": 485306558.0, "step": 12785 }, { "epoch": 1.6265106220582624, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2176201343536377, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8685442209243774, "num_tokens": 485341278.0, "step": 12786 }, { "epoch": 1.6266378323368529, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1105828285217285, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8774416446685791, "num_tokens": 485376803.0, "step": 12787 }, { "epoch": 1.6267650426154434, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3999037742614746, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8599494099617004, "num_tokens": 485411168.0, "step": 12788 }, { "epoch": 1.626892252894034, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.086959362030029, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.863553524017334, "num_tokens": 485456320.0, "step": 12789 }, { "epoch": 1.6270194631726245, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5495829582214355, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8636928796768188, "num_tokens": 485491894.0, "step": 12790 }, { "epoch": 1.627146673451215, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9378979206085205, "learning_rate": 1e-05, "loss": 0.5112, "mean_token_accuracy": 0.8586328029632568, "num_tokens": 485526232.0, "step": 12791 }, { "epoch": 1.6272738837298055, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.594780921936035, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.867985188961029, "num_tokens": 485564752.0, "step": 12792 }, { "epoch": 1.6274010940083958, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.730778455734253, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8780233860015869, "num_tokens": 485600023.0, "step": 12793 }, { "epoch": 1.6275283042869864, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5668888092041016, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8738493323326111, "num_tokens": 485639195.0, "step": 12794 }, { "epoch": 1.6276555145655769, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5604796409606934, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8797267079353333, "num_tokens": 485676619.0, "step": 12795 }, { "epoch": 1.6277827248441674, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5406527519226074, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8563064932823181, "num_tokens": 485713739.0, "step": 12796 }, { "epoch": 1.627909935122758, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.439500331878662, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8698631525039673, "num_tokens": 485750010.0, "step": 12797 }, { "epoch": 1.6280371454013485, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5139126777648926, "learning_rate": 1e-05, "loss": 0.4966, "mean_token_accuracy": 0.855608344078064, "num_tokens": 485790754.0, "step": 12798 }, { "epoch": 1.6281643556799388, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.367316484451294, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8758683204650879, "num_tokens": 485825472.0, "step": 12799 }, { "epoch": 1.6282915659585293, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.492929458618164, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8730819225311279, "num_tokens": 485866349.0, "step": 12800 }, { "epoch": 1.6284187762371198, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.304481267929077, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8734055757522583, "num_tokens": 485908323.0, "step": 12801 }, { "epoch": 1.6285459865157104, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5215699672698975, "learning_rate": 1e-05, "loss": 0.4167, "mean_token_accuracy": 0.8809085488319397, "num_tokens": 485948637.0, "step": 12802 }, { "epoch": 1.628673196794301, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.248136043548584, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.870756983757019, "num_tokens": 485988824.0, "step": 12803 }, { "epoch": 1.6288004070728914, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.582622528076172, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8752251267433167, "num_tokens": 486026364.0, "step": 12804 }, { "epoch": 1.628927617351482, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.17964768409729, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8676763772964478, "num_tokens": 486069578.0, "step": 12805 }, { "epoch": 1.6290548276300725, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.707773447036743, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.869621753692627, "num_tokens": 486105847.0, "step": 12806 }, { "epoch": 1.629182037908663, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1117804050445557, "learning_rate": 1e-05, "loss": 0.4042, "mean_token_accuracy": 0.8850156664848328, "num_tokens": 486146680.0, "step": 12807 }, { "epoch": 1.6293092481872535, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9023196697235107, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8800348043441772, "num_tokens": 486184461.0, "step": 12808 }, { "epoch": 1.629436458465844, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.113771915435791, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.8813613653182983, "num_tokens": 486214084.0, "step": 12809 }, { "epoch": 1.6295636687444346, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.154586315155029, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8718931674957275, "num_tokens": 486250932.0, "step": 12810 }, { "epoch": 1.6296908790230251, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.24995756149292, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8794646263122559, "num_tokens": 486287964.0, "step": 12811 }, { "epoch": 1.6298180893016156, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.332319259643555, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8693719506263733, "num_tokens": 486321312.0, "step": 12812 }, { "epoch": 1.6299452995802062, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.562373399734497, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8704398274421692, "num_tokens": 486358992.0, "step": 12813 }, { "epoch": 1.6300725098587967, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.230573654174805, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.8590178489685059, "num_tokens": 486399632.0, "step": 12814 }, { "epoch": 1.6301997201373872, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7173712253570557, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8722571134567261, "num_tokens": 486438129.0, "step": 12815 }, { "epoch": 1.6303269304159778, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.016294956207275, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8719935417175293, "num_tokens": 486476167.0, "step": 12816 }, { "epoch": 1.630454140694568, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.699045181274414, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8781362771987915, "num_tokens": 486512065.0, "step": 12817 }, { "epoch": 1.6305813509731586, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.823988676071167, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.8558214902877808, "num_tokens": 486555723.0, "step": 12818 }, { "epoch": 1.6307085612517491, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6125118732452393, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8592021465301514, "num_tokens": 486598598.0, "step": 12819 }, { "epoch": 1.6308357715303397, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.689196825027466, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.869158148765564, "num_tokens": 486633426.0, "step": 12820 }, { "epoch": 1.6309629818089302, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.51954984664917, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8753329515457153, "num_tokens": 486675584.0, "step": 12821 }, { "epoch": 1.6310901920875207, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6070730686187744, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8738284707069397, "num_tokens": 486722083.0, "step": 12822 }, { "epoch": 1.6312174023661112, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.42374324798584, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8808914422988892, "num_tokens": 486762658.0, "step": 12823 }, { "epoch": 1.6313446126447015, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.592592477798462, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8690481185913086, "num_tokens": 486799441.0, "step": 12824 }, { "epoch": 1.631471822923292, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3443377017974854, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8615886569023132, "num_tokens": 486836742.0, "step": 12825 }, { "epoch": 1.6315990332018826, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6298136711120605, "learning_rate": 1e-05, "loss": 0.526, "mean_token_accuracy": 0.8465007543563843, "num_tokens": 486873460.0, "step": 12826 }, { "epoch": 1.6317262434804731, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2556328773498535, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8724031448364258, "num_tokens": 486909411.0, "step": 12827 }, { "epoch": 1.6318534537590637, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.711578369140625, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8632640242576599, "num_tokens": 486948427.0, "step": 12828 }, { "epoch": 1.6319806640376542, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.205822467803955, "learning_rate": 1e-05, "loss": 0.5035, "mean_token_accuracy": 0.8538719415664673, "num_tokens": 486984594.0, "step": 12829 }, { "epoch": 1.6321078743162447, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.841515064239502, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8783971071243286, "num_tokens": 487028122.0, "step": 12830 }, { "epoch": 1.6322350845948352, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.18330717086792, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8612086176872253, "num_tokens": 487068731.0, "step": 12831 }, { "epoch": 1.6323622948734258, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.033026218414307, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8749389052391052, "num_tokens": 487103993.0, "step": 12832 }, { "epoch": 1.6324895051520163, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2836875915527344, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.870360255241394, "num_tokens": 487139206.0, "step": 12833 }, { "epoch": 1.6326167154306068, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.125838279724121, "learning_rate": 1e-05, "loss": 0.5607, "mean_token_accuracy": 0.8409661054611206, "num_tokens": 487181553.0, "step": 12834 }, { "epoch": 1.6327439257091974, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.4653871059417725, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8680155277252197, "num_tokens": 487218931.0, "step": 12835 }, { "epoch": 1.6328711359877879, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.05761194229126, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8656206130981445, "num_tokens": 487252829.0, "step": 12836 }, { "epoch": 1.6329983462663784, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5962419509887695, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8774695992469788, "num_tokens": 487292098.0, "step": 12837 }, { "epoch": 1.633125556544969, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.848923444747925, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.872747540473938, "num_tokens": 487330639.0, "step": 12838 }, { "epoch": 1.6332527668235595, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.606403112411499, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8784184455871582, "num_tokens": 487368499.0, "step": 12839 }, { "epoch": 1.63337997710215, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.644582748413086, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8798748254776001, "num_tokens": 487406626.0, "step": 12840 }, { "epoch": 1.6335071873807405, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5942866802215576, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8741270303726196, "num_tokens": 487441055.0, "step": 12841 }, { "epoch": 1.6336343976593308, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4559402465820312, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.8929434418678284, "num_tokens": 487475973.0, "step": 12842 }, { "epoch": 1.6337616079379214, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5944786071777344, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8763561248779297, "num_tokens": 487513626.0, "step": 12843 }, { "epoch": 1.6338888182165119, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.332965135574341, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8799161911010742, "num_tokens": 487550053.0, "step": 12844 }, { "epoch": 1.6340160284951024, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.608778715133667, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8795534372329712, "num_tokens": 487586999.0, "step": 12845 }, { "epoch": 1.634143238773693, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2579379081726074, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8821985721588135, "num_tokens": 487631433.0, "step": 12846 }, { "epoch": 1.6342704490522835, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.680769205093384, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8746602535247803, "num_tokens": 487663259.0, "step": 12847 }, { "epoch": 1.6343976593308738, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.199883222579956, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8647950291633606, "num_tokens": 487697616.0, "step": 12848 }, { "epoch": 1.6345248696094643, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.784695863723755, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8820096254348755, "num_tokens": 487735693.0, "step": 12849 }, { "epoch": 1.6346520798880548, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.152392864227295, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8741978406906128, "num_tokens": 487775519.0, "step": 12850 }, { "epoch": 1.6347792901666454, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.94643235206604, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8629497289657593, "num_tokens": 487814058.0, "step": 12851 }, { "epoch": 1.6349065004452359, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.1893422603607178, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8812549114227295, "num_tokens": 487850128.0, "step": 12852 }, { "epoch": 1.6350337107238264, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.105722427368164, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8856933116912842, "num_tokens": 487889465.0, "step": 12853 }, { "epoch": 1.635160921002417, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.33748722076416, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8671151399612427, "num_tokens": 487926264.0, "step": 12854 }, { "epoch": 1.6352881312810075, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.187473773956299, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.86959308385849, "num_tokens": 487959872.0, "step": 12855 }, { "epoch": 1.635415341559598, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.527578115463257, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8616217374801636, "num_tokens": 487997554.0, "step": 12856 }, { "epoch": 1.6355425518381885, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0807647705078125, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8860554695129395, "num_tokens": 488029172.0, "step": 12857 }, { "epoch": 1.635669762116779, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.628516435623169, "learning_rate": 1e-05, "loss": 0.5234, "mean_token_accuracy": 0.8528481721878052, "num_tokens": 488066651.0, "step": 12858 }, { "epoch": 1.6357969723953696, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9005022048950195, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8626144528388977, "num_tokens": 488102207.0, "step": 12859 }, { "epoch": 1.6359241826739601, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.59896183013916, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8822312355041504, "num_tokens": 488142502.0, "step": 12860 }, { "epoch": 1.6360513929525506, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.729015350341797, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.858345091342926, "num_tokens": 488182334.0, "step": 12861 }, { "epoch": 1.6361786032311412, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5459744930267334, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8759539127349854, "num_tokens": 488219434.0, "step": 12862 }, { "epoch": 1.6363058135097317, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6064250469207764, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8754528164863586, "num_tokens": 488256961.0, "step": 12863 }, { "epoch": 1.6364330237883222, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4881608486175537, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8649843335151672, "num_tokens": 488297069.0, "step": 12864 }, { "epoch": 1.6365602340669128, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.530687093734741, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.8561025261878967, "num_tokens": 488336436.0, "step": 12865 }, { "epoch": 1.636687444345503, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4203720092773438, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8669614791870117, "num_tokens": 488376145.0, "step": 12866 }, { "epoch": 1.6368146546240936, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.503826856613159, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8740752935409546, "num_tokens": 488418319.0, "step": 12867 }, { "epoch": 1.6369418649026841, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3569154739379883, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8726036548614502, "num_tokens": 488460377.0, "step": 12868 }, { "epoch": 1.6370690751812746, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5374906063079834, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8659877777099609, "num_tokens": 488491794.0, "step": 12869 }, { "epoch": 1.6371962854598652, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2917697429656982, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.86469566822052, "num_tokens": 488534106.0, "step": 12870 }, { "epoch": 1.6373234957384557, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5867927074432373, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.863783597946167, "num_tokens": 488568198.0, "step": 12871 }, { "epoch": 1.6374507060170462, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.224107503890991, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8670743703842163, "num_tokens": 488606112.0, "step": 12872 }, { "epoch": 1.6375779162956365, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.697343111038208, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8684654235839844, "num_tokens": 488639585.0, "step": 12873 }, { "epoch": 1.637705126574227, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1621968746185303, "learning_rate": 1e-05, "loss": 0.3949, "mean_token_accuracy": 0.8853804469108582, "num_tokens": 488670789.0, "step": 12874 }, { "epoch": 1.6378323368528176, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.844681739807129, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8711972236633301, "num_tokens": 488707837.0, "step": 12875 }, { "epoch": 1.6379595471314081, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.143296241760254, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8706768751144409, "num_tokens": 488747719.0, "step": 12876 }, { "epoch": 1.6380867574099987, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0481157302856445, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8661549091339111, "num_tokens": 488786493.0, "step": 12877 }, { "epoch": 1.6382139676885892, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.233815908432007, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8663297891616821, "num_tokens": 488826390.0, "step": 12878 }, { "epoch": 1.6383411779671797, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.215287685394287, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8682873249053955, "num_tokens": 488861458.0, "step": 12879 }, { "epoch": 1.6384683882457702, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4666531085968018, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8741289973258972, "num_tokens": 488902762.0, "step": 12880 }, { "epoch": 1.6385955985243608, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.156500339508057, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8835619688034058, "num_tokens": 488938139.0, "step": 12881 }, { "epoch": 1.6387228088029513, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.639435291290283, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8747448325157166, "num_tokens": 488976036.0, "step": 12882 }, { "epoch": 1.6388500190815418, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9634904861450195, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.86008220911026, "num_tokens": 489014333.0, "step": 12883 }, { "epoch": 1.6389772293601323, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.65169095993042, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.873617947101593, "num_tokens": 489050060.0, "step": 12884 }, { "epoch": 1.6391044396387229, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7523696422576904, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8709595203399658, "num_tokens": 489090053.0, "step": 12885 }, { "epoch": 1.6392316499173134, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5961015224456787, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8687548637390137, "num_tokens": 489129081.0, "step": 12886 }, { "epoch": 1.639358860195904, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5810980796813965, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8621096014976501, "num_tokens": 489174811.0, "step": 12887 }, { "epoch": 1.6394860704744945, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.544797897338867, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8704752922058105, "num_tokens": 489214694.0, "step": 12888 }, { "epoch": 1.639613280753085, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4741151332855225, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8759317994117737, "num_tokens": 489255299.0, "step": 12889 }, { "epoch": 1.6397404910316755, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4899489879608154, "learning_rate": 1e-05, "loss": 0.3933, "mean_token_accuracy": 0.8884017467498779, "num_tokens": 489292409.0, "step": 12890 }, { "epoch": 1.6398677013102658, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4226698875427246, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8620603680610657, "num_tokens": 489335354.0, "step": 12891 }, { "epoch": 1.6399949115888564, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.476470708847046, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8498566746711731, "num_tokens": 489369253.0, "step": 12892 }, { "epoch": 1.6401221218674469, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3700592517852783, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8725173473358154, "num_tokens": 489408453.0, "step": 12893 }, { "epoch": 1.6402493321460374, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.455489158630371, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8820871710777283, "num_tokens": 489441609.0, "step": 12894 }, { "epoch": 1.640376542424628, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3449103832244873, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8724814057350159, "num_tokens": 489477228.0, "step": 12895 }, { "epoch": 1.6405037527032185, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4517476558685303, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8580092191696167, "num_tokens": 489518622.0, "step": 12896 }, { "epoch": 1.6406309629818088, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.3097546100616455, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8600028157234192, "num_tokens": 489557707.0, "step": 12897 }, { "epoch": 1.6407581732603993, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4805455207824707, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.865985631942749, "num_tokens": 489591180.0, "step": 12898 }, { "epoch": 1.6408853835389898, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.264129638671875, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8707408905029297, "num_tokens": 489631732.0, "step": 12899 }, { "epoch": 1.6410125938175804, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5272459983825684, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8674865961074829, "num_tokens": 489671484.0, "step": 12900 }, { "epoch": 1.6411398040961709, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.2120580673217773, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8718166947364807, "num_tokens": 489706929.0, "step": 12901 }, { "epoch": 1.6412670143747614, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.63927960395813, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8706017732620239, "num_tokens": 489741830.0, "step": 12902 }, { "epoch": 1.641394224653352, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.113440752029419, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8814958333969116, "num_tokens": 489782808.0, "step": 12903 }, { "epoch": 1.6415214349319425, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8322181701660156, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.862046480178833, "num_tokens": 489823536.0, "step": 12904 }, { "epoch": 1.641648645210533, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.060591697692871, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8751592040061951, "num_tokens": 489854730.0, "step": 12905 }, { "epoch": 1.6417758554891235, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.118175029754639, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8731734752655029, "num_tokens": 489886836.0, "step": 12906 }, { "epoch": 1.641903065767714, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.142667055130005, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8738409280776978, "num_tokens": 489928227.0, "step": 12907 }, { "epoch": 1.6420302760463046, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.416485786437988, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8577049970626831, "num_tokens": 489965697.0, "step": 12908 }, { "epoch": 1.642157486324895, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5187535285949707, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8788799047470093, "num_tokens": 490000344.0, "step": 12909 }, { "epoch": 1.6422846966034856, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.426451206207275, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8921489119529724, "num_tokens": 490030965.0, "step": 12910 }, { "epoch": 1.6424119068820762, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.829291820526123, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8797036409378052, "num_tokens": 490069386.0, "step": 12911 }, { "epoch": 1.6425391171606667, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1653876304626465, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8800003528594971, "num_tokens": 490104040.0, "step": 12912 }, { "epoch": 1.6426663274392572, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.8204965591430664, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8626691102981567, "num_tokens": 490147883.0, "step": 12913 }, { "epoch": 1.6427935377178478, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.950747013092041, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8718509674072266, "num_tokens": 490183889.0, "step": 12914 }, { "epoch": 1.642920747996438, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.697685956954956, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8732126951217651, "num_tokens": 490224132.0, "step": 12915 }, { "epoch": 1.6430479582750286, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8094472885131836, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.870071530342102, "num_tokens": 490262745.0, "step": 12916 }, { "epoch": 1.6431751685536191, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5687553882598877, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.867082417011261, "num_tokens": 490301449.0, "step": 12917 }, { "epoch": 1.6433023788322096, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7534148693084717, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8591822385787964, "num_tokens": 490328873.0, "step": 12918 }, { "epoch": 1.6434295891108002, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4547414779663086, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8756648302078247, "num_tokens": 490361174.0, "step": 12919 }, { "epoch": 1.6435567993893907, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7140567302703857, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8635802268981934, "num_tokens": 490399708.0, "step": 12920 }, { "epoch": 1.6436840096679812, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3412139415740967, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8843186497688293, "num_tokens": 490433159.0, "step": 12921 }, { "epoch": 1.6438112199465715, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7697160243988037, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.873031735420227, "num_tokens": 490473335.0, "step": 12922 }, { "epoch": 1.643938430225162, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2520127296447754, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8626517057418823, "num_tokens": 490517091.0, "step": 12923 }, { "epoch": 1.6440656405037526, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.917919158935547, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8680197596549988, "num_tokens": 490550336.0, "step": 12924 }, { "epoch": 1.6441928507823431, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.234589099884033, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8874106407165527, "num_tokens": 490593592.0, "step": 12925 }, { "epoch": 1.6443200610609336, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.084605693817139, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8818748593330383, "num_tokens": 490629354.0, "step": 12926 }, { "epoch": 1.6444472713395242, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4270503520965576, "learning_rate": 1e-05, "loss": 0.3792, "mean_token_accuracy": 0.8941221237182617, "num_tokens": 490668319.0, "step": 12927 }, { "epoch": 1.6445744816181147, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.075152397155762, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8785226345062256, "num_tokens": 490705978.0, "step": 12928 }, { "epoch": 1.6447016918967052, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5282697677612305, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8730167746543884, "num_tokens": 490742455.0, "step": 12929 }, { "epoch": 1.6448289021752958, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.948089599609375, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8570756912231445, "num_tokens": 490781733.0, "step": 12930 }, { "epoch": 1.6449561124538863, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5945184230804443, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8740816712379456, "num_tokens": 490821077.0, "step": 12931 }, { "epoch": 1.6450833227324768, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7400124073028564, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8760076761245728, "num_tokens": 490862725.0, "step": 12932 }, { "epoch": 1.6452105330110673, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.596320152282715, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8579460382461548, "num_tokens": 490909612.0, "step": 12933 }, { "epoch": 1.6453377432896579, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.552333354949951, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8740397095680237, "num_tokens": 490943202.0, "step": 12934 }, { "epoch": 1.6454649535682484, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5931856632232666, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8703190088272095, "num_tokens": 490978481.0, "step": 12935 }, { "epoch": 1.645592163846839, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3894548416137695, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8754204511642456, "num_tokens": 491018672.0, "step": 12936 }, { "epoch": 1.6457193741254295, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5872268676757812, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.874866783618927, "num_tokens": 491055032.0, "step": 12937 }, { "epoch": 1.64584658440402, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.325927734375, "learning_rate": 1e-05, "loss": 0.518, "mean_token_accuracy": 0.8461430668830872, "num_tokens": 491089780.0, "step": 12938 }, { "epoch": 1.6459737946826105, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6065597534179688, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8759558200836182, "num_tokens": 491124187.0, "step": 12939 }, { "epoch": 1.6461010049612008, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.218567371368408, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8690546154975891, "num_tokens": 491167558.0, "step": 12940 }, { "epoch": 1.6462282152397913, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6902875900268555, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8677686452865601, "num_tokens": 491210054.0, "step": 12941 }, { "epoch": 1.6463554255183819, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1433093547821045, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8766962289810181, "num_tokens": 491256535.0, "step": 12942 }, { "epoch": 1.6464826357969724, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.855708122253418, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8781362771987915, "num_tokens": 491295909.0, "step": 12943 }, { "epoch": 1.646609846075563, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1378276348114014, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8608821630477905, "num_tokens": 491332547.0, "step": 12944 }, { "epoch": 1.6467370563541535, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.054713249206543, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8875660300254822, "num_tokens": 491373187.0, "step": 12945 }, { "epoch": 1.6468642666327438, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.287325620651245, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8725686073303223, "num_tokens": 491412476.0, "step": 12946 }, { "epoch": 1.6469914769113343, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.1409687995910645, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8784002065658569, "num_tokens": 491456502.0, "step": 12947 }, { "epoch": 1.6471186871899248, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.3988096714019775, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8787338137626648, "num_tokens": 491492974.0, "step": 12948 }, { "epoch": 1.6472458974685154, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.154033660888672, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8722161054611206, "num_tokens": 491523621.0, "step": 12949 }, { "epoch": 1.6473731077471059, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.571040630340576, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8598780035972595, "num_tokens": 491563366.0, "step": 12950 }, { "epoch": 1.6475003180256964, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.04034948348999, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8707227110862732, "num_tokens": 491594221.0, "step": 12951 }, { "epoch": 1.647627528304287, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6258792877197266, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8625966310501099, "num_tokens": 491632039.0, "step": 12952 }, { "epoch": 1.6477547385828775, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.818345069885254, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8774914741516113, "num_tokens": 491672282.0, "step": 12953 }, { "epoch": 1.647881948861468, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5994720458984375, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8589825630187988, "num_tokens": 491707969.0, "step": 12954 }, { "epoch": 1.6480091591400585, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.660733938217163, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8759862184524536, "num_tokens": 491743347.0, "step": 12955 }, { "epoch": 1.648136369418649, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.545989990234375, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8690221309661865, "num_tokens": 491777165.0, "step": 12956 }, { "epoch": 1.6482635796972396, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5380001068115234, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8850683569908142, "num_tokens": 491813859.0, "step": 12957 }, { "epoch": 1.64839078997583, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4824066162109375, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8655515909194946, "num_tokens": 491853373.0, "step": 12958 }, { "epoch": 1.6485180002544206, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.476552963256836, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8757028579711914, "num_tokens": 491889401.0, "step": 12959 }, { "epoch": 1.6486452105330112, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4248344898223877, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8676139116287231, "num_tokens": 491933973.0, "step": 12960 }, { "epoch": 1.6487724208116017, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4549193382263184, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8640574216842651, "num_tokens": 491979198.0, "step": 12961 }, { "epoch": 1.6488996310901922, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3969287872314453, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8610898852348328, "num_tokens": 492009127.0, "step": 12962 }, { "epoch": 1.6490268413687827, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4402124881744385, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8725757598876953, "num_tokens": 492045022.0, "step": 12963 }, { "epoch": 1.649154051647373, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.336357831954956, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8694876432418823, "num_tokens": 492092155.0, "step": 12964 }, { "epoch": 1.6492812619259636, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.459209442138672, "learning_rate": 1e-05, "loss": 0.4973, "mean_token_accuracy": 0.854827880859375, "num_tokens": 492131807.0, "step": 12965 }, { "epoch": 1.649408472204554, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2922093868255615, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8728004693984985, "num_tokens": 492166794.0, "step": 12966 }, { "epoch": 1.6495356824831446, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.512845516204834, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8773177266120911, "num_tokens": 492208375.0, "step": 12967 }, { "epoch": 1.6496628927617352, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.248379945755005, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8596768975257874, "num_tokens": 492241166.0, "step": 12968 }, { "epoch": 1.6497901030403257, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.580193281173706, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8856677412986755, "num_tokens": 492278887.0, "step": 12969 }, { "epoch": 1.6499173133189162, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.171424388885498, "learning_rate": 1e-05, "loss": 0.3877, "mean_token_accuracy": 0.8872265815734863, "num_tokens": 492316523.0, "step": 12970 }, { "epoch": 1.6500445235975065, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7261648178100586, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8653795123100281, "num_tokens": 492354094.0, "step": 12971 }, { "epoch": 1.650171733876097, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1274473667144775, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8619779348373413, "num_tokens": 492387250.0, "step": 12972 }, { "epoch": 1.6502989441546876, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.922593832015991, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8687185049057007, "num_tokens": 492423770.0, "step": 12973 }, { "epoch": 1.6504261544332781, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.087209939956665, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8692858219146729, "num_tokens": 492465970.0, "step": 12974 }, { "epoch": 1.6505533647118686, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.20562219619751, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8637760877609253, "num_tokens": 492501668.0, "step": 12975 }, { "epoch": 1.6506805749904592, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.272960662841797, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8660967350006104, "num_tokens": 492541032.0, "step": 12976 }, { "epoch": 1.6508077852690497, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.370794296264648, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8676638603210449, "num_tokens": 492579639.0, "step": 12977 }, { "epoch": 1.6509349955476402, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.621969223022461, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.868248462677002, "num_tokens": 492612324.0, "step": 12978 }, { "epoch": 1.6510622058262308, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.243150234222412, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8643734455108643, "num_tokens": 492648967.0, "step": 12979 }, { "epoch": 1.6511894161048213, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.753042697906494, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8847277164459229, "num_tokens": 492687117.0, "step": 12980 }, { "epoch": 1.6513166263834118, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.034877300262451, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8733221888542175, "num_tokens": 492718760.0, "step": 12981 }, { "epoch": 1.6514438366620023, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7129502296447754, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8741254210472107, "num_tokens": 492760038.0, "step": 12982 }, { "epoch": 1.6515710469405929, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8407251834869385, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.8603015542030334, "num_tokens": 492798450.0, "step": 12983 }, { "epoch": 1.6516982572191834, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.608821153640747, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.87442547082901, "num_tokens": 492835081.0, "step": 12984 }, { "epoch": 1.651825467497774, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7121951580047607, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8796257972717285, "num_tokens": 492878255.0, "step": 12985 }, { "epoch": 1.6519526777763645, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.503026247024536, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8612579703330994, "num_tokens": 492920800.0, "step": 12986 }, { "epoch": 1.652079888054955, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.652413845062256, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8690810203552246, "num_tokens": 492957170.0, "step": 12987 }, { "epoch": 1.6522070983335455, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.405452013015747, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8730875849723816, "num_tokens": 492995123.0, "step": 12988 }, { "epoch": 1.6523343086121358, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6524155139923096, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8743706941604614, "num_tokens": 493034538.0, "step": 12989 }, { "epoch": 1.6524615188907263, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.317676067352295, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8572369813919067, "num_tokens": 493073218.0, "step": 12990 }, { "epoch": 1.6525887291693169, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7086613178253174, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8687683343887329, "num_tokens": 493110056.0, "step": 12991 }, { "epoch": 1.6527159394479074, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2408411502838135, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8785264492034912, "num_tokens": 493144324.0, "step": 12992 }, { "epoch": 1.652843149726498, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.83186411857605, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8628085851669312, "num_tokens": 493181685.0, "step": 12993 }, { "epoch": 1.6529703600050885, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.199228525161743, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8782762289047241, "num_tokens": 493217712.0, "step": 12994 }, { "epoch": 1.6530975702836788, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.003721237182617, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8563675284385681, "num_tokens": 493253852.0, "step": 12995 }, { "epoch": 1.6532247805622693, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.283492088317871, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8715826869010925, "num_tokens": 493287797.0, "step": 12996 }, { "epoch": 1.6533519908408598, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.110642433166504, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8737480044364929, "num_tokens": 493326072.0, "step": 12997 }, { "epoch": 1.6534792011194503, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.453747510910034, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8550364971160889, "num_tokens": 493365374.0, "step": 12998 }, { "epoch": 1.6536064113980409, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.043322563171387, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8873987197875977, "num_tokens": 493408526.0, "step": 12999 }, { "epoch": 1.6537336216766314, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.590207576751709, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.860117495059967, "num_tokens": 493445689.0, "step": 13000 }, { "epoch": 1.653860831955222, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.869521379470825, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.862143874168396, "num_tokens": 493480884.0, "step": 13001 }, { "epoch": 1.6539880422338125, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.604524612426758, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8731749057769775, "num_tokens": 493523917.0, "step": 13002 }, { "epoch": 1.654115252512403, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6520426273345947, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8702210187911987, "num_tokens": 493560438.0, "step": 13003 }, { "epoch": 1.6542424627909935, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.59763765335083, "learning_rate": 1e-05, "loss": 0.4985, "mean_token_accuracy": 0.8599866628646851, "num_tokens": 493596609.0, "step": 13004 }, { "epoch": 1.654369673069584, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.463589668273926, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8648831248283386, "num_tokens": 493634369.0, "step": 13005 }, { "epoch": 1.6544968833481746, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6002860069274902, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8658667802810669, "num_tokens": 493669568.0, "step": 13006 }, { "epoch": 1.654624093626765, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3322913646698, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8762418031692505, "num_tokens": 493710136.0, "step": 13007 }, { "epoch": 1.6547513039053556, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.614610195159912, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8672133088111877, "num_tokens": 493750028.0, "step": 13008 }, { "epoch": 1.6548785141839462, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2587521076202393, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8595407009124756, "num_tokens": 493785167.0, "step": 13009 }, { "epoch": 1.6550057244625367, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.676074504852295, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8595963716506958, "num_tokens": 493822724.0, "step": 13010 }, { "epoch": 1.6551329347411272, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.192814826965332, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8627810478210449, "num_tokens": 493863910.0, "step": 13011 }, { "epoch": 1.6552601450197177, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.778115749359131, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8790762424468994, "num_tokens": 493901863.0, "step": 13012 }, { "epoch": 1.655387355298308, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.166903018951416, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8731526136398315, "num_tokens": 493938174.0, "step": 13013 }, { "epoch": 1.6555145655768986, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.935068368911743, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8696995973587036, "num_tokens": 493973483.0, "step": 13014 }, { "epoch": 1.655641775855489, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.181379556655884, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8837235569953918, "num_tokens": 494016481.0, "step": 13015 }, { "epoch": 1.6557689861340796, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.114716529846191, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8731905221939087, "num_tokens": 494047543.0, "step": 13016 }, { "epoch": 1.6558961964126702, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002613067626953125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.339881658554077, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8692442774772644, "num_tokens": 494082076.0, "step": 13017 }, { "epoch": 1.6560234066912607, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.176036834716797, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8677509427070618, "num_tokens": 494123822.0, "step": 13018 }, { "epoch": 1.6561506169698512, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5315325260162354, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8852820992469788, "num_tokens": 494152644.0, "step": 13019 }, { "epoch": 1.6562778272484415, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.08278751373291, "learning_rate": 1e-05, "loss": 0.5214, "mean_token_accuracy": 0.8552967309951782, "num_tokens": 494185667.0, "step": 13020 }, { "epoch": 1.656405037527032, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6176581382751465, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8809061050415039, "num_tokens": 494224277.0, "step": 13021 }, { "epoch": 1.6565322478056226, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9071881771087646, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.865192174911499, "num_tokens": 494254475.0, "step": 13022 }, { "epoch": 1.656659458084213, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6089556217193604, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8702738285064697, "num_tokens": 494291891.0, "step": 13023 }, { "epoch": 1.6567866683628036, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7291603088378906, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8647282123565674, "num_tokens": 494327659.0, "step": 13024 }, { "epoch": 1.6569138786413942, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.555279016494751, "learning_rate": 1e-05, "loss": 0.4028, "mean_token_accuracy": 0.8871643543243408, "num_tokens": 494358149.0, "step": 13025 }, { "epoch": 1.6570410889199847, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6093358993530273, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8766700625419617, "num_tokens": 494393673.0, "step": 13026 }, { "epoch": 1.6571682991985752, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.514880418777466, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8727701902389526, "num_tokens": 494422811.0, "step": 13027 }, { "epoch": 1.6572955094771658, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5071141719818115, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8706411123275757, "num_tokens": 494467758.0, "step": 13028 }, { "epoch": 1.6574227197557563, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.436143636703491, "learning_rate": 1e-05, "loss": 0.3982, "mean_token_accuracy": 0.8863978385925293, "num_tokens": 494504698.0, "step": 13029 }, { "epoch": 1.6575499300343468, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4851667881011963, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8614047765731812, "num_tokens": 494541235.0, "step": 13030 }, { "epoch": 1.6576771403129373, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3679563999176025, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8714604377746582, "num_tokens": 494584499.0, "step": 13031 }, { "epoch": 1.6578043505915279, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.480612277984619, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.869755744934082, "num_tokens": 494625845.0, "step": 13032 }, { "epoch": 1.6579315608701184, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3318402767181396, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8733786940574646, "num_tokens": 494660820.0, "step": 13033 }, { "epoch": 1.658058771148709, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5015604496002197, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8840504884719849, "num_tokens": 494701215.0, "step": 13034 }, { "epoch": 1.6581859814272994, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.287100076675415, "learning_rate": 1e-05, "loss": 0.5383, "mean_token_accuracy": 0.8424584269523621, "num_tokens": 494739062.0, "step": 13035 }, { "epoch": 1.65831319170589, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5482301712036133, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8695278167724609, "num_tokens": 494779467.0, "step": 13036 }, { "epoch": 1.6584404019844805, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2171051502227783, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8709902763366699, "num_tokens": 494817237.0, "step": 13037 }, { "epoch": 1.6585676122630708, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6483142375946045, "learning_rate": 1e-05, "loss": 0.3971, "mean_token_accuracy": 0.8875609636306763, "num_tokens": 494853485.0, "step": 13038 }, { "epoch": 1.6586948225416613, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1504435539245605, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8757134079933167, "num_tokens": 494890456.0, "step": 13039 }, { "epoch": 1.6588220328202519, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8291196823120117, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8585607409477234, "num_tokens": 494929441.0, "step": 13040 }, { "epoch": 1.6589492430988424, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.11372447013855, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8634297251701355, "num_tokens": 494963642.0, "step": 13041 }, { "epoch": 1.659076453377433, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.063655376434326, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8624759316444397, "num_tokens": 494998138.0, "step": 13042 }, { "epoch": 1.6592036636560235, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.1897711753845215, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8725947737693787, "num_tokens": 495038080.0, "step": 13043 }, { "epoch": 1.6593308739346138, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2660908699035645, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.865754246711731, "num_tokens": 495079103.0, "step": 13044 }, { "epoch": 1.6594580842132043, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4412760734558105, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8716083765029907, "num_tokens": 495119385.0, "step": 13045 }, { "epoch": 1.6595852944917948, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.274885654449463, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8651472330093384, "num_tokens": 495157829.0, "step": 13046 }, { "epoch": 1.6597125047703853, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6886885166168213, "learning_rate": 1e-05, "loss": 0.5484, "mean_token_accuracy": 0.8464919328689575, "num_tokens": 495198554.0, "step": 13047 }, { "epoch": 1.6598397150489759, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0594096183776855, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8778531551361084, "num_tokens": 495241389.0, "step": 13048 }, { "epoch": 1.6599669253275664, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7047998905181885, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8820042014122009, "num_tokens": 495274216.0, "step": 13049 }, { "epoch": 1.660094135606157, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8687734603881836, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8639254570007324, "num_tokens": 495308941.0, "step": 13050 }, { "epoch": 1.6602213458847475, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.628401756286621, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8743565082550049, "num_tokens": 495351668.0, "step": 13051 }, { "epoch": 1.660348556163338, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.715484380722046, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8625063300132751, "num_tokens": 495383402.0, "step": 13052 }, { "epoch": 1.6604757664419285, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5310535430908203, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8826906681060791, "num_tokens": 495421371.0, "step": 13053 }, { "epoch": 1.660602976720519, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.614616870880127, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8649730086326599, "num_tokens": 495460692.0, "step": 13054 }, { "epoch": 1.6607301869991096, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.454200506210327, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8749856948852539, "num_tokens": 495493990.0, "step": 13055 }, { "epoch": 1.6608573972777, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5703353881835938, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8680854439735413, "num_tokens": 495531616.0, "step": 13056 }, { "epoch": 1.6609846075562906, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.375265121459961, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8615421056747437, "num_tokens": 495572065.0, "step": 13057 }, { "epoch": 1.6611118178348812, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5702834129333496, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8547301888465881, "num_tokens": 495612582.0, "step": 13058 }, { "epoch": 1.6612390281134717, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.294447898864746, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8678168058395386, "num_tokens": 495649211.0, "step": 13059 }, { "epoch": 1.6613662383920622, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6281819343566895, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8693031668663025, "num_tokens": 495689805.0, "step": 13060 }, { "epoch": 1.6614934486706527, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2444474697113037, "learning_rate": 1e-05, "loss": 0.5292, "mean_token_accuracy": 0.8456141352653503, "num_tokens": 495727560.0, "step": 13061 }, { "epoch": 1.661620658949243, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7368361949920654, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8721057176589966, "num_tokens": 495759827.0, "step": 13062 }, { "epoch": 1.6617478692278336, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.172088623046875, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8723466396331787, "num_tokens": 495797106.0, "step": 13063 }, { "epoch": 1.661875079506424, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8834972381591797, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8643794655799866, "num_tokens": 495842331.0, "step": 13064 }, { "epoch": 1.6620022897850146, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.166731595993042, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8662169575691223, "num_tokens": 495883491.0, "step": 13065 }, { "epoch": 1.6621295000636052, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.080495357513428, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8591434955596924, "num_tokens": 495924536.0, "step": 13066 }, { "epoch": 1.6622567103421957, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3063056468963623, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8590192198753357, "num_tokens": 495961783.0, "step": 13067 }, { "epoch": 1.662383920620786, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.163956165313721, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8719093203544617, "num_tokens": 495994131.0, "step": 13068 }, { "epoch": 1.6625111308993765, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.509178638458252, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8680516481399536, "num_tokens": 496035981.0, "step": 13069 }, { "epoch": 1.662638341177967, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0578436851501465, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8594397306442261, "num_tokens": 496075622.0, "step": 13070 }, { "epoch": 1.6627655514565576, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6176319122314453, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8687784671783447, "num_tokens": 496117130.0, "step": 13071 }, { "epoch": 1.662892761735148, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8554224967956543, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.8926609754562378, "num_tokens": 496153309.0, "step": 13072 }, { "epoch": 1.6630199720137386, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.620388984680176, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8831119537353516, "num_tokens": 496194955.0, "step": 13073 }, { "epoch": 1.6631471822923292, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6374635696411133, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8684217929840088, "num_tokens": 496238787.0, "step": 13074 }, { "epoch": 1.6632743925709197, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5908873081207275, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8605860471725464, "num_tokens": 496279606.0, "step": 13075 }, { "epoch": 1.6634016028495102, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.484529733657837, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8767348527908325, "num_tokens": 496315724.0, "step": 13076 }, { "epoch": 1.6635288131281007, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5585339069366455, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8738938570022583, "num_tokens": 496357254.0, "step": 13077 }, { "epoch": 1.6636560234066913, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.379424810409546, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8729375600814819, "num_tokens": 496394356.0, "step": 13078 }, { "epoch": 1.6637832336852818, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.559701681137085, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8666813969612122, "num_tokens": 496431569.0, "step": 13079 }, { "epoch": 1.6639104439638723, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3083951473236084, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8730705380439758, "num_tokens": 496465854.0, "step": 13080 }, { "epoch": 1.6640376542424629, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5751097202301025, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8725045919418335, "num_tokens": 496506023.0, "step": 13081 }, { "epoch": 1.6641648645210534, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2444164752960205, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.858684778213501, "num_tokens": 496545950.0, "step": 13082 }, { "epoch": 1.664292074799644, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6342484951019287, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.874366819858551, "num_tokens": 496585728.0, "step": 13083 }, { "epoch": 1.6644192850782344, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1845145225524902, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.8528943061828613, "num_tokens": 496628491.0, "step": 13084 }, { "epoch": 1.664546495356825, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7615222930908203, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8623666763305664, "num_tokens": 496665952.0, "step": 13085 }, { "epoch": 1.6646737056354155, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.121344804763794, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8816136717796326, "num_tokens": 496708188.0, "step": 13086 }, { "epoch": 1.6648009159140058, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9667868614196777, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8797538876533508, "num_tokens": 496743079.0, "step": 13087 }, { "epoch": 1.6649281261925963, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1478986740112305, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8694285750389099, "num_tokens": 496781364.0, "step": 13088 }, { "epoch": 1.6650553364711869, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1897711753845215, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.882670521736145, "num_tokens": 496816419.0, "step": 13089 }, { "epoch": 1.6651825467497774, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.342841148376465, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8761603236198425, "num_tokens": 496851941.0, "step": 13090 }, { "epoch": 1.665309757028368, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.283559322357178, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8751764297485352, "num_tokens": 496884974.0, "step": 13091 }, { "epoch": 1.6654369673069584, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6006669998168945, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8761375546455383, "num_tokens": 496920331.0, "step": 13092 }, { "epoch": 1.6655641775855488, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1382598876953125, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8680135011672974, "num_tokens": 496957599.0, "step": 13093 }, { "epoch": 1.6656913878641393, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.700141191482544, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8838874697685242, "num_tokens": 496992804.0, "step": 13094 }, { "epoch": 1.6658185981427298, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9268651008605957, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8836756944656372, "num_tokens": 497025376.0, "step": 13095 }, { "epoch": 1.6659458084213203, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.654961109161377, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.87892085313797, "num_tokens": 497067923.0, "step": 13096 }, { "epoch": 1.6660730186999109, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.740163564682007, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.8613221645355225, "num_tokens": 497113753.0, "step": 13097 }, { "epoch": 1.6662002289785014, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.593547821044922, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8743061423301697, "num_tokens": 497148001.0, "step": 13098 }, { "epoch": 1.666327439257092, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6118195056915283, "learning_rate": 1e-05, "loss": 0.5106, "mean_token_accuracy": 0.8518285155296326, "num_tokens": 497186702.0, "step": 13099 }, { "epoch": 1.6664546495356825, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.508648633956909, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8712067604064941, "num_tokens": 497224550.0, "step": 13100 }, { "epoch": 1.666581859814273, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5591559410095215, "learning_rate": 1e-05, "loss": 0.5025, "mean_token_accuracy": 0.8520016670227051, "num_tokens": 497261922.0, "step": 13101 }, { "epoch": 1.6667090700928635, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4262735843658447, "learning_rate": 1e-05, "loss": 0.5243, "mean_token_accuracy": 0.8469607830047607, "num_tokens": 497302297.0, "step": 13102 }, { "epoch": 1.666836280371454, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5302412509918213, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8703795075416565, "num_tokens": 497337788.0, "step": 13103 }, { "epoch": 1.6669634906500446, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.352215051651001, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8575783371925354, "num_tokens": 497373455.0, "step": 13104 }, { "epoch": 1.667090700928635, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5439279079437256, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8716622591018677, "num_tokens": 497414743.0, "step": 13105 }, { "epoch": 1.6672179112072256, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2823665142059326, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8744797110557556, "num_tokens": 497446769.0, "step": 13106 }, { "epoch": 1.6673451214858162, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5996294021606445, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8897744417190552, "num_tokens": 497488693.0, "step": 13107 }, { "epoch": 1.6674723317644067, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2184946537017822, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8784283399581909, "num_tokens": 497522258.0, "step": 13108 }, { "epoch": 1.6675995420429972, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7197153568267822, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.865171492099762, "num_tokens": 497557175.0, "step": 13109 }, { "epoch": 1.6677267523215877, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1456620693206787, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8835055232048035, "num_tokens": 497592880.0, "step": 13110 }, { "epoch": 1.667853962600178, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9066481590270996, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8747269511222839, "num_tokens": 497628416.0, "step": 13111 }, { "epoch": 1.6679811728787686, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.152848243713379, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8582097291946411, "num_tokens": 497666739.0, "step": 13112 }, { "epoch": 1.668108383157359, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.110714435577393, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8698985576629639, "num_tokens": 497706307.0, "step": 13113 }, { "epoch": 1.6682355934359496, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.2985270023345947, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8570668697357178, "num_tokens": 497747623.0, "step": 13114 }, { "epoch": 1.6683628037145402, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.211854457855225, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8689208030700684, "num_tokens": 497783354.0, "step": 13115 }, { "epoch": 1.6684900139931307, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.544872999191284, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8720653057098389, "num_tokens": 497818178.0, "step": 13116 }, { "epoch": 1.668617224271721, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.096611022949219, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8714965581893921, "num_tokens": 497852899.0, "step": 13117 }, { "epoch": 1.6687444345503115, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6489675045013428, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8721911907196045, "num_tokens": 497896743.0, "step": 13118 }, { "epoch": 1.668871644828902, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8744192123413086, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.875214695930481, "num_tokens": 497938837.0, "step": 13119 }, { "epoch": 1.6689988551074926, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6505956649780273, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8788958787918091, "num_tokens": 497970335.0, "step": 13120 }, { "epoch": 1.669126065386083, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6864876747131348, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8665258884429932, "num_tokens": 498003484.0, "step": 13121 }, { "epoch": 1.6692532756646736, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000263214111328125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5774505138397217, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8824135065078735, "num_tokens": 498040932.0, "step": 13122 }, { "epoch": 1.6693804859432642, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5437076091766357, "learning_rate": 1e-05, "loss": 0.5377, "mean_token_accuracy": 0.8464312553405762, "num_tokens": 498078911.0, "step": 13123 }, { "epoch": 1.6695076962218547, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.529024600982666, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8801828026771545, "num_tokens": 498119917.0, "step": 13124 }, { "epoch": 1.6696349065004452, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.435068130493164, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8770492672920227, "num_tokens": 498160359.0, "step": 13125 }, { "epoch": 1.6697621167790357, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.51859974861145, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8646386861801147, "num_tokens": 498197048.0, "step": 13126 }, { "epoch": 1.6698893270576263, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3710811138153076, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.8537474870681763, "num_tokens": 498237007.0, "step": 13127 }, { "epoch": 1.6700165373362168, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.496777057647705, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.87046879529953, "num_tokens": 498275324.0, "step": 13128 }, { "epoch": 1.6701437476148073, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3203744888305664, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8527545928955078, "num_tokens": 498317018.0, "step": 13129 }, { "epoch": 1.6702709578933979, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5093636512756348, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8725938200950623, "num_tokens": 498353903.0, "step": 13130 }, { "epoch": 1.6703981681719884, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2680704593658447, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8618505001068115, "num_tokens": 498390959.0, "step": 13131 }, { "epoch": 1.670525378450579, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.549766778945923, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8794682025909424, "num_tokens": 498433067.0, "step": 13132 }, { "epoch": 1.6706525887291694, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.205913782119751, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8645568490028381, "num_tokens": 498476020.0, "step": 13133 }, { "epoch": 1.67077979900776, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.649693489074707, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8742556571960449, "num_tokens": 498515644.0, "step": 13134 }, { "epoch": 1.6709070092863505, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.14011287689209, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8740687966346741, "num_tokens": 498551063.0, "step": 13135 }, { "epoch": 1.6710342195649408, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.817636489868164, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8680150508880615, "num_tokens": 498588849.0, "step": 13136 }, { "epoch": 1.6711614298435313, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.0917930603027344, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8654555082321167, "num_tokens": 498626073.0, "step": 13137 }, { "epoch": 1.6712886401221219, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.065096855163574, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8747225999832153, "num_tokens": 498661523.0, "step": 13138 }, { "epoch": 1.6714158504007124, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1549336910247803, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8625415563583374, "num_tokens": 498702075.0, "step": 13139 }, { "epoch": 1.671543060679303, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.32578706741333, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8633061647415161, "num_tokens": 498737425.0, "step": 13140 }, { "epoch": 1.6716702709578934, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.4493439197540283, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8498332500457764, "num_tokens": 498779715.0, "step": 13141 }, { "epoch": 1.6717974812364838, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.347552299499512, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8797235488891602, "num_tokens": 498817312.0, "step": 13142 }, { "epoch": 1.6719246915150743, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.733927011489868, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8609402179718018, "num_tokens": 498857990.0, "step": 13143 }, { "epoch": 1.6720519017936648, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.145017147064209, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8674731850624084, "num_tokens": 498896275.0, "step": 13144 }, { "epoch": 1.6721791120722553, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7762694358825684, "learning_rate": 1e-05, "loss": 0.5131, "mean_token_accuracy": 0.85988450050354, "num_tokens": 498933662.0, "step": 13145 }, { "epoch": 1.6723063223508459, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9109513759613037, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.872006893157959, "num_tokens": 498974216.0, "step": 13146 }, { "epoch": 1.6724335326294364, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6735219955444336, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8686098456382751, "num_tokens": 499015889.0, "step": 13147 }, { "epoch": 1.672560742908027, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.739619493484497, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8802476525306702, "num_tokens": 499056889.0, "step": 13148 }, { "epoch": 1.6726879531866174, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5706984996795654, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8717628717422485, "num_tokens": 499097440.0, "step": 13149 }, { "epoch": 1.672815163465208, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.640087127685547, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8779963254928589, "num_tokens": 499135450.0, "step": 13150 }, { "epoch": 1.6729423737437985, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4763145446777344, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8610341548919678, "num_tokens": 499178487.0, "step": 13151 }, { "epoch": 1.673069584022389, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5975303649902344, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8656579256057739, "num_tokens": 499213824.0, "step": 13152 }, { "epoch": 1.6731967943009796, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.384777069091797, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8693647384643555, "num_tokens": 499252406.0, "step": 13153 }, { "epoch": 1.67332400457957, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5992441177368164, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8626508712768555, "num_tokens": 499295249.0, "step": 13154 }, { "epoch": 1.6734512148581606, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2841641902923584, "learning_rate": 1e-05, "loss": 0.4121, "mean_token_accuracy": 0.881016194820404, "num_tokens": 499334115.0, "step": 13155 }, { "epoch": 1.6735784251367511, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.674222707748413, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8658270239830017, "num_tokens": 499372181.0, "step": 13156 }, { "epoch": 1.6737056354153417, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2049829959869385, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8627461791038513, "num_tokens": 499413617.0, "step": 13157 }, { "epoch": 1.6738328456939322, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.830815076828003, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8717052340507507, "num_tokens": 499443976.0, "step": 13158 }, { "epoch": 1.6739600559725227, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1753437519073486, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8653264045715332, "num_tokens": 499477973.0, "step": 13159 }, { "epoch": 1.674087266251113, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 4.001631259918213, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.8570066690444946, "num_tokens": 499523163.0, "step": 13160 }, { "epoch": 1.6742144765297036, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2319626808166504, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8556613326072693, "num_tokens": 499567583.0, "step": 13161 }, { "epoch": 1.674341686808294, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.173321723937988, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8641506433486938, "num_tokens": 499603257.0, "step": 13162 }, { "epoch": 1.6744688970868846, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4364280700683594, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8739607930183411, "num_tokens": 499648881.0, "step": 13163 }, { "epoch": 1.6745961073654752, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.143311500549316, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8685534596443176, "num_tokens": 499688632.0, "step": 13164 }, { "epoch": 1.6747233176440657, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6321213245391846, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8707168698310852, "num_tokens": 499723285.0, "step": 13165 }, { "epoch": 1.674850527922656, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9462525844573975, "learning_rate": 1e-05, "loss": 0.5202, "mean_token_accuracy": 0.8527684211730957, "num_tokens": 499761472.0, "step": 13166 }, { "epoch": 1.6749777382012465, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6622657775878906, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8621293902397156, "num_tokens": 499796997.0, "step": 13167 }, { "epoch": 1.675104948479837, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7271721363067627, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.87043297290802, "num_tokens": 499834307.0, "step": 13168 }, { "epoch": 1.6752321587584276, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6206510066986084, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8695915937423706, "num_tokens": 499876497.0, "step": 13169 }, { "epoch": 1.675359369037018, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5353126525878906, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8597023487091064, "num_tokens": 499917620.0, "step": 13170 }, { "epoch": 1.6754865793156086, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5794289112091064, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8761270642280579, "num_tokens": 499957098.0, "step": 13171 }, { "epoch": 1.6756137895941992, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.407144784927368, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8677182793617249, "num_tokens": 500001046.0, "step": 13172 }, { "epoch": 1.6757409998727897, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.560762643814087, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.866005539894104, "num_tokens": 500041443.0, "step": 13173 }, { "epoch": 1.6758682101513802, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3270232677459717, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8864989280700684, "num_tokens": 500076624.0, "step": 13174 }, { "epoch": 1.6759954204299707, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5795795917510986, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.8622840046882629, "num_tokens": 500118995.0, "step": 13175 }, { "epoch": 1.6761226307085613, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.260465383529663, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8658857941627502, "num_tokens": 500155134.0, "step": 13176 }, { "epoch": 1.6762498409871518, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6445093154907227, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8769171237945557, "num_tokens": 500189921.0, "step": 13177 }, { "epoch": 1.6763770512657423, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1918046474456787, "learning_rate": 1e-05, "loss": 0.4996, "mean_token_accuracy": 0.8538485765457153, "num_tokens": 500230655.0, "step": 13178 }, { "epoch": 1.6765042615443329, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7568609714508057, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8691229224205017, "num_tokens": 500270827.0, "step": 13179 }, { "epoch": 1.6766314718229234, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1360089778900146, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8579148650169373, "num_tokens": 500312810.0, "step": 13180 }, { "epoch": 1.676758682101514, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9507899284362793, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8600447177886963, "num_tokens": 500345013.0, "step": 13181 }, { "epoch": 1.6768858923801044, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002651214599609375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.170029640197754, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8712295889854431, "num_tokens": 500375628.0, "step": 13182 }, { "epoch": 1.677013102658695, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.14689826965332, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.886399507522583, "num_tokens": 500410084.0, "step": 13183 }, { "epoch": 1.6771403129372855, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.302741289138794, "learning_rate": 1e-05, "loss": 0.3994, "mean_token_accuracy": 0.887354850769043, "num_tokens": 500452221.0, "step": 13184 }, { "epoch": 1.6772675232158758, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.255795001983643, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8697264790534973, "num_tokens": 500491358.0, "step": 13185 }, { "epoch": 1.6773947334944663, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5693562030792236, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8649834394454956, "num_tokens": 500530363.0, "step": 13186 }, { "epoch": 1.6775219437730569, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.144367218017578, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8781400918960571, "num_tokens": 500568922.0, "step": 13187 }, { "epoch": 1.6776491540516474, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.679619550704956, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.8612735867500305, "num_tokens": 500616445.0, "step": 13188 }, { "epoch": 1.677776364330238, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.947885274887085, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8779971599578857, "num_tokens": 500650692.0, "step": 13189 }, { "epoch": 1.6779035746088284, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.655937910079956, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8614547252655029, "num_tokens": 500692465.0, "step": 13190 }, { "epoch": 1.6780307848874187, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7611496448516846, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.858586311340332, "num_tokens": 500728715.0, "step": 13191 }, { "epoch": 1.6781579951660093, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5887725353240967, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8661953806877136, "num_tokens": 500768310.0, "step": 13192 }, { "epoch": 1.6782852054445998, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6252567768096924, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8747116327285767, "num_tokens": 500809777.0, "step": 13193 }, { "epoch": 1.6784124157231903, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.512075185775757, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8603806495666504, "num_tokens": 500851412.0, "step": 13194 }, { "epoch": 1.6785396260017809, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5460994243621826, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8522055149078369, "num_tokens": 500890896.0, "step": 13195 }, { "epoch": 1.6786668362803714, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4391913414001465, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8794723749160767, "num_tokens": 500926826.0, "step": 13196 }, { "epoch": 1.678794046558962, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.529820442199707, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8623242974281311, "num_tokens": 500960755.0, "step": 13197 }, { "epoch": 1.6789212568375524, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3721697330474854, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8692233562469482, "num_tokens": 500993177.0, "step": 13198 }, { "epoch": 1.679048467116143, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5227627754211426, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8697926998138428, "num_tokens": 501031391.0, "step": 13199 }, { "epoch": 1.6791756773947335, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2859292030334473, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8793798089027405, "num_tokens": 501072010.0, "step": 13200 }, { "epoch": 1.679302887673324, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.581136465072632, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8754985332489014, "num_tokens": 501112677.0, "step": 13201 }, { "epoch": 1.6794300979519146, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2110979557037354, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8728220462799072, "num_tokens": 501156634.0, "step": 13202 }, { "epoch": 1.679557308230505, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7236154079437256, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8718518018722534, "num_tokens": 501188194.0, "step": 13203 }, { "epoch": 1.6796845185090956, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1686978340148926, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8652389049530029, "num_tokens": 501216962.0, "step": 13204 }, { "epoch": 1.6798117287876861, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.871084690093994, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8841972947120667, "num_tokens": 501261964.0, "step": 13205 }, { "epoch": 1.6799389390662767, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.128845691680908, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8720479607582092, "num_tokens": 501303701.0, "step": 13206 }, { "epoch": 1.6800661493448672, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.100706100463867, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8791308403015137, "num_tokens": 501338531.0, "step": 13207 }, { "epoch": 1.6801933596234577, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2726433277130127, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8678120374679565, "num_tokens": 501375968.0, "step": 13208 }, { "epoch": 1.680320569902048, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2306623458862305, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8658320307731628, "num_tokens": 501414646.0, "step": 13209 }, { "epoch": 1.6804477801806386, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.602923631668091, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8682328462600708, "num_tokens": 501455035.0, "step": 13210 }, { "epoch": 1.680574990459229, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.059712886810303, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8677148818969727, "num_tokens": 501493198.0, "step": 13211 }, { "epoch": 1.6807022007378196, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6174747943878174, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8750892877578735, "num_tokens": 501534293.0, "step": 13212 }, { "epoch": 1.6808294110164101, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.880692720413208, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8686947226524353, "num_tokens": 501566614.0, "step": 13213 }, { "epoch": 1.6809566212950007, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.612535238265991, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8594851493835449, "num_tokens": 501607535.0, "step": 13214 }, { "epoch": 1.681083831573591, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6938085556030273, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8656103610992432, "num_tokens": 501644208.0, "step": 13215 }, { "epoch": 1.6812110418521815, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5596628189086914, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.857914388179779, "num_tokens": 501684980.0, "step": 13216 }, { "epoch": 1.681338252130772, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5561888217926025, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8714271783828735, "num_tokens": 501722603.0, "step": 13217 }, { "epoch": 1.6814654624093626, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5218043327331543, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.873227059841156, "num_tokens": 501755650.0, "step": 13218 }, { "epoch": 1.681592672687953, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4725420475006104, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8686218857765198, "num_tokens": 501788848.0, "step": 13219 }, { "epoch": 1.6817198829665436, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.465858221054077, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8676356077194214, "num_tokens": 501826396.0, "step": 13220 }, { "epoch": 1.6818470932451342, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4320333003997803, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8582507371902466, "num_tokens": 501863780.0, "step": 13221 }, { "epoch": 1.6819743035237247, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.423450469970703, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8664587736129761, "num_tokens": 501903275.0, "step": 13222 }, { "epoch": 1.6821015138023152, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4017772674560547, "learning_rate": 1e-05, "loss": 0.3757, "mean_token_accuracy": 0.8914837837219238, "num_tokens": 501940397.0, "step": 13223 }, { "epoch": 1.6822287240809057, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.400940418243408, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8701285719871521, "num_tokens": 501979814.0, "step": 13224 }, { "epoch": 1.6823559343594963, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3944084644317627, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8739601969718933, "num_tokens": 502017349.0, "step": 13225 }, { "epoch": 1.6824831446380868, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3734006881713867, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8611053228378296, "num_tokens": 502062441.0, "step": 13226 }, { "epoch": 1.6826103549166773, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3817105293273926, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8695693612098694, "num_tokens": 502099897.0, "step": 13227 }, { "epoch": 1.6827375651952678, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3574302196502686, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8733841776847839, "num_tokens": 502139682.0, "step": 13228 }, { "epoch": 1.6828647754738584, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.399789571762085, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8621687293052673, "num_tokens": 502179534.0, "step": 13229 }, { "epoch": 1.682991985752449, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.331515073776245, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8751912117004395, "num_tokens": 502218031.0, "step": 13230 }, { "epoch": 1.6831191960310394, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.425475597381592, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8765602111816406, "num_tokens": 502251409.0, "step": 13231 }, { "epoch": 1.68324640630963, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.303190231323242, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8745189905166626, "num_tokens": 502285757.0, "step": 13232 }, { "epoch": 1.6833736165882205, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.454620838165283, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.872719407081604, "num_tokens": 502325247.0, "step": 13233 }, { "epoch": 1.6835008268668108, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2477972507476807, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8708242177963257, "num_tokens": 502367691.0, "step": 13234 }, { "epoch": 1.6836280371454013, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5288302898406982, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8580064177513123, "num_tokens": 502409084.0, "step": 13235 }, { "epoch": 1.6837552474239919, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.175177574157715, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8726784586906433, "num_tokens": 502448478.0, "step": 13236 }, { "epoch": 1.6838824577025824, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6654510498046875, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8636373281478882, "num_tokens": 502481459.0, "step": 13237 }, { "epoch": 1.684009667981173, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.08760142326355, "learning_rate": 1e-05, "loss": 0.5189, "mean_token_accuracy": 0.8527340292930603, "num_tokens": 502522300.0, "step": 13238 }, { "epoch": 1.6841368782597634, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8852527141571045, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8732832670211792, "num_tokens": 502560257.0, "step": 13239 }, { "epoch": 1.6842640885383537, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0216526985168457, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8604868054389954, "num_tokens": 502602815.0, "step": 13240 }, { "epoch": 1.6843912988169443, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.232973575592041, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8728660345077515, "num_tokens": 502640067.0, "step": 13241 }, { "epoch": 1.6845185090955348, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.1976943016052246, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8649011254310608, "num_tokens": 502680351.0, "step": 13242 }, { "epoch": 1.6846457193741253, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.540073394775391, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8777285814285278, "num_tokens": 502715279.0, "step": 13243 }, { "epoch": 1.6847729296527159, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.6465816497802734, "learning_rate": 1e-05, "loss": 0.5046, "mean_token_accuracy": 0.8583557605743408, "num_tokens": 502762078.0, "step": 13244 }, { "epoch": 1.6849001399313064, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.4577717781066895, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8837900757789612, "num_tokens": 502799886.0, "step": 13245 }, { "epoch": 1.685027350209897, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.9360036849975586, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8603715300559998, "num_tokens": 502838414.0, "step": 13246 }, { "epoch": 1.6851545604884874, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.175694465637207, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8609521985054016, "num_tokens": 502879504.0, "step": 13247 }, { "epoch": 1.685281770767078, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.859340190887451, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8899049758911133, "num_tokens": 502919167.0, "step": 13248 }, { "epoch": 1.6854089810456685, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9642558097839355, "learning_rate": 1e-05, "loss": 0.3925, "mean_token_accuracy": 0.8934680223464966, "num_tokens": 502956548.0, "step": 13249 }, { "epoch": 1.685536191324259, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.71227765083313, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8752871751785278, "num_tokens": 502993001.0, "step": 13250 }, { "epoch": 1.6856634016028496, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8354556560516357, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8721768856048584, "num_tokens": 503030690.0, "step": 13251 }, { "epoch": 1.68579061188144, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.561342477798462, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8650727272033691, "num_tokens": 503074010.0, "step": 13252 }, { "epoch": 1.6859178221600306, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.786959648132324, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8725775480270386, "num_tokens": 503111077.0, "step": 13253 }, { "epoch": 1.6860450324386211, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.431427240371704, "learning_rate": 1e-05, "loss": 0.385, "mean_token_accuracy": 0.8935732245445251, "num_tokens": 503146594.0, "step": 13254 }, { "epoch": 1.6861722427172117, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.8084795475006104, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8853652477264404, "num_tokens": 503181470.0, "step": 13255 }, { "epoch": 1.6862994529958022, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.332834243774414, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8835203647613525, "num_tokens": 503215839.0, "step": 13256 }, { "epoch": 1.6864266632743927, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.911262035369873, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8795602321624756, "num_tokens": 503253131.0, "step": 13257 }, { "epoch": 1.686553873552983, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.303090810775757, "learning_rate": 1e-05, "loss": 0.5171, "mean_token_accuracy": 0.8487440943717957, "num_tokens": 503290331.0, "step": 13258 }, { "epoch": 1.6866810838315736, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.062136173248291, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8570308685302734, "num_tokens": 503328551.0, "step": 13259 }, { "epoch": 1.686808294110164, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.386094331741333, "learning_rate": 1e-05, "loss": 0.5892, "mean_token_accuracy": 0.8355845212936401, "num_tokens": 503365288.0, "step": 13260 }, { "epoch": 1.6869355043887546, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.118135929107666, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8780882954597473, "num_tokens": 503402941.0, "step": 13261 }, { "epoch": 1.6870627146673451, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.523406982421875, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8607777953147888, "num_tokens": 503443174.0, "step": 13262 }, { "epoch": 1.6871899249459357, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.0333356857299805, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8754551410675049, "num_tokens": 503481790.0, "step": 13263 }, { "epoch": 1.687317135224526, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.629680633544922, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8730484247207642, "num_tokens": 503518386.0, "step": 13264 }, { "epoch": 1.6874443455031165, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.829134702682495, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8765690326690674, "num_tokens": 503558985.0, "step": 13265 }, { "epoch": 1.687571555781707, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6327548027038574, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8775607347488403, "num_tokens": 503602423.0, "step": 13266 }, { "epoch": 1.6876987660602976, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.635826587677002, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8781833648681641, "num_tokens": 503634300.0, "step": 13267 }, { "epoch": 1.687825976338888, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.633676528930664, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.859921932220459, "num_tokens": 503668225.0, "step": 13268 }, { "epoch": 1.6879531866174786, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4503543376922607, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8558881282806396, "num_tokens": 503705158.0, "step": 13269 }, { "epoch": 1.6880803968960691, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.606834888458252, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8715847730636597, "num_tokens": 503741322.0, "step": 13270 }, { "epoch": 1.6882076071746597, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3507916927337646, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8831223845481873, "num_tokens": 503777932.0, "step": 13271 }, { "epoch": 1.6883348174532502, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6069586277008057, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8803071975708008, "num_tokens": 503810175.0, "step": 13272 }, { "epoch": 1.6884620277318407, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.265770435333252, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8589490652084351, "num_tokens": 503844498.0, "step": 13273 }, { "epoch": 1.6885892380104313, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6721253395080566, "learning_rate": 1e-05, "loss": 0.3939, "mean_token_accuracy": 0.8898287415504456, "num_tokens": 503878898.0, "step": 13274 }, { "epoch": 1.6887164482890218, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.176783323287964, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8811662197113037, "num_tokens": 503915473.0, "step": 13275 }, { "epoch": 1.6888436585676123, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8079636096954346, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.872575044631958, "num_tokens": 503954234.0, "step": 13276 }, { "epoch": 1.6889708688462028, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.15820574760437, "learning_rate": 1e-05, "loss": 0.5214, "mean_token_accuracy": 0.850050151348114, "num_tokens": 503988740.0, "step": 13277 }, { "epoch": 1.6890980791247934, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.978111743927002, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8737936019897461, "num_tokens": 504025755.0, "step": 13278 }, { "epoch": 1.689225289403384, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1866402626037598, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8559659123420715, "num_tokens": 504062639.0, "step": 13279 }, { "epoch": 1.6893524996819744, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.178462982177734, "learning_rate": 1e-05, "loss": 0.5258, "mean_token_accuracy": 0.85111403465271, "num_tokens": 504099645.0, "step": 13280 }, { "epoch": 1.689479709960565, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3634328842163086, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8813462257385254, "num_tokens": 504137678.0, "step": 13281 }, { "epoch": 1.6896069202391555, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.230058193206787, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8707668781280518, "num_tokens": 504174037.0, "step": 13282 }, { "epoch": 1.6897341305177458, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5963134765625, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8545533418655396, "num_tokens": 504220117.0, "step": 13283 }, { "epoch": 1.6898613407963363, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.083437442779541, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8778695464134216, "num_tokens": 504259104.0, "step": 13284 }, { "epoch": 1.6899885510749268, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.670536994934082, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8708416223526001, "num_tokens": 504302867.0, "step": 13285 }, { "epoch": 1.6901157613535174, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.878403902053833, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8802546262741089, "num_tokens": 504337555.0, "step": 13286 }, { "epoch": 1.690242971632108, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.648451566696167, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.881737470626831, "num_tokens": 504372334.0, "step": 13287 }, { "epoch": 1.6903701819106984, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6847403049468994, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8757815957069397, "num_tokens": 504412894.0, "step": 13288 }, { "epoch": 1.6904973921892887, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5873122215270996, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8736583590507507, "num_tokens": 504452919.0, "step": 13289 }, { "epoch": 1.6906246024678793, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.553654909133911, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8730634450912476, "num_tokens": 504490694.0, "step": 13290 }, { "epoch": 1.6907518127464698, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.53511381149292, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.867003321647644, "num_tokens": 504526606.0, "step": 13291 }, { "epoch": 1.6908790230250603, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.45810866355896, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8753968477249146, "num_tokens": 504568885.0, "step": 13292 }, { "epoch": 1.6910062333036509, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.47811222076416, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8859990835189819, "num_tokens": 504604947.0, "step": 13293 }, { "epoch": 1.6911334435822414, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.435640811920166, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.86406409740448, "num_tokens": 504643937.0, "step": 13294 }, { "epoch": 1.691260653860832, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4258649349212646, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8705297112464905, "num_tokens": 504684298.0, "step": 13295 }, { "epoch": 1.6913878641394224, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4174485206604004, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8572945594787598, "num_tokens": 504723578.0, "step": 13296 }, { "epoch": 1.691515074418013, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.382286787033081, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8635966181755066, "num_tokens": 504768190.0, "step": 13297 }, { "epoch": 1.6916422846966035, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.4105870723724365, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8813278675079346, "num_tokens": 504807656.0, "step": 13298 }, { "epoch": 1.691769494975194, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.354201555252075, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8693905472755432, "num_tokens": 504846631.0, "step": 13299 }, { "epoch": 1.6918967052537845, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.422414541244507, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.872584342956543, "num_tokens": 504890451.0, "step": 13300 }, { "epoch": 1.692023915532375, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.319549083709717, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8852508068084717, "num_tokens": 504931863.0, "step": 13301 }, { "epoch": 1.6921511258109656, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4604640007019043, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8781299591064453, "num_tokens": 504970247.0, "step": 13302 }, { "epoch": 1.6922783360895561, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.274923801422119, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8769810199737549, "num_tokens": 505009987.0, "step": 13303 }, { "epoch": 1.6924055463681467, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.513700008392334, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8767406940460205, "num_tokens": 505051933.0, "step": 13304 }, { "epoch": 1.6925327566467372, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2169220447540283, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8735915422439575, "num_tokens": 505086216.0, "step": 13305 }, { "epoch": 1.6926599669253277, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.605999708175659, "learning_rate": 1e-05, "loss": 0.38, "mean_token_accuracy": 0.8921105861663818, "num_tokens": 505123040.0, "step": 13306 }, { "epoch": 1.692787177203918, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.139371633529663, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8740660548210144, "num_tokens": 505158546.0, "step": 13307 }, { "epoch": 1.6929143874825086, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7861132621765137, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8729212880134583, "num_tokens": 505195433.0, "step": 13308 }, { "epoch": 1.693041597761099, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.38690185546875e-05, "grad_norm": 3.0618233680725098, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8759256601333618, "num_tokens": 505232411.0, "step": 13309 }, { "epoch": 1.6931688080396896, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.068809509277344, "learning_rate": 1e-05, "loss": 0.5091, "mean_token_accuracy": 0.8575727939605713, "num_tokens": 505270590.0, "step": 13310 }, { "epoch": 1.6932960183182801, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.0997133255004883, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8731226921081543, "num_tokens": 505307705.0, "step": 13311 }, { "epoch": 1.6934232285968707, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.414120197296143, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8718501329421997, "num_tokens": 505342676.0, "step": 13312 }, { "epoch": 1.693550438875461, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.454815149307251, "learning_rate": 1e-05, "loss": 0.5043, "mean_token_accuracy": 0.8601648807525635, "num_tokens": 505387058.0, "step": 13313 }, { "epoch": 1.6936776491540515, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.488990783691406, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.862921416759491, "num_tokens": 505429584.0, "step": 13314 }, { "epoch": 1.693804859432642, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.8349697589874268, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8814539313316345, "num_tokens": 505464619.0, "step": 13315 }, { "epoch": 1.6939320697112326, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.236856937408447, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8666977882385254, "num_tokens": 505507390.0, "step": 13316 }, { "epoch": 1.694059279989823, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.868887424468994, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8704341650009155, "num_tokens": 505543599.0, "step": 13317 }, { "epoch": 1.6941864902684136, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.01644229888916, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8688207268714905, "num_tokens": 505576283.0, "step": 13318 }, { "epoch": 1.6943137005470041, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.741722822189331, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8656336069107056, "num_tokens": 505615244.0, "step": 13319 }, { "epoch": 1.6944409108255947, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8512628078460693, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8618175983428955, "num_tokens": 505653505.0, "step": 13320 }, { "epoch": 1.6945681211041852, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6310884952545166, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8722166419029236, "num_tokens": 505678666.0, "step": 13321 }, { "epoch": 1.6946953313827757, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.746356248855591, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8619893789291382, "num_tokens": 505718191.0, "step": 13322 }, { "epoch": 1.6948225416613663, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4751505851745605, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8856594562530518, "num_tokens": 505759174.0, "step": 13323 }, { "epoch": 1.6949497519399568, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.722811460494995, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.856752872467041, "num_tokens": 505804483.0, "step": 13324 }, { "epoch": 1.6950769622185473, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3723883628845215, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8649554252624512, "num_tokens": 505839767.0, "step": 13325 }, { "epoch": 1.6952041724971378, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.782886505126953, "learning_rate": 1e-05, "loss": 0.5204, "mean_token_accuracy": 0.8520829081535339, "num_tokens": 505880207.0, "step": 13326 }, { "epoch": 1.6953313827757284, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2823612689971924, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8753370046615601, "num_tokens": 505914686.0, "step": 13327 }, { "epoch": 1.695458593054319, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8987579345703125, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8733914494514465, "num_tokens": 505959029.0, "step": 13328 }, { "epoch": 1.6955858033329094, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2497785091400146, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8595529794692993, "num_tokens": 505997501.0, "step": 13329 }, { "epoch": 1.6957130136115, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.081096172332764, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8589354753494263, "num_tokens": 506029097.0, "step": 13330 }, { "epoch": 1.6958402238900905, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.352651357650757, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8837942481040955, "num_tokens": 506065912.0, "step": 13331 }, { "epoch": 1.6959674341686808, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.146651744842529, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8764827251434326, "num_tokens": 506108843.0, "step": 13332 }, { "epoch": 1.6960946444472713, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5352635383605957, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8835675120353699, "num_tokens": 506147505.0, "step": 13333 }, { "epoch": 1.6962218547258618, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.0489678382873535, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.876803994178772, "num_tokens": 506186491.0, "step": 13334 }, { "epoch": 1.6963490650044524, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6577911376953125, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8778027296066284, "num_tokens": 506223343.0, "step": 13335 }, { "epoch": 1.696476275283043, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8245503902435303, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8771533370018005, "num_tokens": 506262328.0, "step": 13336 }, { "epoch": 1.6966034855616334, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6725504398345947, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8603092432022095, "num_tokens": 506294758.0, "step": 13337 }, { "epoch": 1.6967306958402237, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5900752544403076, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8799991607666016, "num_tokens": 506325423.0, "step": 13338 }, { "epoch": 1.6968579061188143, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6411869525909424, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8605313301086426, "num_tokens": 506367722.0, "step": 13339 }, { "epoch": 1.6969851163974048, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4040918350219727, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8608959913253784, "num_tokens": 506404638.0, "step": 13340 }, { "epoch": 1.6971123266759953, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6568057537078857, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8704752922058105, "num_tokens": 506440976.0, "step": 13341 }, { "epoch": 1.6972395369545858, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2869176864624023, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8662777543067932, "num_tokens": 506475392.0, "step": 13342 }, { "epoch": 1.6973667472331764, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6933510303497314, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8720923662185669, "num_tokens": 506510933.0, "step": 13343 }, { "epoch": 1.697493957511767, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2090420722961426, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.86476731300354, "num_tokens": 506550217.0, "step": 13344 }, { "epoch": 1.6976211677903574, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8182852268218994, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8753834962844849, "num_tokens": 506583836.0, "step": 13345 }, { "epoch": 1.697748378068948, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.169553756713867, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8776156902313232, "num_tokens": 506616889.0, "step": 13346 }, { "epoch": 1.6978755883475385, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9797921180725098, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.87712562084198, "num_tokens": 506657198.0, "step": 13347 }, { "epoch": 1.698002798626129, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2077572345733643, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.88031005859375, "num_tokens": 506696439.0, "step": 13348 }, { "epoch": 1.6981300089047195, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.155854225158691, "learning_rate": 1e-05, "loss": 0.4993, "mean_token_accuracy": 0.8588420152664185, "num_tokens": 506736140.0, "step": 13349 }, { "epoch": 1.69825721918331, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4012510776519775, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8792077302932739, "num_tokens": 506770976.0, "step": 13350 }, { "epoch": 1.6983844294619006, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.172333240509033, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8726316690444946, "num_tokens": 506804463.0, "step": 13351 }, { "epoch": 1.6985116397404911, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5676286220550537, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.88734370470047, "num_tokens": 506839498.0, "step": 13352 }, { "epoch": 1.6986388500190817, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.045051097869873, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8666382431983948, "num_tokens": 506875899.0, "step": 13353 }, { "epoch": 1.6987660602976722, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.641242265701294, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8646121025085449, "num_tokens": 506914760.0, "step": 13354 }, { "epoch": 1.6988932705762627, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.859171152114868, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8749428987503052, "num_tokens": 506950022.0, "step": 13355 }, { "epoch": 1.699020480854853, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5889904499053955, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8752449750900269, "num_tokens": 506993349.0, "step": 13356 }, { "epoch": 1.6991476911334435, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.705404043197632, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8657263517379761, "num_tokens": 507037519.0, "step": 13357 }, { "epoch": 1.699274901412034, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.551100492477417, "learning_rate": 1e-05, "loss": 0.5385, "mean_token_accuracy": 0.8445801138877869, "num_tokens": 507071932.0, "step": 13358 }, { "epoch": 1.6994021116906246, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.594344139099121, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.871401309967041, "num_tokens": 507108140.0, "step": 13359 }, { "epoch": 1.6995293219692151, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4612793922424316, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.864837646484375, "num_tokens": 507148241.0, "step": 13360 }, { "epoch": 1.6996565322478057, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.563169002532959, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8712537884712219, "num_tokens": 507183200.0, "step": 13361 }, { "epoch": 1.699783742526396, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3888320922851562, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8679622411727905, "num_tokens": 507219036.0, "step": 13362 }, { "epoch": 1.6999109528049865, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5603408813476562, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8825421333312988, "num_tokens": 507254767.0, "step": 13363 }, { "epoch": 1.700038163083577, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3116607666015625, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8848684430122375, "num_tokens": 507289945.0, "step": 13364 }, { "epoch": 1.7001653733621676, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.608433961868286, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8714621663093567, "num_tokens": 507328350.0, "step": 13365 }, { "epoch": 1.700292583640758, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2454991340637207, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8580286502838135, "num_tokens": 507367151.0, "step": 13366 }, { "epoch": 1.7004197939193486, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.717177391052246, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8766500949859619, "num_tokens": 507404181.0, "step": 13367 }, { "epoch": 1.7005470041979391, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.186009407043457, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8828151226043701, "num_tokens": 507438402.0, "step": 13368 }, { "epoch": 1.7006742144765297, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8638217449188232, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8730091452598572, "num_tokens": 507478091.0, "step": 13369 }, { "epoch": 1.7008014247551202, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.17878794670105, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.864459216594696, "num_tokens": 507518285.0, "step": 13370 }, { "epoch": 1.7009286350337107, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.044532299041748, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8705228567123413, "num_tokens": 507555250.0, "step": 13371 }, { "epoch": 1.7010558453123013, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2889676094055176, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8765281438827515, "num_tokens": 507591255.0, "step": 13372 }, { "epoch": 1.7011830555908918, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.146854877471924, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8767038583755493, "num_tokens": 507623805.0, "step": 13373 }, { "epoch": 1.7013102658694823, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4830493927001953, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8616131544113159, "num_tokens": 507660312.0, "step": 13374 }, { "epoch": 1.7014374761480728, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0547051429748535, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8737059831619263, "num_tokens": 507700132.0, "step": 13375 }, { "epoch": 1.7015646864266634, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5953023433685303, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8715627789497375, "num_tokens": 507737521.0, "step": 13376 }, { "epoch": 1.701691896705254, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.860322952270508, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8907520771026611, "num_tokens": 507779327.0, "step": 13377 }, { "epoch": 1.7018191069838444, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.608614683151245, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8730169534683228, "num_tokens": 507821899.0, "step": 13378 }, { "epoch": 1.701946317262435, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.65488338470459, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8863489627838135, "num_tokens": 507862946.0, "step": 13379 }, { "epoch": 1.7020735275410255, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.603062152862549, "learning_rate": 1e-05, "loss": 0.5157, "mean_token_accuracy": 0.8514112234115601, "num_tokens": 507898409.0, "step": 13380 }, { "epoch": 1.7022007378196158, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.476933240890503, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8771904706954956, "num_tokens": 507936670.0, "step": 13381 }, { "epoch": 1.7023279480982063, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.590467929840088, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8705135583877563, "num_tokens": 507976160.0, "step": 13382 }, { "epoch": 1.7024551583767968, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3505942821502686, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8733783960342407, "num_tokens": 508015500.0, "step": 13383 }, { "epoch": 1.7025823686553874, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6032660007476807, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8853964805603027, "num_tokens": 508050860.0, "step": 13384 }, { "epoch": 1.702709578933978, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2809267044067383, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8693042993545532, "num_tokens": 508085794.0, "step": 13385 }, { "epoch": 1.7028367892125684, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6448051929473877, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8844956159591675, "num_tokens": 508122731.0, "step": 13386 }, { "epoch": 1.7029639994911587, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2032053470611572, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8683154582977295, "num_tokens": 508168513.0, "step": 13387 }, { "epoch": 1.7030912097697493, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7542312145233154, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8882130980491638, "num_tokens": 508209758.0, "step": 13388 }, { "epoch": 1.7032184200483398, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1561715602874756, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8833028078079224, "num_tokens": 508248172.0, "step": 13389 }, { "epoch": 1.7033456303269303, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9191060066223145, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8651383519172668, "num_tokens": 508288694.0, "step": 13390 }, { "epoch": 1.7034728406055208, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1731183528900146, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8645858764648438, "num_tokens": 508329021.0, "step": 13391 }, { "epoch": 1.7036000508841114, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.106637954711914, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8555344939231873, "num_tokens": 508363706.0, "step": 13392 }, { "epoch": 1.703727261162702, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3068184852600098, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8823825120925903, "num_tokens": 508399879.0, "step": 13393 }, { "epoch": 1.7038544714412924, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.185368061065674, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8643545508384705, "num_tokens": 508440172.0, "step": 13394 }, { "epoch": 1.703981681719883, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5253541469573975, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8678640127182007, "num_tokens": 508475531.0, "step": 13395 }, { "epoch": 1.7041088919984735, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0974602699279785, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8790587186813354, "num_tokens": 508509290.0, "step": 13396 }, { "epoch": 1.704236102277064, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6113975048065186, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8753383159637451, "num_tokens": 508545052.0, "step": 13397 }, { "epoch": 1.7043633125556545, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9439804553985596, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8754761219024658, "num_tokens": 508577922.0, "step": 13398 }, { "epoch": 1.704490522834245, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.610037326812744, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8762516975402832, "num_tokens": 508618135.0, "step": 13399 }, { "epoch": 1.7046177331128356, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7741096019744873, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8600021004676819, "num_tokens": 508656665.0, "step": 13400 }, { "epoch": 1.7047449433914261, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5626637935638428, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8720278739929199, "num_tokens": 508693898.0, "step": 13401 }, { "epoch": 1.7048721536700167, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6469218730926514, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.8637741208076477, "num_tokens": 508726865.0, "step": 13402 }, { "epoch": 1.7049993639486072, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.502969980239868, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8744505643844604, "num_tokens": 508762040.0, "step": 13403 }, { "epoch": 1.7051265742271977, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5481646060943604, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8700338006019592, "num_tokens": 508803288.0, "step": 13404 }, { "epoch": 1.705253784505788, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.438066244125366, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.877474308013916, "num_tokens": 508840729.0, "step": 13405 }, { "epoch": 1.7053809947843785, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5174343585968018, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8806070685386658, "num_tokens": 508874219.0, "step": 13406 }, { "epoch": 1.705508205062969, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3892171382904053, "learning_rate": 1e-05, "loss": 0.3913, "mean_token_accuracy": 0.8905364274978638, "num_tokens": 508911746.0, "step": 13407 }, { "epoch": 1.7056354153415596, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5107204914093018, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8771988153457642, "num_tokens": 508944996.0, "step": 13408 }, { "epoch": 1.7057626256201501, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.323439121246338, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8639425039291382, "num_tokens": 508986416.0, "step": 13409 }, { "epoch": 1.7058898358987407, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.616661548614502, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8700966238975525, "num_tokens": 509023690.0, "step": 13410 }, { "epoch": 1.706017046177331, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2467002868652344, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8742655515670776, "num_tokens": 509065405.0, "step": 13411 }, { "epoch": 1.7061442564559215, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6854207515716553, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8680530190467834, "num_tokens": 509104871.0, "step": 13412 }, { "epoch": 1.706271466734512, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.217662811279297, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8543825149536133, "num_tokens": 509140938.0, "step": 13413 }, { "epoch": 1.7063986770131025, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7966935634613037, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8755908608436584, "num_tokens": 509175324.0, "step": 13414 }, { "epoch": 1.706525887291693, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1495754718780518, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8740787506103516, "num_tokens": 509214983.0, "step": 13415 }, { "epoch": 1.7066530975702836, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.987095594406128, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8645967245101929, "num_tokens": 509249418.0, "step": 13416 }, { "epoch": 1.7067803078488741, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.20664644241333, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8802495002746582, "num_tokens": 509282884.0, "step": 13417 }, { "epoch": 1.7069075181274647, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.153409957885742, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.875812828540802, "num_tokens": 509317880.0, "step": 13418 }, { "epoch": 1.7070347284060552, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.3980233669281006, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8638935685157776, "num_tokens": 509355760.0, "step": 13419 }, { "epoch": 1.7071619386846457, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.154310703277588, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8740971088409424, "num_tokens": 509393088.0, "step": 13420 }, { "epoch": 1.7072891489632362, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.5747971534729004, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8651759624481201, "num_tokens": 509431568.0, "step": 13421 }, { "epoch": 1.7074163592418268, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.995051622390747, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8682618141174316, "num_tokens": 509471044.0, "step": 13422 }, { "epoch": 1.7075435695204173, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6279942989349365, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8804594874382019, "num_tokens": 509507746.0, "step": 13423 }, { "epoch": 1.7076707797990078, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.7857649326324463, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8851678967475891, "num_tokens": 509545753.0, "step": 13424 }, { "epoch": 1.7077979900775984, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6031606197357178, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8745989799499512, "num_tokens": 509581155.0, "step": 13425 }, { "epoch": 1.7079252003561889, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6129508018493652, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8650735020637512, "num_tokens": 509615540.0, "step": 13426 }, { "epoch": 1.7080524106347794, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5541493892669678, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8678172826766968, "num_tokens": 509651257.0, "step": 13427 }, { "epoch": 1.70817962091337, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.500459671020508, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8548676371574402, "num_tokens": 509688088.0, "step": 13428 }, { "epoch": 1.7083068311919605, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.515998601913452, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8737131357192993, "num_tokens": 509721122.0, "step": 13429 }, { "epoch": 1.7084340414705508, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4275307655334473, "learning_rate": 1e-05, "loss": 0.5094, "mean_token_accuracy": 0.851751446723938, "num_tokens": 509756711.0, "step": 13430 }, { "epoch": 1.7085612517491413, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4798686504364014, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.871077835559845, "num_tokens": 509793868.0, "step": 13431 }, { "epoch": 1.7086884620277318, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3747565746307373, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8812639713287354, "num_tokens": 509828741.0, "step": 13432 }, { "epoch": 1.7088156723063224, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4642460346221924, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.880172848701477, "num_tokens": 509866534.0, "step": 13433 }, { "epoch": 1.708942882584913, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3450379371643066, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8702782392501831, "num_tokens": 509897664.0, "step": 13434 }, { "epoch": 1.7090700928635034, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.48514723777771, "learning_rate": 1e-05, "loss": 0.5374, "mean_token_accuracy": 0.8463945388793945, "num_tokens": 509933349.0, "step": 13435 }, { "epoch": 1.7091973031420937, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2963314056396484, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8682465553283691, "num_tokens": 509970912.0, "step": 13436 }, { "epoch": 1.7093245134206843, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.522874355316162, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8651629686355591, "num_tokens": 510007431.0, "step": 13437 }, { "epoch": 1.7094517236992748, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2335400581359863, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8672427535057068, "num_tokens": 510044109.0, "step": 13438 }, { "epoch": 1.7095789339778653, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.598313093185425, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.855063796043396, "num_tokens": 510084665.0, "step": 13439 }, { "epoch": 1.7097061442564558, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1706738471984863, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8573832511901855, "num_tokens": 510123108.0, "step": 13440 }, { "epoch": 1.7098333545350464, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.732363700866699, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8657915592193604, "num_tokens": 510167918.0, "step": 13441 }, { "epoch": 1.709960564813637, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.100020408630371, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8807103633880615, "num_tokens": 510204555.0, "step": 13442 }, { "epoch": 1.7100877750922274, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9623312950134277, "learning_rate": 1e-05, "loss": 0.5188, "mean_token_accuracy": 0.8510780930519104, "num_tokens": 510239508.0, "step": 13443 }, { "epoch": 1.710214985370818, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.089585781097412, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8737050294876099, "num_tokens": 510275246.0, "step": 13444 }, { "epoch": 1.7103421956494085, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.258183479309082, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8729796409606934, "num_tokens": 510316558.0, "step": 13445 }, { "epoch": 1.710469405927999, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3044979572296143, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8774930238723755, "num_tokens": 510351784.0, "step": 13446 }, { "epoch": 1.7105966162065895, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.428526878356934, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8765480518341064, "num_tokens": 510391044.0, "step": 13447 }, { "epoch": 1.71072382648518, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.6695756912231445, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8720241189002991, "num_tokens": 510431199.0, "step": 13448 }, { "epoch": 1.7108510367637706, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.285616397857666, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8749099969863892, "num_tokens": 510471629.0, "step": 13449 }, { "epoch": 1.7109782470423611, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.811544418334961, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8823300004005432, "num_tokens": 510511546.0, "step": 13450 }, { "epoch": 1.7111054573209517, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.067233562469482, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8623979091644287, "num_tokens": 510548942.0, "step": 13451 }, { "epoch": 1.7112326675995422, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.773411750793457, "learning_rate": 1e-05, "loss": 0.5456, "mean_token_accuracy": 0.8450096845626831, "num_tokens": 510589855.0, "step": 13452 }, { "epoch": 1.7113598778781327, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8575239181518555, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8751196265220642, "num_tokens": 510628960.0, "step": 13453 }, { "epoch": 1.711487088156723, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6384942531585693, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.866316556930542, "num_tokens": 510671107.0, "step": 13454 }, { "epoch": 1.7116142984353135, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7446346282958984, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8684449791908264, "num_tokens": 510708957.0, "step": 13455 }, { "epoch": 1.711741508713904, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5175137519836426, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8726534843444824, "num_tokens": 510746150.0, "step": 13456 }, { "epoch": 1.7118687189924946, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.70371150970459, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8619052767753601, "num_tokens": 510775810.0, "step": 13457 }, { "epoch": 1.7119959292710851, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3992486000061035, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.858690619468689, "num_tokens": 510817438.0, "step": 13458 }, { "epoch": 1.7121231395496757, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7128818035125732, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8683587312698364, "num_tokens": 510854529.0, "step": 13459 }, { "epoch": 1.712250349828266, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.305664300918579, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8700451254844666, "num_tokens": 510891568.0, "step": 13460 }, { "epoch": 1.7123775601068565, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.797239303588867, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8744593262672424, "num_tokens": 510928424.0, "step": 13461 }, { "epoch": 1.712504770385447, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2374253273010254, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8656340837478638, "num_tokens": 510971770.0, "step": 13462 }, { "epoch": 1.7126319806640375, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9707841873168945, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.859958291053772, "num_tokens": 511004087.0, "step": 13463 }, { "epoch": 1.712759190942628, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2748842239379883, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8627884387969971, "num_tokens": 511046494.0, "step": 13464 }, { "epoch": 1.7128864012212186, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.093480110168457, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8643532991409302, "num_tokens": 511088729.0, "step": 13465 }, { "epoch": 1.7130136114998091, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.4081106185913086, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8749220967292786, "num_tokens": 511127581.0, "step": 13466 }, { "epoch": 1.7131408217783997, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.110741138458252, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8634697198867798, "num_tokens": 511166072.0, "step": 13467 }, { "epoch": 1.7132680320569902, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.574136734008789, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8697474002838135, "num_tokens": 511203529.0, "step": 13468 }, { "epoch": 1.7133952423355807, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.953728437423706, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8694810271263123, "num_tokens": 511245846.0, "step": 13469 }, { "epoch": 1.7135224526141712, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.630668878555298, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.8650806546211243, "num_tokens": 511287280.0, "step": 13470 }, { "epoch": 1.7136496628927618, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.735135316848755, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8726601600646973, "num_tokens": 511330905.0, "step": 13471 }, { "epoch": 1.7137768731713523, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.623459577560425, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8587287664413452, "num_tokens": 511369872.0, "step": 13472 }, { "epoch": 1.7139040834499428, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.529238700866699, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8782587051391602, "num_tokens": 511410638.0, "step": 13473 }, { "epoch": 1.7140312937285334, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.61301589012146, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8733964562416077, "num_tokens": 511445517.0, "step": 13474 }, { "epoch": 1.7141585040071239, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.390080213546753, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.8550541400909424, "num_tokens": 511482749.0, "step": 13475 }, { "epoch": 1.7142857142857144, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.618422031402588, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.860015332698822, "num_tokens": 511517001.0, "step": 13476 }, { "epoch": 1.714412924564305, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2906017303466797, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.864820122718811, "num_tokens": 511559437.0, "step": 13477 }, { "epoch": 1.7145401348428955, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6772260665893555, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8709887862205505, "num_tokens": 511593713.0, "step": 13478 }, { "epoch": 1.7146673451214858, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.214431047439575, "learning_rate": 1e-05, "loss": 0.4919, "mean_token_accuracy": 0.857703447341919, "num_tokens": 511633515.0, "step": 13479 }, { "epoch": 1.7147945554000763, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.780524492263794, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8691099882125854, "num_tokens": 511671539.0, "step": 13480 }, { "epoch": 1.7149217656786668, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.160568952560425, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8782951235771179, "num_tokens": 511706972.0, "step": 13481 }, { "epoch": 1.7150489759572574, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9510579109191895, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8646634221076965, "num_tokens": 511740186.0, "step": 13482 }, { "epoch": 1.7151761862358479, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1898252964019775, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8696466684341431, "num_tokens": 511779755.0, "step": 13483 }, { "epoch": 1.7153033965144384, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.128291606903076, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8621799945831299, "num_tokens": 511822531.0, "step": 13484 }, { "epoch": 1.7154306067930287, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3402934074401855, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8681159019470215, "num_tokens": 511852950.0, "step": 13485 }, { "epoch": 1.7155578170716193, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.190120220184326, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8772588968276978, "num_tokens": 511890251.0, "step": 13486 }, { "epoch": 1.7156850273502098, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5352206230163574, "learning_rate": 1e-05, "loss": 0.5126, "mean_token_accuracy": 0.8536912798881531, "num_tokens": 511933103.0, "step": 13487 }, { "epoch": 1.7158122376288003, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.077908992767334, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8776357173919678, "num_tokens": 511975682.0, "step": 13488 }, { "epoch": 1.7159394479073908, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.628486156463623, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8693820834159851, "num_tokens": 512014899.0, "step": 13489 }, { "epoch": 1.7160666581859814, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9066624641418457, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8593600988388062, "num_tokens": 512054951.0, "step": 13490 }, { "epoch": 1.716193868464572, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.614809036254883, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8694499731063843, "num_tokens": 512094273.0, "step": 13491 }, { "epoch": 1.7163210787431624, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.72883677482605, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8658890724182129, "num_tokens": 512133493.0, "step": 13492 }, { "epoch": 1.716448289021753, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5693721771240234, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8708363771438599, "num_tokens": 512171527.0, "step": 13493 }, { "epoch": 1.7165754993003435, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5898962020874023, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8738099336624146, "num_tokens": 512211106.0, "step": 13494 }, { "epoch": 1.716702709578934, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5094239711761475, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8746049404144287, "num_tokens": 512247444.0, "step": 13495 }, { "epoch": 1.7168299198575245, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.52571177482605, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8622093200683594, "num_tokens": 512293912.0, "step": 13496 }, { "epoch": 1.716957130136115, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.447277307510376, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8813542127609253, "num_tokens": 512334149.0, "step": 13497 }, { "epoch": 1.7170843404147056, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4871740341186523, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.875504732131958, "num_tokens": 512371161.0, "step": 13498 }, { "epoch": 1.7172115506932961, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.400580406188965, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8573040962219238, "num_tokens": 512407217.0, "step": 13499 }, { "epoch": 1.7173387609718866, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.484332799911499, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8815340995788574, "num_tokens": 512441975.0, "step": 13500 }, { "epoch": 1.7174659712504772, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.348208427429199, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.861759603023529, "num_tokens": 512482956.0, "step": 13501 }, { "epoch": 1.7175931815290677, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5103366374969482, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8685057163238525, "num_tokens": 512518514.0, "step": 13502 }, { "epoch": 1.717720391807658, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.274956703186035, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8604320287704468, "num_tokens": 512561620.0, "step": 13503 }, { "epoch": 1.7178476020862485, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5640528202056885, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8736789226531982, "num_tokens": 512598234.0, "step": 13504 }, { "epoch": 1.717974812364839, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2053308486938477, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8691152930259705, "num_tokens": 512638219.0, "step": 13505 }, { "epoch": 1.7181020226434296, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6992506980895996, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8588308095932007, "num_tokens": 512672277.0, "step": 13506 }, { "epoch": 1.7182292329220201, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1332592964172363, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8742555379867554, "num_tokens": 512709573.0, "step": 13507 }, { "epoch": 1.7183564432006107, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8867976665496826, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8639532327651978, "num_tokens": 512750243.0, "step": 13508 }, { "epoch": 1.718483653479201, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.131788969039917, "learning_rate": 1e-05, "loss": 0.5238, "mean_token_accuracy": 0.8470546007156372, "num_tokens": 512786001.0, "step": 13509 }, { "epoch": 1.7186108637577915, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.11102294921875, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8738231658935547, "num_tokens": 512823844.0, "step": 13510 }, { "epoch": 1.718738074036382, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2389278411865234, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8713158965110779, "num_tokens": 512859474.0, "step": 13511 }, { "epoch": 1.7188652843149725, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.290790557861328, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8822469711303711, "num_tokens": 512889670.0, "step": 13512 }, { "epoch": 1.718992494593563, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.526679515838623, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8807643055915833, "num_tokens": 512919749.0, "step": 13513 }, { "epoch": 1.7191197048721536, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.221873760223389, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8671098947525024, "num_tokens": 512953206.0, "step": 13514 }, { "epoch": 1.7192469151507441, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.704545497894287, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8659185171127319, "num_tokens": 512989066.0, "step": 13515 }, { "epoch": 1.7193741254293347, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.007445335388184, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8884782791137695, "num_tokens": 513021975.0, "step": 13516 }, { "epoch": 1.7195013357079252, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6852407455444336, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8711669445037842, "num_tokens": 513061669.0, "step": 13517 }, { "epoch": 1.7196285459865157, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8205130100250244, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.868401050567627, "num_tokens": 513095714.0, "step": 13518 }, { "epoch": 1.7197557562651062, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6074130535125732, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8703505992889404, "num_tokens": 513137235.0, "step": 13519 }, { "epoch": 1.7198829665436968, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6712818145751953, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8730401992797852, "num_tokens": 513176339.0, "step": 13520 }, { "epoch": 1.7200101768222873, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.51129150390625, "learning_rate": 1e-05, "loss": 0.3955, "mean_token_accuracy": 0.8892436623573303, "num_tokens": 513218984.0, "step": 13521 }, { "epoch": 1.7201373871008778, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6047422885894775, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8618438839912415, "num_tokens": 513259487.0, "step": 13522 }, { "epoch": 1.7202645973794684, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4399609565734863, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8640472888946533, "num_tokens": 513296270.0, "step": 13523 }, { "epoch": 1.7203918076580589, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5953369140625, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8630131483078003, "num_tokens": 513333282.0, "step": 13524 }, { "epoch": 1.7205190179366494, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3633484840393066, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8725051283836365, "num_tokens": 513365679.0, "step": 13525 }, { "epoch": 1.72064622821524, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.597625970840454, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8708690404891968, "num_tokens": 513403387.0, "step": 13526 }, { "epoch": 1.7207734384938305, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3008737564086914, "learning_rate": 1e-05, "loss": 0.5262, "mean_token_accuracy": 0.8471509218215942, "num_tokens": 513442886.0, "step": 13527 }, { "epoch": 1.7209006487724208, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6535465717315674, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8780921101570129, "num_tokens": 513482401.0, "step": 13528 }, { "epoch": 1.7210278590510113, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2229480743408203, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8707890510559082, "num_tokens": 513517898.0, "step": 13529 }, { "epoch": 1.7211550693296018, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7815966606140137, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8730272650718689, "num_tokens": 513556761.0, "step": 13530 }, { "epoch": 1.7212822796081924, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1961169242858887, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.86155766248703, "num_tokens": 513590803.0, "step": 13531 }, { "epoch": 1.7214094898867829, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9577488899230957, "learning_rate": 1e-05, "loss": 0.4044, "mean_token_accuracy": 0.8924282789230347, "num_tokens": 513626718.0, "step": 13532 }, { "epoch": 1.7215367001653734, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.221552848815918, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8684397339820862, "num_tokens": 513663439.0, "step": 13533 }, { "epoch": 1.7216639104439637, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.097569465637207, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8728405833244324, "num_tokens": 513699916.0, "step": 13534 }, { "epoch": 1.7217911207225542, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.35276198387146, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8726519346237183, "num_tokens": 513739411.0, "step": 13535 }, { "epoch": 1.7219183310011448, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.131746292114258, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8742775917053223, "num_tokens": 513777497.0, "step": 13536 }, { "epoch": 1.7220455412797353, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5361135005950928, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8645186424255371, "num_tokens": 513817658.0, "step": 13537 }, { "epoch": 1.7221727515583258, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.003319263458252, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8837229013442993, "num_tokens": 513853011.0, "step": 13538 }, { "epoch": 1.7222999618369164, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6019790172576904, "learning_rate": 1e-05, "loss": 0.3965, "mean_token_accuracy": 0.8950343132019043, "num_tokens": 513892933.0, "step": 13539 }, { "epoch": 1.7224271721155069, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8053627014160156, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.8575753569602966, "num_tokens": 513932284.0, "step": 13540 }, { "epoch": 1.7225543823940974, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.593919038772583, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8785362243652344, "num_tokens": 513968151.0, "step": 13541 }, { "epoch": 1.722681592672688, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.624425172805786, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8692973852157593, "num_tokens": 514005323.0, "step": 13542 }, { "epoch": 1.7228088029512785, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5686686038970947, "learning_rate": 1e-05, "loss": 0.4982, "mean_token_accuracy": 0.8570020198822021, "num_tokens": 514044158.0, "step": 13543 }, { "epoch": 1.722936013229869, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4769997596740723, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8746261596679688, "num_tokens": 514081076.0, "step": 13544 }, { "epoch": 1.7230632235084595, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5404396057128906, "learning_rate": 1e-05, "loss": 0.4949, "mean_token_accuracy": 0.8574385643005371, "num_tokens": 514122784.0, "step": 13545 }, { "epoch": 1.72319043378705, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.402996301651001, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8697710037231445, "num_tokens": 514158998.0, "step": 13546 }, { "epoch": 1.7233176440656406, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5173492431640625, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8672780394554138, "num_tokens": 514195815.0, "step": 13547 }, { "epoch": 1.7234448543442311, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.348816394805908, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8660362958908081, "num_tokens": 514228353.0, "step": 13548 }, { "epoch": 1.7235720646228216, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5217092037200928, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8653850555419922, "num_tokens": 514267283.0, "step": 13549 }, { "epoch": 1.7236992749014122, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2790796756744385, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8767653107643127, "num_tokens": 514309950.0, "step": 13550 }, { "epoch": 1.7238264851800027, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5704078674316406, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.882993221282959, "num_tokens": 514345043.0, "step": 13551 }, { "epoch": 1.723953695458593, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.22658634185791, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8725236654281616, "num_tokens": 514383692.0, "step": 13552 }, { "epoch": 1.7240809057371835, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6591811180114746, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.8567777276039124, "num_tokens": 514423552.0, "step": 13553 }, { "epoch": 1.724208116015774, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1602394580841064, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8686259388923645, "num_tokens": 514460460.0, "step": 13554 }, { "epoch": 1.7243353262943646, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8097426891326904, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8655288219451904, "num_tokens": 514504751.0, "step": 13555 }, { "epoch": 1.7244625365729551, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.106234312057495, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.86637943983078, "num_tokens": 514539967.0, "step": 13556 }, { "epoch": 1.7245897468515456, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.044956207275391, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8729331493377686, "num_tokens": 514583191.0, "step": 13557 }, { "epoch": 1.724716957130136, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1573774814605713, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8722702860832214, "num_tokens": 514622832.0, "step": 13558 }, { "epoch": 1.7248441674087265, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.292492866516113, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8755296468734741, "num_tokens": 514663387.0, "step": 13559 }, { "epoch": 1.724971377687317, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.426865577697754, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8648451566696167, "num_tokens": 514702711.0, "step": 13560 }, { "epoch": 1.7250985879659075, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.334016799926758, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.8548068404197693, "num_tokens": 514740341.0, "step": 13561 }, { "epoch": 1.725225798244498, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.714503049850464, "learning_rate": 1e-05, "loss": 0.5262, "mean_token_accuracy": 0.8552166223526001, "num_tokens": 514774834.0, "step": 13562 }, { "epoch": 1.7253530085230886, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.131131649017334, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8555209636688232, "num_tokens": 514814988.0, "step": 13563 }, { "epoch": 1.7254802188016791, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7432167530059814, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.875403881072998, "num_tokens": 514854936.0, "step": 13564 }, { "epoch": 1.7256074290802697, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.929870367050171, "learning_rate": 1e-05, "loss": 0.5184, "mean_token_accuracy": 0.8523739576339722, "num_tokens": 514890528.0, "step": 13565 }, { "epoch": 1.7257346393588602, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6604084968566895, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8666871786117554, "num_tokens": 514935455.0, "step": 13566 }, { "epoch": 1.7258618496374507, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.762760877609253, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8640635013580322, "num_tokens": 514975059.0, "step": 13567 }, { "epoch": 1.7259890599160412, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5716233253479004, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8781698346138, "num_tokens": 515007820.0, "step": 13568 }, { "epoch": 1.7261162701946318, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.66140079498291, "learning_rate": 1e-05, "loss": 0.4951, "mean_token_accuracy": 0.8610295057296753, "num_tokens": 515051826.0, "step": 13569 }, { "epoch": 1.7262434804732223, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4762468338012695, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8644444346427917, "num_tokens": 515092316.0, "step": 13570 }, { "epoch": 1.7263706907518128, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6306440830230713, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8844449520111084, "num_tokens": 515124314.0, "step": 13571 }, { "epoch": 1.7264979010304033, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3862645626068115, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8688246011734009, "num_tokens": 515159725.0, "step": 13572 }, { "epoch": 1.7266251113089939, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6338515281677246, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8603376150131226, "num_tokens": 515201095.0, "step": 13573 }, { "epoch": 1.7267523215875844, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2891268730163574, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8742157816886902, "num_tokens": 515241761.0, "step": 13574 }, { "epoch": 1.726879531866175, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7144134044647217, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8687081336975098, "num_tokens": 515281983.0, "step": 13575 }, { "epoch": 1.7270067421447655, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2401325702667236, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8640265464782715, "num_tokens": 515315557.0, "step": 13576 }, { "epoch": 1.7271339524233558, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8591489791870117, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8629323840141296, "num_tokens": 515345969.0, "step": 13577 }, { "epoch": 1.7272611627019463, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.199314594268799, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8626691102981567, "num_tokens": 515385520.0, "step": 13578 }, { "epoch": 1.7273883729805368, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.023741722106934, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8592628240585327, "num_tokens": 515424296.0, "step": 13579 }, { "epoch": 1.7275155832591274, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.282074213027954, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8809570670127869, "num_tokens": 515461602.0, "step": 13580 }, { "epoch": 1.7276427935377179, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.133801460266113, "learning_rate": 1e-05, "loss": 0.5028, "mean_token_accuracy": 0.8554438948631287, "num_tokens": 515504123.0, "step": 13581 }, { "epoch": 1.7277700038163084, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4751124382019043, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8722676038742065, "num_tokens": 515543056.0, "step": 13582 }, { "epoch": 1.7278972140948987, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.066814422607422, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8662962913513184, "num_tokens": 515580194.0, "step": 13583 }, { "epoch": 1.7280244243734892, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6184659004211426, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8751323223114014, "num_tokens": 515612493.0, "step": 13584 }, { "epoch": 1.7281516346520798, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8477978706359863, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8833054304122925, "num_tokens": 515656285.0, "step": 13585 }, { "epoch": 1.7282788449306703, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6357035636901855, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8759958744049072, "num_tokens": 515695326.0, "step": 13586 }, { "epoch": 1.7284060552092608, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6374027729034424, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8541250228881836, "num_tokens": 515730865.0, "step": 13587 }, { "epoch": 1.7285332654878514, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.617805004119873, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8624977469444275, "num_tokens": 515766725.0, "step": 13588 }, { "epoch": 1.7286604757664419, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4568560123443604, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8647341728210449, "num_tokens": 515807660.0, "step": 13589 }, { "epoch": 1.7287876860450324, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6047329902648926, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8794525861740112, "num_tokens": 515844833.0, "step": 13590 }, { "epoch": 1.728914896323623, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3488903045654297, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8709272146224976, "num_tokens": 515876538.0, "step": 13591 }, { "epoch": 1.7290421066022135, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6286683082580566, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8695821166038513, "num_tokens": 515916217.0, "step": 13592 }, { "epoch": 1.729169316880804, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2657079696655273, "learning_rate": 1e-05, "loss": 0.5138, "mean_token_accuracy": 0.8499966859817505, "num_tokens": 515955129.0, "step": 13593 }, { "epoch": 1.7292965271593945, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.69443678855896, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8635691404342651, "num_tokens": 515993910.0, "step": 13594 }, { "epoch": 1.729423737437985, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.187394618988037, "learning_rate": 1e-05, "loss": 0.4102, "mean_token_accuracy": 0.8843362927436829, "num_tokens": 516034376.0, "step": 13595 }, { "epoch": 1.7295509477165756, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8358592987060547, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8649210929870605, "num_tokens": 516070383.0, "step": 13596 }, { "epoch": 1.729678157995166, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1554269790649414, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8689785003662109, "num_tokens": 516114527.0, "step": 13597 }, { "epoch": 1.7298053682737566, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0136027336120605, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8791748285293579, "num_tokens": 516156179.0, "step": 13598 }, { "epoch": 1.7299325785523472, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2260501384735107, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8759361505508423, "num_tokens": 516191000.0, "step": 13599 }, { "epoch": 1.7300597888309377, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.17900276184082, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8738426566123962, "num_tokens": 516224194.0, "step": 13600 }, { "epoch": 1.730186999109528, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4085166454315186, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8741822838783264, "num_tokens": 516262210.0, "step": 13601 }, { "epoch": 1.7303142093881185, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.201347827911377, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8657860159873962, "num_tokens": 516298859.0, "step": 13602 }, { "epoch": 1.730441419666709, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5851004123687744, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8629722595214844, "num_tokens": 516344100.0, "step": 13603 }, { "epoch": 1.7305686299452996, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.060842990875244, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8626503348350525, "num_tokens": 516391313.0, "step": 13604 }, { "epoch": 1.7306958402238901, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6602163314819336, "learning_rate": 1e-05, "loss": 0.505, "mean_token_accuracy": 0.857296884059906, "num_tokens": 516430814.0, "step": 13605 }, { "epoch": 1.7308230505024806, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.860968589782715, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8714204430580139, "num_tokens": 516471380.0, "step": 13606 }, { "epoch": 1.730950260781071, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6135969161987305, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8710551261901855, "num_tokens": 516518374.0, "step": 13607 }, { "epoch": 1.7310774710596615, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7101619243621826, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.865949809551239, "num_tokens": 516549917.0, "step": 13608 }, { "epoch": 1.731204681338252, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5470499992370605, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8569620847702026, "num_tokens": 516592869.0, "step": 13609 }, { "epoch": 1.7313318916168425, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5877437591552734, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8723364472389221, "num_tokens": 516630550.0, "step": 13610 }, { "epoch": 1.731459101895433, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4851207733154297, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8769865036010742, "num_tokens": 516669285.0, "step": 13611 }, { "epoch": 1.7315863121740236, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5559728145599365, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8838793635368347, "num_tokens": 516704616.0, "step": 13612 }, { "epoch": 1.7317135224526141, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.401099681854248, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8777704834938049, "num_tokens": 516749566.0, "step": 13613 }, { "epoch": 1.7318407327312046, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.534858226776123, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8738710880279541, "num_tokens": 516786264.0, "step": 13614 }, { "epoch": 1.7319679430097952, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.34521746635437, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8640314936637878, "num_tokens": 516822471.0, "step": 13615 }, { "epoch": 1.7320951532883857, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.561244487762451, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8690617084503174, "num_tokens": 516861840.0, "step": 13616 }, { "epoch": 1.7322223635669762, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3001039028167725, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8680801391601562, "num_tokens": 516895451.0, "step": 13617 }, { "epoch": 1.7323495738455668, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6052393913269043, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8779399991035461, "num_tokens": 516935926.0, "step": 13618 }, { "epoch": 1.7324767841241573, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2284367084503174, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8707051277160645, "num_tokens": 516976268.0, "step": 13619 }, { "epoch": 1.7326039944027478, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7158517837524414, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8686082363128662, "num_tokens": 517021459.0, "step": 13620 }, { "epoch": 1.7327312046813383, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1754767894744873, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8753097057342529, "num_tokens": 517052441.0, "step": 13621 }, { "epoch": 1.7328584149599289, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8771560192108154, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.85940021276474, "num_tokens": 517094575.0, "step": 13622 }, { "epoch": 1.7329856252385194, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1638357639312744, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8683563470840454, "num_tokens": 517135239.0, "step": 13623 }, { "epoch": 1.73311283551711, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.062337875366211, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8682771921157837, "num_tokens": 517174132.0, "step": 13624 }, { "epoch": 1.7332400457957005, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2718279361724854, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8730348348617554, "num_tokens": 517208693.0, "step": 13625 }, { "epoch": 1.7333672560742908, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.189600467681885, "learning_rate": 1e-05, "loss": 0.5332, "mean_token_accuracy": 0.8522041440010071, "num_tokens": 517249441.0, "step": 13626 }, { "epoch": 1.7334944663528813, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.477855682373047, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8649410009384155, "num_tokens": 517291368.0, "step": 13627 }, { "epoch": 1.7336216766314718, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.126827716827393, "learning_rate": 1e-05, "loss": 0.5025, "mean_token_accuracy": 0.8613925576210022, "num_tokens": 517328292.0, "step": 13628 }, { "epoch": 1.7337488869100623, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.632436752319336, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8771772384643555, "num_tokens": 517361049.0, "step": 13629 }, { "epoch": 1.7338760971886529, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9325883388519287, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8693442344665527, "num_tokens": 517395086.0, "step": 13630 }, { "epoch": 1.7340033074672434, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.63350248336792, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8688771724700928, "num_tokens": 517432919.0, "step": 13631 }, { "epoch": 1.7341305177458337, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7328481674194336, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8704708218574524, "num_tokens": 517474676.0, "step": 13632 }, { "epoch": 1.7342577280244242, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5872106552124023, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8780570030212402, "num_tokens": 517514131.0, "step": 13633 }, { "epoch": 1.7343849383030148, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5793802738189697, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8729750514030457, "num_tokens": 517547704.0, "step": 13634 }, { "epoch": 1.7345121485816053, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5345115661621094, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8687924146652222, "num_tokens": 517586430.0, "step": 13635 }, { "epoch": 1.7346393588601958, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.47951602935791, "learning_rate": 1e-05, "loss": 0.5036, "mean_token_accuracy": 0.8563747406005859, "num_tokens": 517630303.0, "step": 13636 }, { "epoch": 1.7347665691387864, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4884214401245117, "learning_rate": 1e-05, "loss": 0.3876, "mean_token_accuracy": 0.8890944719314575, "num_tokens": 517663276.0, "step": 13637 }, { "epoch": 1.7348937794173769, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.438352584838867, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8664834499359131, "num_tokens": 517697946.0, "step": 13638 }, { "epoch": 1.7350209896959674, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4552571773529053, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.867888867855072, "num_tokens": 517730820.0, "step": 13639 }, { "epoch": 1.735148199974558, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3873963356018066, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8746247887611389, "num_tokens": 517774576.0, "step": 13640 }, { "epoch": 1.7352754102531485, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.43451189994812, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8609446287155151, "num_tokens": 517814748.0, "step": 13641 }, { "epoch": 1.735402620531739, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3722996711730957, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8700513243675232, "num_tokens": 517846582.0, "step": 13642 }, { "epoch": 1.7355298308103295, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4271838665008545, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8556934595108032, "num_tokens": 517883354.0, "step": 13643 }, { "epoch": 1.73565704108892, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3531105518341064, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8541014194488525, "num_tokens": 517920920.0, "step": 13644 }, { "epoch": 1.7357842513675106, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.420872688293457, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8584942817687988, "num_tokens": 517964844.0, "step": 13645 }, { "epoch": 1.735911461646101, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3357815742492676, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8684365749359131, "num_tokens": 517998803.0, "step": 13646 }, { "epoch": 1.7360386719246916, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4247188568115234, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8668583631515503, "num_tokens": 518040012.0, "step": 13647 }, { "epoch": 1.7361658822032822, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.303999185562134, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8608765006065369, "num_tokens": 518082287.0, "step": 13648 }, { "epoch": 1.7362930924818727, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4708118438720703, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8716863989830017, "num_tokens": 518116898.0, "step": 13649 }, { "epoch": 1.736420302760463, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2538065910339355, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8751020431518555, "num_tokens": 518157559.0, "step": 13650 }, { "epoch": 1.7365475130390535, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5264101028442383, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8621455430984497, "num_tokens": 518200691.0, "step": 13651 }, { "epoch": 1.736674723317644, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1940999031066895, "learning_rate": 1e-05, "loss": 0.5017, "mean_token_accuracy": 0.854826807975769, "num_tokens": 518239188.0, "step": 13652 }, { "epoch": 1.7368019335962346, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.656428813934326, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8621823787689209, "num_tokens": 518274294.0, "step": 13653 }, { "epoch": 1.736929143874825, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1013197898864746, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8821624517440796, "num_tokens": 518310960.0, "step": 13654 }, { "epoch": 1.7370563541534156, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.874241352081299, "learning_rate": 1e-05, "loss": 0.5402, "mean_token_accuracy": 0.8484417200088501, "num_tokens": 518352133.0, "step": 13655 }, { "epoch": 1.737183564432006, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.0440526008605957, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8600507378578186, "num_tokens": 518388793.0, "step": 13656 }, { "epoch": 1.7373107747105965, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.204843521118164, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8785519003868103, "num_tokens": 518425581.0, "step": 13657 }, { "epoch": 1.737437984989187, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.1597769260406494, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8729386925697327, "num_tokens": 518465477.0, "step": 13658 }, { "epoch": 1.7375651952677775, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.536416053771973, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8619482517242432, "num_tokens": 518506790.0, "step": 13659 }, { "epoch": 1.737692405546368, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.6258394718170166, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8694283962249756, "num_tokens": 518548240.0, "step": 13660 }, { "epoch": 1.7378196158249586, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.489925861358643, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8580749034881592, "num_tokens": 518583865.0, "step": 13661 }, { "epoch": 1.7379468261035491, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.9426138401031494, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8798028230667114, "num_tokens": 518623303.0, "step": 13662 }, { "epoch": 1.7380740363821396, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.18324089050293, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.8665122985839844, "num_tokens": 518662375.0, "step": 13663 }, { "epoch": 1.7382012466607302, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.87972354888916, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.8629792928695679, "num_tokens": 518702007.0, "step": 13664 }, { "epoch": 1.7383284569393207, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.963881492614746, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8814159035682678, "num_tokens": 518743676.0, "step": 13665 }, { "epoch": 1.7384556672179112, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.7154757976531982, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8758279085159302, "num_tokens": 518783550.0, "step": 13666 }, { "epoch": 1.7385828774965018, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8320140838623047, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8863422274589539, "num_tokens": 518820788.0, "step": 13667 }, { "epoch": 1.7387100877750923, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5726914405822754, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.860190749168396, "num_tokens": 518865195.0, "step": 13668 }, { "epoch": 1.7388372980536828, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.774463176727295, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8667568564414978, "num_tokens": 518903683.0, "step": 13669 }, { "epoch": 1.7389645083322733, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4395487308502197, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8803638219833374, "num_tokens": 518939535.0, "step": 13670 }, { "epoch": 1.7390917186108639, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.795308828353882, "learning_rate": 1e-05, "loss": 0.513, "mean_token_accuracy": 0.8559208512306213, "num_tokens": 518978753.0, "step": 13671 }, { "epoch": 1.7392189288894544, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3510396480560303, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.857846736907959, "num_tokens": 519014916.0, "step": 13672 }, { "epoch": 1.739346139168045, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.879767417907715, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8657776117324829, "num_tokens": 519043537.0, "step": 13673 }, { "epoch": 1.7394733494466355, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2823662757873535, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8722002506256104, "num_tokens": 519081388.0, "step": 13674 }, { "epoch": 1.7396005597252258, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 4.013709545135498, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.87941974401474, "num_tokens": 519119531.0, "step": 13675 }, { "epoch": 1.7397277700038163, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.327611207962036, "learning_rate": 1e-05, "loss": 0.5096, "mean_token_accuracy": 0.8541275858879089, "num_tokens": 519157982.0, "step": 13676 }, { "epoch": 1.7398549802824068, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.139726161956787, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8641879558563232, "num_tokens": 519198928.0, "step": 13677 }, { "epoch": 1.7399821905609973, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4905354976654053, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8638083934783936, "num_tokens": 519237760.0, "step": 13678 }, { "epoch": 1.7401094008395879, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.098705291748047, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8705092072486877, "num_tokens": 519273557.0, "step": 13679 }, { "epoch": 1.7402366111181784, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6281163692474365, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8652682304382324, "num_tokens": 519316082.0, "step": 13680 }, { "epoch": 1.7403638213967687, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9129035472869873, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8842180967330933, "num_tokens": 519352528.0, "step": 13681 }, { "epoch": 1.7404910316753592, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6614184379577637, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8892434239387512, "num_tokens": 519384931.0, "step": 13682 }, { "epoch": 1.7406182419539498, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6890299320220947, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8698015213012695, "num_tokens": 519421974.0, "step": 13683 }, { "epoch": 1.7407454522325403, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6275382041931152, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8757807612419128, "num_tokens": 519460623.0, "step": 13684 }, { "epoch": 1.7408726625111308, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.5005369186401367, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.881626546382904, "num_tokens": 519497079.0, "step": 13685 }, { "epoch": 1.7409998727897213, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6099023818969727, "learning_rate": 1e-05, "loss": 0.5032, "mean_token_accuracy": 0.854475200176239, "num_tokens": 519542680.0, "step": 13686 }, { "epoch": 1.7411270830683119, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.36958646774292, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8670136332511902, "num_tokens": 519577025.0, "step": 13687 }, { "epoch": 1.7412542933469024, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6287949085235596, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8662266731262207, "num_tokens": 519618616.0, "step": 13688 }, { "epoch": 1.741381503625493, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2657909393310547, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8659444451332092, "num_tokens": 519658672.0, "step": 13689 }, { "epoch": 1.7415087139040835, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.688061237335205, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8915113210678101, "num_tokens": 519692830.0, "step": 13690 }, { "epoch": 1.741635924182674, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.190469264984131, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8738702535629272, "num_tokens": 519729949.0, "step": 13691 }, { "epoch": 1.7417631344612645, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.842090129852295, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8677377700805664, "num_tokens": 519763556.0, "step": 13692 }, { "epoch": 1.741890344739855, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1382522583007812, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8617080450057983, "num_tokens": 519804556.0, "step": 13693 }, { "epoch": 1.7420175550184456, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.043504238128662, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8753736615180969, "num_tokens": 519845846.0, "step": 13694 }, { "epoch": 1.742144765297036, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2412548065185547, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.862575113773346, "num_tokens": 519883333.0, "step": 13695 }, { "epoch": 1.7422719755756266, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.204225063323975, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8807622194290161, "num_tokens": 519915927.0, "step": 13696 }, { "epoch": 1.7423991858542172, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4293887615203857, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.870476484298706, "num_tokens": 519951555.0, "step": 13697 }, { "epoch": 1.7425263961328077, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.195423603057861, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8802827596664429, "num_tokens": 519996616.0, "step": 13698 }, { "epoch": 1.742653606411398, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.652076244354248, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8540235161781311, "num_tokens": 520032467.0, "step": 13699 }, { "epoch": 1.7427808166899885, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.017730236053467, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8725962042808533, "num_tokens": 520068893.0, "step": 13700 }, { "epoch": 1.742908026968579, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6640782356262207, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.870272159576416, "num_tokens": 520104745.0, "step": 13701 }, { "epoch": 1.7430352372471696, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.821277618408203, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8559979796409607, "num_tokens": 520146994.0, "step": 13702 }, { "epoch": 1.74316244752576, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.597777843475342, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8783206939697266, "num_tokens": 520188220.0, "step": 13703 }, { "epoch": 1.7432896578043506, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6752803325653076, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.8865804672241211, "num_tokens": 520223410.0, "step": 13704 }, { "epoch": 1.743416868082941, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.535896062850952, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8683289885520935, "num_tokens": 520260017.0, "step": 13705 }, { "epoch": 1.7435440783615315, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.593153715133667, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8767790794372559, "num_tokens": 520296232.0, "step": 13706 }, { "epoch": 1.743671288640122, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4664406776428223, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8592156171798706, "num_tokens": 520330720.0, "step": 13707 }, { "epoch": 1.7437984989187125, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5629708766937256, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8767703175544739, "num_tokens": 520362438.0, "step": 13708 }, { "epoch": 1.743925709197303, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3759829998016357, "learning_rate": 1e-05, "loss": 0.4029, "mean_token_accuracy": 0.8857576847076416, "num_tokens": 520398170.0, "step": 13709 }, { "epoch": 1.7440529194758936, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.564305543899536, "learning_rate": 1e-05, "loss": 0.52, "mean_token_accuracy": 0.854141891002655, "num_tokens": 520438219.0, "step": 13710 }, { "epoch": 1.744180129754484, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3081839084625244, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8746490478515625, "num_tokens": 520481642.0, "step": 13711 }, { "epoch": 1.7443073400330746, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6185524463653564, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8721878528594971, "num_tokens": 520515945.0, "step": 13712 }, { "epoch": 1.7444345503116652, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.237332344055176, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8631691336631775, "num_tokens": 520553259.0, "step": 13713 }, { "epoch": 1.7445617605902557, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7374157905578613, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8815809488296509, "num_tokens": 520590371.0, "step": 13714 }, { "epoch": 1.7446889708688462, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.176745653152466, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8707666397094727, "num_tokens": 520624738.0, "step": 13715 }, { "epoch": 1.7448161811474368, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.895829916000366, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8832485675811768, "num_tokens": 520666631.0, "step": 13716 }, { "epoch": 1.7449433914260273, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1930673122406006, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.87790846824646, "num_tokens": 520696567.0, "step": 13717 }, { "epoch": 1.7450706017046178, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.058332443237305, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8660489916801453, "num_tokens": 520737975.0, "step": 13718 }, { "epoch": 1.7451978119832083, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3067378997802734, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8829303979873657, "num_tokens": 520776884.0, "step": 13719 }, { "epoch": 1.7453250222617989, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.14000940322876, "learning_rate": 1e-05, "loss": 0.5142, "mean_token_accuracy": 0.8588900566101074, "num_tokens": 520817770.0, "step": 13720 }, { "epoch": 1.7454522325403894, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4823310375213623, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8718912601470947, "num_tokens": 520859381.0, "step": 13721 }, { "epoch": 1.74557944281898, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.044090747833252, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8915138244628906, "num_tokens": 520900007.0, "step": 13722 }, { "epoch": 1.7457066530975704, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5922465324401855, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8771156072616577, "num_tokens": 520940707.0, "step": 13723 }, { "epoch": 1.7458338633761608, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8815364837646484, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8638260960578918, "num_tokens": 520972565.0, "step": 13724 }, { "epoch": 1.7459610736547513, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6049840450286865, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8652327656745911, "num_tokens": 521012364.0, "step": 13725 }, { "epoch": 1.7460882839333418, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.668060302734375, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8825943470001221, "num_tokens": 521058761.0, "step": 13726 }, { "epoch": 1.7462154942119323, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.58595871925354, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8704249858856201, "num_tokens": 521096620.0, "step": 13727 }, { "epoch": 1.7463427044905229, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5290439128875732, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8668011426925659, "num_tokens": 521133223.0, "step": 13728 }, { "epoch": 1.7464699147691134, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5443379878997803, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8835456371307373, "num_tokens": 521170334.0, "step": 13729 }, { "epoch": 1.7465971250477037, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4094510078430176, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8777374625205994, "num_tokens": 521207458.0, "step": 13730 }, { "epoch": 1.7467243353262942, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5382027626037598, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8715223073959351, "num_tokens": 521245017.0, "step": 13731 }, { "epoch": 1.7468515456048848, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3384714126586914, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8659644722938538, "num_tokens": 521285792.0, "step": 13732 }, { "epoch": 1.7469787558834753, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.550143003463745, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.881009578704834, "num_tokens": 521327107.0, "step": 13733 }, { "epoch": 1.7471059661620658, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2822346687316895, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8817296028137207, "num_tokens": 521364720.0, "step": 13734 }, { "epoch": 1.7472331764406563, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.600909471511841, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8795145750045776, "num_tokens": 521398183.0, "step": 13735 }, { "epoch": 1.7473603867192469, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.217381238937378, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8692890405654907, "num_tokens": 521436837.0, "step": 13736 }, { "epoch": 1.7474875969978374, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.689091205596924, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8708012700080872, "num_tokens": 521477722.0, "step": 13737 }, { "epoch": 1.747614807276428, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.159949779510498, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8764731884002686, "num_tokens": 521513647.0, "step": 13738 }, { "epoch": 1.7477420175550185, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8614890575408936, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8725893497467041, "num_tokens": 521545480.0, "step": 13739 }, { "epoch": 1.747869227833609, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1178348064422607, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8600058555603027, "num_tokens": 521587363.0, "step": 13740 }, { "epoch": 1.7479964381121995, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.089721202850342, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8648273944854736, "num_tokens": 521629399.0, "step": 13741 }, { "epoch": 1.74812364839079, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.243197202682495, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8686369061470032, "num_tokens": 521667065.0, "step": 13742 }, { "epoch": 1.7482508586693806, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.261805057525635, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8768049478530884, "num_tokens": 521696317.0, "step": 13743 }, { "epoch": 1.748378068947971, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.452687978744507, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8771151304244995, "num_tokens": 521736948.0, "step": 13744 }, { "epoch": 1.7485052792265616, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2531962394714355, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8837032318115234, "num_tokens": 521770320.0, "step": 13745 }, { "epoch": 1.7486324895051522, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.682299852371216, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8843069076538086, "num_tokens": 521803836.0, "step": 13746 }, { "epoch": 1.7487596997837427, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.081072807312012, "learning_rate": 1e-05, "loss": 0.5049, "mean_token_accuracy": 0.8588759303092957, "num_tokens": 521838066.0, "step": 13747 }, { "epoch": 1.748886910062333, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.689495086669922, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8753352165222168, "num_tokens": 521876806.0, "step": 13748 }, { "epoch": 1.7490141203409235, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8779714107513428, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8642262816429138, "num_tokens": 521916241.0, "step": 13749 }, { "epoch": 1.749141330619514, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6377391815185547, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.876876175403595, "num_tokens": 521949369.0, "step": 13750 }, { "epoch": 1.7492685408981046, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7109692096710205, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.860237181186676, "num_tokens": 521989251.0, "step": 13751 }, { "epoch": 1.749395751176695, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.560981273651123, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8767229318618774, "num_tokens": 522019221.0, "step": 13752 }, { "epoch": 1.7495229614552856, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6059203147888184, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.867387056350708, "num_tokens": 522058858.0, "step": 13753 }, { "epoch": 1.749650171733876, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.492893934249878, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8719310760498047, "num_tokens": 522089834.0, "step": 13754 }, { "epoch": 1.7497773820124665, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.542645215988159, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8795566558837891, "num_tokens": 522124474.0, "step": 13755 }, { "epoch": 1.749904592291057, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.423807382583618, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8636960387229919, "num_tokens": 522159046.0, "step": 13756 }, { "epoch": 1.7500318025696475, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.527557134628296, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8696801662445068, "num_tokens": 522200698.0, "step": 13757 }, { "epoch": 1.750159012848238, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.360327959060669, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8573064208030701, "num_tokens": 522233533.0, "step": 13758 }, { "epoch": 1.7502862231268286, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5452089309692383, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.861919105052948, "num_tokens": 522274234.0, "step": 13759 }, { "epoch": 1.750413433405419, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2869112491607666, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8731234073638916, "num_tokens": 522313055.0, "step": 13760 }, { "epoch": 1.7505406436840096, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.611849069595337, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8718616366386414, "num_tokens": 522346372.0, "step": 13761 }, { "epoch": 1.7506678539626002, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.219587802886963, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.872307538986206, "num_tokens": 522381199.0, "step": 13762 }, { "epoch": 1.7507950642411907, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.73555326461792, "learning_rate": 1e-05, "loss": 0.5163, "mean_token_accuracy": 0.8537286520004272, "num_tokens": 522419509.0, "step": 13763 }, { "epoch": 1.7509222745197812, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.141432046890259, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8705773949623108, "num_tokens": 522458433.0, "step": 13764 }, { "epoch": 1.7510494847983717, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.949401378631592, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8718621730804443, "num_tokens": 522492010.0, "step": 13765 }, { "epoch": 1.7511766950769623, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.163356304168701, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8736704587936401, "num_tokens": 522524190.0, "step": 13766 }, { "epoch": 1.7513039053555528, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.16476583480835, "learning_rate": 1e-05, "loss": 0.4946, "mean_token_accuracy": 0.8582800626754761, "num_tokens": 522558218.0, "step": 13767 }, { "epoch": 1.7514311156341433, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3256192207336426, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.874836266040802, "num_tokens": 522594683.0, "step": 13768 }, { "epoch": 1.7515583259127339, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.267277240753174, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.876366138458252, "num_tokens": 522633409.0, "step": 13769 }, { "epoch": 1.7516855361913244, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.579300880432129, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8679196238517761, "num_tokens": 522672469.0, "step": 13770 }, { "epoch": 1.751812746469915, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1405205726623535, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8747060894966125, "num_tokens": 522714676.0, "step": 13771 }, { "epoch": 1.7519399567485054, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.695488929748535, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8838005661964417, "num_tokens": 522748726.0, "step": 13772 }, { "epoch": 1.7520671670270958, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.949744939804077, "learning_rate": 1e-05, "loss": 0.5342, "mean_token_accuracy": 0.8517305850982666, "num_tokens": 522783761.0, "step": 13773 }, { "epoch": 1.7521943773056863, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.671684980392456, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8787695169448853, "num_tokens": 522822027.0, "step": 13774 }, { "epoch": 1.7523215875842768, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7459146976470947, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8764426112174988, "num_tokens": 522856952.0, "step": 13775 }, { "epoch": 1.7524487978628673, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5973572731018066, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8653519153594971, "num_tokens": 522897450.0, "step": 13776 }, { "epoch": 1.7525760081414579, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.621100664138794, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8750067949295044, "num_tokens": 522930547.0, "step": 13777 }, { "epoch": 1.7527032184200484, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.524171829223633, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.869078516960144, "num_tokens": 522970145.0, "step": 13778 }, { "epoch": 1.7528304286986387, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.516151189804077, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8862689137458801, "num_tokens": 523006492.0, "step": 13779 }, { "epoch": 1.7529576389772292, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.467449188232422, "learning_rate": 1e-05, "loss": 0.4848, "mean_token_accuracy": 0.8624933362007141, "num_tokens": 523045524.0, "step": 13780 }, { "epoch": 1.7530848492558198, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4731874465942383, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8752002716064453, "num_tokens": 523086091.0, "step": 13781 }, { "epoch": 1.7532120595344103, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.399935007095337, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8768046498298645, "num_tokens": 523127351.0, "step": 13782 }, { "epoch": 1.7533392698130008, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.476048707962036, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8691517114639282, "num_tokens": 523161775.0, "step": 13783 }, { "epoch": 1.7534664800915913, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3576061725616455, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8679461479187012, "num_tokens": 523198977.0, "step": 13784 }, { "epoch": 1.7535936903701819, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4860239028930664, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8647526502609253, "num_tokens": 523233917.0, "step": 13785 }, { "epoch": 1.7537209006487724, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.312288522720337, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8749368190765381, "num_tokens": 523267679.0, "step": 13786 }, { "epoch": 1.753848110927363, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.532931327819824, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8599357604980469, "num_tokens": 523301512.0, "step": 13787 }, { "epoch": 1.7539753212059535, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2278311252593994, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8788602352142334, "num_tokens": 523342618.0, "step": 13788 }, { "epoch": 1.754102531484544, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6347429752349854, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.872397243976593, "num_tokens": 523379555.0, "step": 13789 }, { "epoch": 1.7542297417631345, "ewc_loss": 0.07080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.410743713378906e-05, "grad_norm": 3.151883363723755, "learning_rate": 1e-05, "loss": 0.3852, "mean_token_accuracy": 0.8904929161071777, "num_tokens": 523414621.0, "step": 13790 }, { "epoch": 1.754356952041725, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.793630599975586, "learning_rate": 1e-05, "loss": 0.3899, "mean_token_accuracy": 0.8925422430038452, "num_tokens": 523459022.0, "step": 13791 }, { "epoch": 1.7544841623203156, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.08422589302063, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.8847746849060059, "num_tokens": 523498718.0, "step": 13792 }, { "epoch": 1.754611372598906, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0781168937683105, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8822823762893677, "num_tokens": 523531926.0, "step": 13793 }, { "epoch": 1.7547385828774966, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.150540351867676, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8727962374687195, "num_tokens": 523572457.0, "step": 13794 }, { "epoch": 1.7548657931560872, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.356045246124268, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8625911474227905, "num_tokens": 523607317.0, "step": 13795 }, { "epoch": 1.7549930034346777, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.481041431427002, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8650909662246704, "num_tokens": 523638979.0, "step": 13796 }, { "epoch": 1.755120213713268, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.355480670928955, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8737972974777222, "num_tokens": 523681839.0, "step": 13797 }, { "epoch": 1.7552474239918585, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.7463488578796387, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8774410486221313, "num_tokens": 523719797.0, "step": 13798 }, { "epoch": 1.755374634270449, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.147114276885986, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8776439428329468, "num_tokens": 523761774.0, "step": 13799 }, { "epoch": 1.7555018445490396, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7807559967041016, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8734179735183716, "num_tokens": 523799522.0, "step": 13800 }, { "epoch": 1.75562905482763, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.945502281188965, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.873865008354187, "num_tokens": 523831927.0, "step": 13801 }, { "epoch": 1.7557562651062206, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.7044405937194824, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8633401989936829, "num_tokens": 523863058.0, "step": 13802 }, { "epoch": 1.755883475384811, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.776158332824707, "learning_rate": 1e-05, "loss": 0.5149, "mean_token_accuracy": 0.850814700126648, "num_tokens": 523907017.0, "step": 13803 }, { "epoch": 1.7560106856634015, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5750813484191895, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8685003519058228, "num_tokens": 523943376.0, "step": 13804 }, { "epoch": 1.756137895941992, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6885573863983154, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8797311782836914, "num_tokens": 523982902.0, "step": 13805 }, { "epoch": 1.7562651062205825, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.481975793838501, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8648951053619385, "num_tokens": 524016423.0, "step": 13806 }, { "epoch": 1.756392316499173, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6491942405700684, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8562305569648743, "num_tokens": 524056686.0, "step": 13807 }, { "epoch": 1.7565195267777636, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.37269926071167, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8714902400970459, "num_tokens": 524093731.0, "step": 13808 }, { "epoch": 1.756646737056354, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.672395944595337, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8697577118873596, "num_tokens": 524139398.0, "step": 13809 }, { "epoch": 1.7567739473349446, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.40069580078125, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8637117147445679, "num_tokens": 524177012.0, "step": 13810 }, { "epoch": 1.7569011576135352, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6638991832733154, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8855602741241455, "num_tokens": 524211785.0, "step": 13811 }, { "epoch": 1.7570283678921257, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.241030216217041, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8841023445129395, "num_tokens": 524252263.0, "step": 13812 }, { "epoch": 1.7571555781707162, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.770491361618042, "learning_rate": 1e-05, "loss": 0.3974, "mean_token_accuracy": 0.8892432451248169, "num_tokens": 524286561.0, "step": 13813 }, { "epoch": 1.7572827884493067, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2014000415802, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8759840726852417, "num_tokens": 524320910.0, "step": 13814 }, { "epoch": 1.7574099987278973, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9359264373779297, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8800536394119263, "num_tokens": 524362416.0, "step": 13815 }, { "epoch": 1.7575372090064878, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2070517539978027, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8792880773544312, "num_tokens": 524405432.0, "step": 13816 }, { "epoch": 1.7576644192850783, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.126613616943359, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8752166032791138, "num_tokens": 524440605.0, "step": 13817 }, { "epoch": 1.7577916295636689, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3790857791900635, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8702208995819092, "num_tokens": 524483126.0, "step": 13818 }, { "epoch": 1.7579188398422594, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.168336391448975, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8788617253303528, "num_tokens": 524516675.0, "step": 13819 }, { "epoch": 1.75804605012085, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5834593772888184, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8745905756950378, "num_tokens": 524554394.0, "step": 13820 }, { "epoch": 1.7581732603994404, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.003286361694336, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.867944598197937, "num_tokens": 524591195.0, "step": 13821 }, { "epoch": 1.7583004706780307, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.661552667617798, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8634132146835327, "num_tokens": 524629077.0, "step": 13822 }, { "epoch": 1.7584276809566213, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.76839542388916, "learning_rate": 1e-05, "loss": 0.5136, "mean_token_accuracy": 0.8554548025131226, "num_tokens": 524672072.0, "step": 13823 }, { "epoch": 1.7585548912352118, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.657742500305176, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8740441799163818, "num_tokens": 524706735.0, "step": 13824 }, { "epoch": 1.7586821015138023, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5459978580474854, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8767725229263306, "num_tokens": 524741750.0, "step": 13825 }, { "epoch": 1.7588093117923929, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.636904716491699, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.888715386390686, "num_tokens": 524774565.0, "step": 13826 }, { "epoch": 1.7589365220709834, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.39823317527771, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8814570903778076, "num_tokens": 524803731.0, "step": 13827 }, { "epoch": 1.7590637323495737, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6465628147125244, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8683888912200928, "num_tokens": 524842570.0, "step": 13828 }, { "epoch": 1.7591909426281642, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2873878479003906, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8715789318084717, "num_tokens": 524880695.0, "step": 13829 }, { "epoch": 1.7593181529067548, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6979868412017822, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.874384343624115, "num_tokens": 524921331.0, "step": 13830 }, { "epoch": 1.7594453631853453, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.226024866104126, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.864279568195343, "num_tokens": 524955769.0, "step": 13831 }, { "epoch": 1.7595725734639358, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.80446457862854, "learning_rate": 1e-05, "loss": 0.5011, "mean_token_accuracy": 0.857721209526062, "num_tokens": 524998299.0, "step": 13832 }, { "epoch": 1.7596997837425263, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1803627014160156, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8600941896438599, "num_tokens": 525038590.0, "step": 13833 }, { "epoch": 1.7598269940211169, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9689524173736572, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8709560036659241, "num_tokens": 525073688.0, "step": 13834 }, { "epoch": 1.7599542042997074, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2038002014160156, "learning_rate": 1e-05, "loss": 0.3876, "mean_token_accuracy": 0.8908953666687012, "num_tokens": 525112109.0, "step": 13835 }, { "epoch": 1.760081414578298, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.135249137878418, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8769378662109375, "num_tokens": 525147554.0, "step": 13836 }, { "epoch": 1.7602086248568884, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.36161208152771, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8607270121574402, "num_tokens": 525187233.0, "step": 13837 }, { "epoch": 1.760335835135479, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.20273494720459, "learning_rate": 1e-05, "loss": 0.5034, "mean_token_accuracy": 0.8619539141654968, "num_tokens": 525218104.0, "step": 13838 }, { "epoch": 1.7604630454140695, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.557663917541504, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8774763941764832, "num_tokens": 525255690.0, "step": 13839 }, { "epoch": 1.76059025569266, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.081575393676758, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8713409900665283, "num_tokens": 525287972.0, "step": 13840 }, { "epoch": 1.7607174659712506, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6657509803771973, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8644430637359619, "num_tokens": 525328003.0, "step": 13841 }, { "epoch": 1.760844676249841, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.867011547088623, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8781291246414185, "num_tokens": 525370642.0, "step": 13842 }, { "epoch": 1.7609718865284316, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6186671257019043, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8584520816802979, "num_tokens": 525408668.0, "step": 13843 }, { "epoch": 1.7610990968070221, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.696035385131836, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.871776819229126, "num_tokens": 525447566.0, "step": 13844 }, { "epoch": 1.7612263070856127, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5491745471954346, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8648894429206848, "num_tokens": 525485456.0, "step": 13845 }, { "epoch": 1.761353517364203, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5936105251312256, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8571262359619141, "num_tokens": 525521190.0, "step": 13846 }, { "epoch": 1.7614807276427935, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.480724811553955, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8731389045715332, "num_tokens": 525559331.0, "step": 13847 }, { "epoch": 1.761607937921384, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.529242753982544, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8641716241836548, "num_tokens": 525598380.0, "step": 13848 }, { "epoch": 1.7617351481999746, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4109904766082764, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8628866076469421, "num_tokens": 525637901.0, "step": 13849 }, { "epoch": 1.761862358478565, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.525991678237915, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8667163848876953, "num_tokens": 525676072.0, "step": 13850 }, { "epoch": 1.7619895687571556, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3457863330841064, "learning_rate": 1e-05, "loss": 0.3921, "mean_token_accuracy": 0.8878886699676514, "num_tokens": 525710655.0, "step": 13851 }, { "epoch": 1.762116779035746, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.548971176147461, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8762774467468262, "num_tokens": 525745714.0, "step": 13852 }, { "epoch": 1.7622439893143365, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2757956981658936, "learning_rate": 1e-05, "loss": 0.3937, "mean_token_accuracy": 0.8865385055541992, "num_tokens": 525782246.0, "step": 13853 }, { "epoch": 1.762371199592927, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6191012859344482, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.886991024017334, "num_tokens": 525814232.0, "step": 13854 }, { "epoch": 1.7624984098715175, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2239396572113037, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8717102408409119, "num_tokens": 525847775.0, "step": 13855 }, { "epoch": 1.762625620150108, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7212016582489014, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8746649622917175, "num_tokens": 525890760.0, "step": 13856 }, { "epoch": 1.7627528304286986, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1660780906677246, "learning_rate": 1e-05, "loss": 0.4097, "mean_token_accuracy": 0.8822389245033264, "num_tokens": 525926893.0, "step": 13857 }, { "epoch": 1.762880040707289, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8872528076171875, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8625980615615845, "num_tokens": 525968907.0, "step": 13858 }, { "epoch": 1.7630072509858796, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.168400764465332, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8761460781097412, "num_tokens": 526003988.0, "step": 13859 }, { "epoch": 1.7631344612644702, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0759596824646, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8759106397628784, "num_tokens": 526044575.0, "step": 13860 }, { "epoch": 1.7632616715430607, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2732653617858887, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8728911280632019, "num_tokens": 526087803.0, "step": 13861 }, { "epoch": 1.7633888818216512, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.207126617431641, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8572942018508911, "num_tokens": 526123432.0, "step": 13862 }, { "epoch": 1.7635160921002417, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4952337741851807, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8823652267456055, "num_tokens": 526161285.0, "step": 13863 }, { "epoch": 1.7636433023788323, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.126330375671387, "learning_rate": 1e-05, "loss": 0.525, "mean_token_accuracy": 0.8552433848381042, "num_tokens": 526203096.0, "step": 13864 }, { "epoch": 1.7637705126574228, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.635439395904541, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8831138014793396, "num_tokens": 526237741.0, "step": 13865 }, { "epoch": 1.7638977229360133, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9230432510375977, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.874116063117981, "num_tokens": 526279704.0, "step": 13866 }, { "epoch": 1.7640249332146039, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.646214485168457, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.870240330696106, "num_tokens": 526315708.0, "step": 13867 }, { "epoch": 1.7641521434931944, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.718475818634033, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8686342239379883, "num_tokens": 526360725.0, "step": 13868 }, { "epoch": 1.764279353771785, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5982933044433594, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8660739660263062, "num_tokens": 526398920.0, "step": 13869 }, { "epoch": 1.7644065640503754, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.564112663269043, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8874369859695435, "num_tokens": 526433453.0, "step": 13870 }, { "epoch": 1.7645337743289657, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5356085300445557, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8798325061798096, "num_tokens": 526476256.0, "step": 13871 }, { "epoch": 1.7646609846075563, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4628303050994873, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8745308518409729, "num_tokens": 526517707.0, "step": 13872 }, { "epoch": 1.7647881948861468, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5120344161987305, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8739615082740784, "num_tokens": 526551696.0, "step": 13873 }, { "epoch": 1.7649154051647373, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3984856605529785, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8826039433479309, "num_tokens": 526587884.0, "step": 13874 }, { "epoch": 1.7650426154433279, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.499873638153076, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8680371046066284, "num_tokens": 526625398.0, "step": 13875 }, { "epoch": 1.7651698257219184, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3458988666534424, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8698247671127319, "num_tokens": 526662481.0, "step": 13876 }, { "epoch": 1.7652970360005087, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.5038695335388184, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.862484335899353, "num_tokens": 526700719.0, "step": 13877 }, { "epoch": 1.7654242462790992, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.312701463699341, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8718209862709045, "num_tokens": 526732941.0, "step": 13878 }, { "epoch": 1.7655514565576897, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5103745460510254, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8745191097259521, "num_tokens": 526777825.0, "step": 13879 }, { "epoch": 1.7656786668362803, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2842493057250977, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8661153316497803, "num_tokens": 526811331.0, "step": 13880 }, { "epoch": 1.7658058771148708, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5555386543273926, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8728684186935425, "num_tokens": 526846788.0, "step": 13881 }, { "epoch": 1.7659330873934613, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.216123580932617, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8785731792449951, "num_tokens": 526880127.0, "step": 13882 }, { "epoch": 1.7660602976720519, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.653841495513916, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.862445592880249, "num_tokens": 526914099.0, "step": 13883 }, { "epoch": 1.7661875079506424, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.144113540649414, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8646633625030518, "num_tokens": 526956409.0, "step": 13884 }, { "epoch": 1.766314718229233, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.822486400604248, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8710291981697083, "num_tokens": 526995175.0, "step": 13885 }, { "epoch": 1.7664419285078234, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.083101749420166, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8719359636306763, "num_tokens": 527035833.0, "step": 13886 }, { "epoch": 1.766569138786414, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.084229469299316, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8722106218338013, "num_tokens": 527078780.0, "step": 13887 }, { "epoch": 1.7666963490650045, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.155102491378784, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8795638084411621, "num_tokens": 527119494.0, "step": 13888 }, { "epoch": 1.766823559343595, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.361037731170654, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8707103133201599, "num_tokens": 527158112.0, "step": 13889 }, { "epoch": 1.7669507696221856, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.4801478385925293, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8750631809234619, "num_tokens": 527195885.0, "step": 13890 }, { "epoch": 1.767077979900776, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.384671211242676, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8719639778137207, "num_tokens": 527229681.0, "step": 13891 }, { "epoch": 1.7672051901793666, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.765380859375, "learning_rate": 1e-05, "loss": 0.5222, "mean_token_accuracy": 0.8483567237854004, "num_tokens": 527267248.0, "step": 13892 }, { "epoch": 1.7673324004579571, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.177975654602051, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8723019361495972, "num_tokens": 527299695.0, "step": 13893 }, { "epoch": 1.7674596107365477, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.800794839859009, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8690959215164185, "num_tokens": 527334398.0, "step": 13894 }, { "epoch": 1.767586821015138, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9562535285949707, "learning_rate": 1e-05, "loss": 0.5083, "mean_token_accuracy": 0.8582251071929932, "num_tokens": 527376954.0, "step": 13895 }, { "epoch": 1.7677140312937285, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.690934896469116, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8698552250862122, "num_tokens": 527419737.0, "step": 13896 }, { "epoch": 1.767841241572319, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.804443120956421, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.8526023626327515, "num_tokens": 527455874.0, "step": 13897 }, { "epoch": 1.7679684518509096, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5806562900543213, "learning_rate": 1e-05, "loss": 0.5273, "mean_token_accuracy": 0.8465498685836792, "num_tokens": 527494796.0, "step": 13898 }, { "epoch": 1.7680956621295, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.714710235595703, "learning_rate": 1e-05, "loss": 0.5093, "mean_token_accuracy": 0.8540263175964355, "num_tokens": 527536955.0, "step": 13899 }, { "epoch": 1.7682228724080906, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4853110313415527, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8686092495918274, "num_tokens": 527569472.0, "step": 13900 }, { "epoch": 1.768350082686681, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6732900142669678, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.873786211013794, "num_tokens": 527603641.0, "step": 13901 }, { "epoch": 1.7684772929652715, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.372523784637451, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8640339374542236, "num_tokens": 527642942.0, "step": 13902 }, { "epoch": 1.768604503243862, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7008256912231445, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.865955650806427, "num_tokens": 527682687.0, "step": 13903 }, { "epoch": 1.7687317135224525, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.280785322189331, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.883144736289978, "num_tokens": 527721355.0, "step": 13904 }, { "epoch": 1.768858923801043, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8088526725769043, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8738521337509155, "num_tokens": 527760887.0, "step": 13905 }, { "epoch": 1.7689861340796336, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2092080116271973, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.860633134841919, "num_tokens": 527808669.0, "step": 13906 }, { "epoch": 1.769113344358224, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.012115955352783, "learning_rate": 1e-05, "loss": 0.5076, "mean_token_accuracy": 0.859228789806366, "num_tokens": 527840262.0, "step": 13907 }, { "epoch": 1.7692405546368146, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2685658931732178, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8605204820632935, "num_tokens": 527880822.0, "step": 13908 }, { "epoch": 1.7693677649154052, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.144371032714844, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8765385150909424, "num_tokens": 527923677.0, "step": 13909 }, { "epoch": 1.7694949751939957, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4400229454040527, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8576126098632812, "num_tokens": 527966004.0, "step": 13910 }, { "epoch": 1.7696221854725862, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.135324001312256, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8708887100219727, "num_tokens": 528004725.0, "step": 13911 }, { "epoch": 1.7697493957511767, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.616626739501953, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8699543476104736, "num_tokens": 528043262.0, "step": 13912 }, { "epoch": 1.7698766060297673, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9384796619415283, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8663570284843445, "num_tokens": 528082810.0, "step": 13913 }, { "epoch": 1.7700038163083578, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6750400066375732, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8651970624923706, "num_tokens": 528117672.0, "step": 13914 }, { "epoch": 1.7701310265869483, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.692307710647583, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8705345392227173, "num_tokens": 528155756.0, "step": 13915 }, { "epoch": 1.7702582368655388, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6471028327941895, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8791911005973816, "num_tokens": 528194664.0, "step": 13916 }, { "epoch": 1.7703854471441294, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4855966567993164, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8820269107818604, "num_tokens": 528234334.0, "step": 13917 }, { "epoch": 1.77051265742272, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6400692462921143, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8866760730743408, "num_tokens": 528270543.0, "step": 13918 }, { "epoch": 1.7706398677013102, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.341956377029419, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8639060854911804, "num_tokens": 528309698.0, "step": 13919 }, { "epoch": 1.7707670779799007, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6725587844848633, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8656410574913025, "num_tokens": 528356396.0, "step": 13920 }, { "epoch": 1.7708942882584913, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2469072341918945, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8641417026519775, "num_tokens": 528394542.0, "step": 13921 }, { "epoch": 1.7710214985370818, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7630820274353027, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8829861879348755, "num_tokens": 528435433.0, "step": 13922 }, { "epoch": 1.7711487088156723, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1763522624969482, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8860522508621216, "num_tokens": 528471152.0, "step": 13923 }, { "epoch": 1.7712759190942629, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.924241542816162, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8800404667854309, "num_tokens": 528505084.0, "step": 13924 }, { "epoch": 1.7714031293728534, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1934516429901123, "learning_rate": 1e-05, "loss": 0.5347, "mean_token_accuracy": 0.8439440131187439, "num_tokens": 528545234.0, "step": 13925 }, { "epoch": 1.7715303396514437, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.088629245758057, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8755790591239929, "num_tokens": 528584099.0, "step": 13926 }, { "epoch": 1.7716575499300342, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.306694507598877, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8628592491149902, "num_tokens": 528624143.0, "step": 13927 }, { "epoch": 1.7717847602086247, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.183165550231934, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.870632529258728, "num_tokens": 528661521.0, "step": 13928 }, { "epoch": 1.7719119704872153, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4995875358581543, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.882369875907898, "num_tokens": 528699317.0, "step": 13929 }, { "epoch": 1.7720391807658058, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.124007701873779, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.8572490215301514, "num_tokens": 528737277.0, "step": 13930 }, { "epoch": 1.7721663910443963, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.63165020942688, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8596036434173584, "num_tokens": 528777076.0, "step": 13931 }, { "epoch": 1.7722936013229869, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.937303066253662, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8723322749137878, "num_tokens": 528813940.0, "step": 13932 }, { "epoch": 1.7724208116015774, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.62363600730896, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8897738456726074, "num_tokens": 528852883.0, "step": 13933 }, { "epoch": 1.772548021880168, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7708559036254883, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8610236644744873, "num_tokens": 528891561.0, "step": 13934 }, { "epoch": 1.7726752321587584, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.574610471725464, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8675411939620972, "num_tokens": 528930755.0, "step": 13935 }, { "epoch": 1.772802442437349, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6435482501983643, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8660697937011719, "num_tokens": 528969230.0, "step": 13936 }, { "epoch": 1.7729296527159395, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.499565839767456, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8760104179382324, "num_tokens": 529005157.0, "step": 13937 }, { "epoch": 1.77305686299453, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5632951259613037, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8770910501480103, "num_tokens": 529042934.0, "step": 13938 }, { "epoch": 1.7731840732731206, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4220736026763916, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.874067485332489, "num_tokens": 529087811.0, "step": 13939 }, { "epoch": 1.773311283551711, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5559394359588623, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.861961841583252, "num_tokens": 529133143.0, "step": 13940 }, { "epoch": 1.7734384938303016, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3509938716888428, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.884247899055481, "num_tokens": 529172737.0, "step": 13941 }, { "epoch": 1.7735657041088921, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.592620849609375, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.872289776802063, "num_tokens": 529208913.0, "step": 13942 }, { "epoch": 1.7736929143874827, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.270327091217041, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8673461079597473, "num_tokens": 529250905.0, "step": 13943 }, { "epoch": 1.773820124666073, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.672218084335327, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8883694410324097, "num_tokens": 529289542.0, "step": 13944 }, { "epoch": 1.7739473349446635, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2082808017730713, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8789908289909363, "num_tokens": 529324759.0, "step": 13945 }, { "epoch": 1.774074545223254, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8174986839294434, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8717188835144043, "num_tokens": 529361845.0, "step": 13946 }, { "epoch": 1.7742017555018446, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1739766597747803, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8715291023254395, "num_tokens": 529402220.0, "step": 13947 }, { "epoch": 1.774328965780435, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.013294696807861, "learning_rate": 1e-05, "loss": 0.4999, "mean_token_accuracy": 0.8578898906707764, "num_tokens": 529440532.0, "step": 13948 }, { "epoch": 1.7744561760590256, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.234896421432495, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8758764266967773, "num_tokens": 529480357.0, "step": 13949 }, { "epoch": 1.774583386337616, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.169281959533691, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8653851747512817, "num_tokens": 529514075.0, "step": 13950 }, { "epoch": 1.7747105966162064, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4485599994659424, "learning_rate": 1e-05, "loss": 0.5087, "mean_token_accuracy": 0.8578430414199829, "num_tokens": 529550738.0, "step": 13951 }, { "epoch": 1.774837806894797, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.118976593017578, "learning_rate": 1e-05, "loss": 0.4945, "mean_token_accuracy": 0.8635466694831848, "num_tokens": 529594151.0, "step": 13952 }, { "epoch": 1.7749650171733875, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.593759059906006, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8742817640304565, "num_tokens": 529633631.0, "step": 13953 }, { "epoch": 1.775092227451978, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.943335771560669, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8683242797851562, "num_tokens": 529674377.0, "step": 13954 }, { "epoch": 1.7752194377305686, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6462104320526123, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8592672348022461, "num_tokens": 529707209.0, "step": 13955 }, { "epoch": 1.775346648009159, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7387032508850098, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.868506908416748, "num_tokens": 529737546.0, "step": 13956 }, { "epoch": 1.7754738582877496, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.612516164779663, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8715242147445679, "num_tokens": 529775532.0, "step": 13957 }, { "epoch": 1.7756010685663401, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5476884841918945, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8777755498886108, "num_tokens": 529814705.0, "step": 13958 }, { "epoch": 1.7757282788449307, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5776097774505615, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8709391951560974, "num_tokens": 529850912.0, "step": 13959 }, { "epoch": 1.7758554891235212, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4266538619995117, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8786137104034424, "num_tokens": 529889042.0, "step": 13960 }, { "epoch": 1.7759826994021117, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5636496543884277, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8732409477233887, "num_tokens": 529926749.0, "step": 13961 }, { "epoch": 1.7761099096807023, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.370762825012207, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8748130798339844, "num_tokens": 529951775.0, "step": 13962 }, { "epoch": 1.7762371199592928, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.545349359512329, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8861967325210571, "num_tokens": 529996172.0, "step": 13963 }, { "epoch": 1.7763643302378833, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3362088203430176, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8529360294342041, "num_tokens": 530039577.0, "step": 13964 }, { "epoch": 1.7764915405164738, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5547337532043457, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8810532093048096, "num_tokens": 530076516.0, "step": 13965 }, { "epoch": 1.7766187507950644, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2612850666046143, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8817940354347229, "num_tokens": 530118552.0, "step": 13966 }, { "epoch": 1.776745961073655, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.619955539703369, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8643556237220764, "num_tokens": 530154971.0, "step": 13967 }, { "epoch": 1.7768731713522452, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.20617938041687, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8840386867523193, "num_tokens": 530190509.0, "step": 13968 }, { "epoch": 1.7770003816308357, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.729271411895752, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8773517608642578, "num_tokens": 530229814.0, "step": 13969 }, { "epoch": 1.7771275919094263, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1435134410858154, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8661519885063171, "num_tokens": 530269426.0, "step": 13970 }, { "epoch": 1.7772548021880168, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.918243408203125, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8645731806755066, "num_tokens": 530305036.0, "step": 13971 }, { "epoch": 1.7773820124666073, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1287996768951416, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8688987493515015, "num_tokens": 530344995.0, "step": 13972 }, { "epoch": 1.7775092227451978, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.1679582595825195, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8774950504302979, "num_tokens": 530378283.0, "step": 13973 }, { "epoch": 1.7776364330237884, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.283851146697998, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.8537577390670776, "num_tokens": 530416147.0, "step": 13974 }, { "epoch": 1.7777636433023787, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.31181526184082, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8696285486221313, "num_tokens": 530452096.0, "step": 13975 }, { "epoch": 1.7778908535809692, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.55702543258667, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8598412871360779, "num_tokens": 530489347.0, "step": 13976 }, { "epoch": 1.7780180638595597, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.231403827667236, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8668910264968872, "num_tokens": 530528265.0, "step": 13977 }, { "epoch": 1.7781452741381503, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.704986810684204, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.882630467414856, "num_tokens": 530567501.0, "step": 13978 }, { "epoch": 1.7782724844167408, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.030508995056152, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8689112067222595, "num_tokens": 530610038.0, "step": 13979 }, { "epoch": 1.7783996946953313, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6930410861968994, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8669740557670593, "num_tokens": 530649039.0, "step": 13980 }, { "epoch": 1.7785269049739219, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8465142250061035, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8768537044525146, "num_tokens": 530687142.0, "step": 13981 }, { "epoch": 1.7786541152525124, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6111974716186523, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8804406523704529, "num_tokens": 530724399.0, "step": 13982 }, { "epoch": 1.778781325531103, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7000441551208496, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.871059775352478, "num_tokens": 530764552.0, "step": 13983 }, { "epoch": 1.7789085358096934, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5214450359344482, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8658750057220459, "num_tokens": 530808587.0, "step": 13984 }, { "epoch": 1.779035746088284, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.636775493621826, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8673382997512817, "num_tokens": 530845887.0, "step": 13985 }, { "epoch": 1.7791629563668745, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4422240257263184, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8599090576171875, "num_tokens": 530887398.0, "step": 13986 }, { "epoch": 1.779290166645465, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.621206760406494, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8811937570571899, "num_tokens": 530916377.0, "step": 13987 }, { "epoch": 1.7794173769240555, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.373652935028076, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8599420785903931, "num_tokens": 530954651.0, "step": 13988 }, { "epoch": 1.779544587202646, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6262288093566895, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8599216341972351, "num_tokens": 530999406.0, "step": 13989 }, { "epoch": 1.7796717974812366, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.281193971633911, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8759932518005371, "num_tokens": 531036925.0, "step": 13990 }, { "epoch": 1.7797990077598271, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7114298343658447, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8729464411735535, "num_tokens": 531076168.0, "step": 13991 }, { "epoch": 1.7799262180384177, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2138009071350098, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.874199628829956, "num_tokens": 531113217.0, "step": 13992 }, { "epoch": 1.780053428317008, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.844758987426758, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8695952892303467, "num_tokens": 531155938.0, "step": 13993 }, { "epoch": 1.7801806385955985, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.193587064743042, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8733835220336914, "num_tokens": 531194556.0, "step": 13994 }, { "epoch": 1.780307848874189, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.025206565856934, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.8588292002677917, "num_tokens": 531230704.0, "step": 13995 }, { "epoch": 1.7804350591527796, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.277641773223877, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8736076354980469, "num_tokens": 531270672.0, "step": 13996 }, { "epoch": 1.78056226943137, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.149112224578857, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8820087909698486, "num_tokens": 531302043.0, "step": 13997 }, { "epoch": 1.7806894797099606, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.4649717807769775, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8852231502532959, "num_tokens": 531337014.0, "step": 13998 }, { "epoch": 1.780816689988551, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.087928295135498, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8758736252784729, "num_tokens": 531371537.0, "step": 13999 }, { "epoch": 1.7809439002671414, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6044187545776367, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8818863034248352, "num_tokens": 531406171.0, "step": 14000 }, { "epoch": 1.781071110545732, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8938357830047607, "learning_rate": 1e-05, "loss": 0.5094, "mean_token_accuracy": 0.8578795194625854, "num_tokens": 531445625.0, "step": 14001 }, { "epoch": 1.7811983208243225, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6208486557006836, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8807159662246704, "num_tokens": 531490748.0, "step": 14002 }, { "epoch": 1.781325531102913, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6935977935791016, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8713040351867676, "num_tokens": 531526142.0, "step": 14003 }, { "epoch": 1.7814527413815036, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.595012903213501, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.874864399433136, "num_tokens": 531568325.0, "step": 14004 }, { "epoch": 1.781579951660094, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.522486448287964, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8682868480682373, "num_tokens": 531608898.0, "step": 14005 }, { "epoch": 1.7817071619386846, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.580815315246582, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8719996213912964, "num_tokens": 531640860.0, "step": 14006 }, { "epoch": 1.7818343722172751, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.393256902694702, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8755842447280884, "num_tokens": 531680121.0, "step": 14007 }, { "epoch": 1.7819615824958657, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6058719158172607, "learning_rate": 1e-05, "loss": 0.5052, "mean_token_accuracy": 0.8534969091415405, "num_tokens": 531713496.0, "step": 14008 }, { "epoch": 1.7820887927744562, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.30306339263916, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8637337684631348, "num_tokens": 531749989.0, "step": 14009 }, { "epoch": 1.7822160030530467, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.641087532043457, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8685271143913269, "num_tokens": 531788032.0, "step": 14010 }, { "epoch": 1.7823432133316373, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.237004280090332, "learning_rate": 1e-05, "loss": 0.5049, "mean_token_accuracy": 0.8485677242279053, "num_tokens": 531822178.0, "step": 14011 }, { "epoch": 1.7824704236102278, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.73697829246521, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.864082932472229, "num_tokens": 531855713.0, "step": 14012 }, { "epoch": 1.7825976338888183, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1679625511169434, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.859235942363739, "num_tokens": 531898714.0, "step": 14013 }, { "epoch": 1.7827248441674088, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8976426124572754, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8668861389160156, "num_tokens": 531933968.0, "step": 14014 }, { "epoch": 1.7828520544459994, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1532273292541504, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8715161681175232, "num_tokens": 531980672.0, "step": 14015 }, { "epoch": 1.78297926472459, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.099179267883301, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8733529448509216, "num_tokens": 532022461.0, "step": 14016 }, { "epoch": 1.7831064750031802, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.267608165740967, "learning_rate": 1e-05, "loss": 0.3983, "mean_token_accuracy": 0.8880547285079956, "num_tokens": 532059090.0, "step": 14017 }, { "epoch": 1.7832336852817707, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.245025157928467, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.8604965806007385, "num_tokens": 532095926.0, "step": 14018 }, { "epoch": 1.7833608955603613, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5191447734832764, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8733698129653931, "num_tokens": 532131623.0, "step": 14019 }, { "epoch": 1.7834881058389518, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.175560474395752, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.854841411113739, "num_tokens": 532171594.0, "step": 14020 }, { "epoch": 1.7836153161175423, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6504268646240234, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.878901481628418, "num_tokens": 532209884.0, "step": 14021 }, { "epoch": 1.7837425263961328, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9929039478302, "learning_rate": 1e-05, "loss": 0.508, "mean_token_accuracy": 0.8617105484008789, "num_tokens": 532246182.0, "step": 14022 }, { "epoch": 1.7838697366747234, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6640982627868652, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8726493120193481, "num_tokens": 532281133.0, "step": 14023 }, { "epoch": 1.7839969469533137, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.798444986343384, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8672181367874146, "num_tokens": 532318593.0, "step": 14024 }, { "epoch": 1.7841241572319042, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6092288494110107, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8599149584770203, "num_tokens": 532353473.0, "step": 14025 }, { "epoch": 1.7842513675104947, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6346819400787354, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8572165966033936, "num_tokens": 532395462.0, "step": 14026 }, { "epoch": 1.7843785777890853, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5381555557250977, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8769568204879761, "num_tokens": 532434677.0, "step": 14027 }, { "epoch": 1.7845057880676758, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5565695762634277, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8670324087142944, "num_tokens": 532470585.0, "step": 14028 }, { "epoch": 1.7846329983462663, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.483477830886841, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8635762929916382, "num_tokens": 532502657.0, "step": 14029 }, { "epoch": 1.7847602086248568, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.482919931411743, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.86214280128479, "num_tokens": 532548372.0, "step": 14030 }, { "epoch": 1.7848874189034474, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4366133213043213, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8608906269073486, "num_tokens": 532584707.0, "step": 14031 }, { "epoch": 1.785014629182038, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4616732597351074, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8698863983154297, "num_tokens": 532623179.0, "step": 14032 }, { "epoch": 1.7851418394606284, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3831729888916016, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8730647563934326, "num_tokens": 532663054.0, "step": 14033 }, { "epoch": 1.785269049739219, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4693763256073, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8683215975761414, "num_tokens": 532696456.0, "step": 14034 }, { "epoch": 1.7853962600178095, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3340110778808594, "learning_rate": 1e-05, "loss": 0.394, "mean_token_accuracy": 0.8847479224205017, "num_tokens": 532730726.0, "step": 14035 }, { "epoch": 1.7855234702964, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.50049090385437, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8688567876815796, "num_tokens": 532765561.0, "step": 14036 }, { "epoch": 1.7856506805749905, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.279343605041504, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8787862062454224, "num_tokens": 532803225.0, "step": 14037 }, { "epoch": 1.785777890853581, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.570213556289673, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8698486089706421, "num_tokens": 532843284.0, "step": 14038 }, { "epoch": 1.7859051011321716, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2308788299560547, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.866267204284668, "num_tokens": 532877637.0, "step": 14039 }, { "epoch": 1.7860323114107621, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.675999164581299, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8708541393280029, "num_tokens": 532914678.0, "step": 14040 }, { "epoch": 1.7861595216893527, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1532418727874756, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8693050146102905, "num_tokens": 532952309.0, "step": 14041 }, { "epoch": 1.786286731967943, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.854116916656494, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8774648904800415, "num_tokens": 532986440.0, "step": 14042 }, { "epoch": 1.7864139422465335, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.124042272567749, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8639290928840637, "num_tokens": 533020488.0, "step": 14043 }, { "epoch": 1.786541152525124, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.094414710998535, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8747410178184509, "num_tokens": 533052535.0, "step": 14044 }, { "epoch": 1.7866683628037145, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.1778674125671387, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8782663345336914, "num_tokens": 533092924.0, "step": 14045 }, { "epoch": 1.786795573082305, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.353843688964844, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8666585087776184, "num_tokens": 533130497.0, "step": 14046 }, { "epoch": 1.7869227833608956, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.507093906402588, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.8502774238586426, "num_tokens": 533170012.0, "step": 14047 }, { "epoch": 1.787049993639486, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.3325090408325195, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8689679503440857, "num_tokens": 533207433.0, "step": 14048 }, { "epoch": 1.7871772039180764, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.730642080307007, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8848796486854553, "num_tokens": 533248809.0, "step": 14049 }, { "epoch": 1.787304414196667, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1352925300598145, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8790286183357239, "num_tokens": 533283381.0, "step": 14050 }, { "epoch": 1.7874316244752575, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.757162094116211, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8810210227966309, "num_tokens": 533320951.0, "step": 14051 }, { "epoch": 1.787558834753848, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9266979694366455, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8723998069763184, "num_tokens": 533361080.0, "step": 14052 }, { "epoch": 1.7876860450324386, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6664369106292725, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8834604620933533, "num_tokens": 533396448.0, "step": 14053 }, { "epoch": 1.787813255311029, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.783998727798462, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8720473051071167, "num_tokens": 533434924.0, "step": 14054 }, { "epoch": 1.7879404655896196, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5635740756988525, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8732810020446777, "num_tokens": 533471248.0, "step": 14055 }, { "epoch": 1.7880676758682101, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.679060697555542, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8799452781677246, "num_tokens": 533512521.0, "step": 14056 }, { "epoch": 1.7881948861468007, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4666831493377686, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8782989978790283, "num_tokens": 533550385.0, "step": 14057 }, { "epoch": 1.7883220964253912, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.677716016769409, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8725354671478271, "num_tokens": 533580048.0, "step": 14058 }, { "epoch": 1.7884493067039817, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3621766567230225, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8822088837623596, "num_tokens": 533615115.0, "step": 14059 }, { "epoch": 1.7885765169825723, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.704181432723999, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8730149865150452, "num_tokens": 533653774.0, "step": 14060 }, { "epoch": 1.7887037272611628, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.276163339614868, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8831619024276733, "num_tokens": 533692628.0, "step": 14061 }, { "epoch": 1.7888309375397533, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.802800178527832, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.889549970626831, "num_tokens": 533734935.0, "step": 14062 }, { "epoch": 1.7889581478183438, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2309811115264893, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8664611577987671, "num_tokens": 533772541.0, "step": 14063 }, { "epoch": 1.7890853580969344, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.954286813735962, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8771152496337891, "num_tokens": 533816036.0, "step": 14064 }, { "epoch": 1.789212568375525, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2775161266326904, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8706576824188232, "num_tokens": 533848174.0, "step": 14065 }, { "epoch": 1.7893397786541152, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0645060539245605, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.859697699546814, "num_tokens": 533890011.0, "step": 14066 }, { "epoch": 1.7894669889327057, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4015345573425293, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.860427975654602, "num_tokens": 533927650.0, "step": 14067 }, { "epoch": 1.7895941992112963, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.079826831817627, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8646309971809387, "num_tokens": 533959117.0, "step": 14068 }, { "epoch": 1.7897214094898868, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5422489643096924, "learning_rate": 1e-05, "loss": 0.4977, "mean_token_accuracy": 0.8601646423339844, "num_tokens": 533995396.0, "step": 14069 }, { "epoch": 1.7898486197684773, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.93066668510437, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8814552426338196, "num_tokens": 534032498.0, "step": 14070 }, { "epoch": 1.7899758300470678, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6049411296844482, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.86269211769104, "num_tokens": 534071790.0, "step": 14071 }, { "epoch": 1.7901030403256584, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7205636501312256, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8632074594497681, "num_tokens": 534112778.0, "step": 14072 }, { "epoch": 1.7902302506042487, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.600898265838623, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8766525983810425, "num_tokens": 534150815.0, "step": 14073 }, { "epoch": 1.7903574608828392, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5359280109405518, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8675159811973572, "num_tokens": 534186148.0, "step": 14074 }, { "epoch": 1.7904846711614297, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.599553346633911, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8656572103500366, "num_tokens": 534219496.0, "step": 14075 }, { "epoch": 1.7906118814400203, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3800511360168457, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.877647340297699, "num_tokens": 534263093.0, "step": 14076 }, { "epoch": 1.7907390917186108, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6247403621673584, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8829753994941711, "num_tokens": 534298515.0, "step": 14077 }, { "epoch": 1.7908663019972013, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.282927989959717, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.874347448348999, "num_tokens": 534337210.0, "step": 14078 }, { "epoch": 1.7909935122757918, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.691105604171753, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8614608645439148, "num_tokens": 534373440.0, "step": 14079 }, { "epoch": 1.7911207225543824, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2058470249176025, "learning_rate": 1e-05, "loss": 0.3761, "mean_token_accuracy": 0.8947097659111023, "num_tokens": 534405560.0, "step": 14080 }, { "epoch": 1.791247932832973, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.80242919921875, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8732137680053711, "num_tokens": 534442824.0, "step": 14081 }, { "epoch": 1.7913751431115634, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1573452949523926, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8725610971450806, "num_tokens": 534482569.0, "step": 14082 }, { "epoch": 1.791502353390154, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.990482807159424, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8638572692871094, "num_tokens": 534518481.0, "step": 14083 }, { "epoch": 1.7916295636687445, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2111289501190186, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.8498796224594116, "num_tokens": 534553273.0, "step": 14084 }, { "epoch": 1.791756773947335, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.159919261932373, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8655794262886047, "num_tokens": 534587322.0, "step": 14085 }, { "epoch": 1.7918839842259255, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.364076614379883, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8808537721633911, "num_tokens": 534622999.0, "step": 14086 }, { "epoch": 1.792011194504516, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.226529598236084, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8631947636604309, "num_tokens": 534655953.0, "step": 14087 }, { "epoch": 1.7921384047831066, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5669753551483154, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8651629686355591, "num_tokens": 534699495.0, "step": 14088 }, { "epoch": 1.7922656150616971, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.105883598327637, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8795102834701538, "num_tokens": 534737650.0, "step": 14089 }, { "epoch": 1.7923928253402877, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.664301633834839, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8743846416473389, "num_tokens": 534771169.0, "step": 14090 }, { "epoch": 1.792520035618878, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.911928176879883, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8809906840324402, "num_tokens": 534802968.0, "step": 14091 }, { "epoch": 1.7926472458974685, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6322832107543945, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.874906063079834, "num_tokens": 534846975.0, "step": 14092 }, { "epoch": 1.792774456176059, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7359731197357178, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8746461868286133, "num_tokens": 534880969.0, "step": 14093 }, { "epoch": 1.7929016664546495, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5730254650115967, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.866156816482544, "num_tokens": 534918423.0, "step": 14094 }, { "epoch": 1.79302887673324, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.611283540725708, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.878665566444397, "num_tokens": 534952733.0, "step": 14095 }, { "epoch": 1.7931560870118306, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.517133951187134, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.877943754196167, "num_tokens": 534981868.0, "step": 14096 }, { "epoch": 1.793283297290421, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.540057420730591, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8694822192192078, "num_tokens": 535020940.0, "step": 14097 }, { "epoch": 1.7934105075690114, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4463298320770264, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8758885264396667, "num_tokens": 535053073.0, "step": 14098 }, { "epoch": 1.793537717847602, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.501070976257324, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8699436187744141, "num_tokens": 535097731.0, "step": 14099 }, { "epoch": 1.7936649281261925, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3860929012298584, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8648080825805664, "num_tokens": 535145571.0, "step": 14100 }, { "epoch": 1.793792138404783, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4988811016082764, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8676478266716003, "num_tokens": 535184226.0, "step": 14101 }, { "epoch": 1.7939193486833735, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3563387393951416, "learning_rate": 1e-05, "loss": 0.5145, "mean_token_accuracy": 0.8517778515815735, "num_tokens": 535221990.0, "step": 14102 }, { "epoch": 1.794046558961964, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5181381702423096, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8641201257705688, "num_tokens": 535261588.0, "step": 14103 }, { "epoch": 1.7941737692405546, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2937769889831543, "learning_rate": 1e-05, "loss": 0.5025, "mean_token_accuracy": 0.8535665273666382, "num_tokens": 535299023.0, "step": 14104 }, { "epoch": 1.7943009795191451, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5824975967407227, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8656376600265503, "num_tokens": 535332174.0, "step": 14105 }, { "epoch": 1.7944281897977357, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2286376953125, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8757350444793701, "num_tokens": 535367242.0, "step": 14106 }, { "epoch": 1.7945554000763262, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.690723419189453, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8731021881103516, "num_tokens": 535401962.0, "step": 14107 }, { "epoch": 1.7946826103549167, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.161910057067871, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8607600331306458, "num_tokens": 535441481.0, "step": 14108 }, { "epoch": 1.7948098206335072, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.869154930114746, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8780339360237122, "num_tokens": 535473619.0, "step": 14109 }, { "epoch": 1.7949370309120978, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1188719272613525, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8837943077087402, "num_tokens": 535515340.0, "step": 14110 }, { "epoch": 1.7950642411906883, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.124989032745361, "learning_rate": 1e-05, "loss": 0.5097, "mean_token_accuracy": 0.8556874394416809, "num_tokens": 535555090.0, "step": 14111 }, { "epoch": 1.7951914514692788, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2432966232299805, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8708314299583435, "num_tokens": 535595484.0, "step": 14112 }, { "epoch": 1.7953186617478694, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.306108474731445, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8689767122268677, "num_tokens": 535634170.0, "step": 14113 }, { "epoch": 1.7954458720264599, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5173263549804688, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8749620914459229, "num_tokens": 535671771.0, "step": 14114 }, { "epoch": 1.7955730823050502, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.262181758880615, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8788453340530396, "num_tokens": 535710280.0, "step": 14115 }, { "epoch": 1.7957002925836407, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7171671390533447, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8671263456344604, "num_tokens": 535749509.0, "step": 14116 }, { "epoch": 1.7958275028622313, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.042586326599121, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8782364130020142, "num_tokens": 535794184.0, "step": 14117 }, { "epoch": 1.7959547131408218, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.722693920135498, "learning_rate": 1e-05, "loss": 0.5144, "mean_token_accuracy": 0.8514434099197388, "num_tokens": 535828818.0, "step": 14118 }, { "epoch": 1.7960819234194123, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8508474826812744, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8644883632659912, "num_tokens": 535869461.0, "step": 14119 }, { "epoch": 1.7962091336980028, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.628661870956421, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8725463151931763, "num_tokens": 535907287.0, "step": 14120 }, { "epoch": 1.7963363439765934, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6988978385925293, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8795895576477051, "num_tokens": 535949063.0, "step": 14121 }, { "epoch": 1.7964635542551837, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.53653883934021, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8724403977394104, "num_tokens": 535987202.0, "step": 14122 }, { "epoch": 1.7965907645337742, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6227924823760986, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8811144828796387, "num_tokens": 536021790.0, "step": 14123 }, { "epoch": 1.7967179748123647, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4502406120300293, "learning_rate": 1e-05, "loss": 0.55, "mean_token_accuracy": 0.8429864645004272, "num_tokens": 536069932.0, "step": 14124 }, { "epoch": 1.7968451850909553, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.597835063934326, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8791872262954712, "num_tokens": 536110428.0, "step": 14125 }, { "epoch": 1.7969723953695458, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3606152534484863, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8678959012031555, "num_tokens": 536147711.0, "step": 14126 }, { "epoch": 1.7970996056481363, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.627110242843628, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8770930767059326, "num_tokens": 536184422.0, "step": 14127 }, { "epoch": 1.7972268159267268, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.274493455886841, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.8827966451644897, "num_tokens": 536224802.0, "step": 14128 }, { "epoch": 1.7973540262053174, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7091174125671387, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8666942119598389, "num_tokens": 536267596.0, "step": 14129 }, { "epoch": 1.797481236483908, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.21207857131958, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8747566938400269, "num_tokens": 536306130.0, "step": 14130 }, { "epoch": 1.7976084467624984, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.861720561981201, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8698034882545471, "num_tokens": 536345451.0, "step": 14131 }, { "epoch": 1.797735657041089, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.199218511581421, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.864836573600769, "num_tokens": 536377953.0, "step": 14132 }, { "epoch": 1.7978628673196795, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.0376362800598145, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8817238807678223, "num_tokens": 536410879.0, "step": 14133 }, { "epoch": 1.79799007759827, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2741858959198, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8584128022193909, "num_tokens": 536448351.0, "step": 14134 }, { "epoch": 1.7981172878768605, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.259698867797852, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8663711547851562, "num_tokens": 536488968.0, "step": 14135 }, { "epoch": 1.798244498155451, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.497546434402466, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8794222474098206, "num_tokens": 536529189.0, "step": 14136 }, { "epoch": 1.7983717084340416, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.127246379852295, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8863895535469055, "num_tokens": 536567778.0, "step": 14137 }, { "epoch": 1.7984989187126321, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6440389156341553, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8701785802841187, "num_tokens": 536609193.0, "step": 14138 }, { "epoch": 1.7986261289912227, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9406352043151855, "learning_rate": 1e-05, "loss": 0.5105, "mean_token_accuracy": 0.853995680809021, "num_tokens": 536646122.0, "step": 14139 }, { "epoch": 1.798753339269813, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.66552734375, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8690600395202637, "num_tokens": 536682146.0, "step": 14140 }, { "epoch": 1.7988805495484035, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.727085828781128, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8623447418212891, "num_tokens": 536715449.0, "step": 14141 }, { "epoch": 1.799007759826994, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.623692512512207, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8664064407348633, "num_tokens": 536752430.0, "step": 14142 }, { "epoch": 1.7991349701055845, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00026702880859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.561755657196045, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8576884269714355, "num_tokens": 536785577.0, "step": 14143 }, { "epoch": 1.799262180384175, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5761477947235107, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8823951482772827, "num_tokens": 536824074.0, "step": 14144 }, { "epoch": 1.7993893906627656, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4329206943511963, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8866307735443115, "num_tokens": 536865819.0, "step": 14145 }, { "epoch": 1.799516600941356, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.556504487991333, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8714920282363892, "num_tokens": 536905538.0, "step": 14146 }, { "epoch": 1.7996438112199464, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3594110012054443, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8801910281181335, "num_tokens": 536944895.0, "step": 14147 }, { "epoch": 1.799771021498537, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.576265573501587, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8677647113800049, "num_tokens": 536981094.0, "step": 14148 }, { "epoch": 1.7998982317771275, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3005545139312744, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8581593632698059, "num_tokens": 537020242.0, "step": 14149 }, { "epoch": 1.800025442055718, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6107442378997803, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8599599599838257, "num_tokens": 537062475.0, "step": 14150 }, { "epoch": 1.8001526523343085, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2364490032196045, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8667005896568298, "num_tokens": 537100969.0, "step": 14151 }, { "epoch": 1.800279862612899, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.704073667526245, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8694550395011902, "num_tokens": 537141828.0, "step": 14152 }, { "epoch": 1.8004070728914896, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.16355299949646, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8759346604347229, "num_tokens": 537181324.0, "step": 14153 }, { "epoch": 1.8005342831700801, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8536579608917236, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8756163120269775, "num_tokens": 537224593.0, "step": 14154 }, { "epoch": 1.8006614934486707, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.144329309463501, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8655010461807251, "num_tokens": 537262835.0, "step": 14155 }, { "epoch": 1.8007887037272612, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.075183391571045, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8667272329330444, "num_tokens": 537299722.0, "step": 14156 }, { "epoch": 1.8009159140058517, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.211099147796631, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8744087219238281, "num_tokens": 537345445.0, "step": 14157 }, { "epoch": 1.8010431242844422, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.279314994812012, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8735896944999695, "num_tokens": 537382327.0, "step": 14158 }, { "epoch": 1.8011703345630328, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4748358726501465, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8707042932510376, "num_tokens": 537419013.0, "step": 14159 }, { "epoch": 1.8012975448416233, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.256271839141846, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8691016435623169, "num_tokens": 537459841.0, "step": 14160 }, { "epoch": 1.8014247551202138, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.679326295852661, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8825904130935669, "num_tokens": 537494124.0, "step": 14161 }, { "epoch": 1.8015519653988044, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.073238372802734, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.876811146736145, "num_tokens": 537532798.0, "step": 14162 }, { "epoch": 1.8016791756773949, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.711400032043457, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8705861568450928, "num_tokens": 537568548.0, "step": 14163 }, { "epoch": 1.8018063859559852, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8601155281066895, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8853934407234192, "num_tokens": 537604413.0, "step": 14164 }, { "epoch": 1.8019335962345757, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6418302059173584, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8720325231552124, "num_tokens": 537640588.0, "step": 14165 }, { "epoch": 1.8020608065131662, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6916754245758057, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8629910349845886, "num_tokens": 537682664.0, "step": 14166 }, { "epoch": 1.8021880167917568, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.580740451812744, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8544409275054932, "num_tokens": 537716020.0, "step": 14167 }, { "epoch": 1.8023152270703473, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5785434246063232, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8767093420028687, "num_tokens": 537756895.0, "step": 14168 }, { "epoch": 1.8024424373489378, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5118637084960938, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8666419982910156, "num_tokens": 537794130.0, "step": 14169 }, { "epoch": 1.8025696476275284, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.530338764190674, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8846032023429871, "num_tokens": 537822865.0, "step": 14170 }, { "epoch": 1.8026968579061187, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4397363662719727, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8764515519142151, "num_tokens": 537854594.0, "step": 14171 }, { "epoch": 1.8028240681847092, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4916200637817383, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8717526197433472, "num_tokens": 537889667.0, "step": 14172 }, { "epoch": 1.8029512784632997, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.391227960586548, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8596535325050354, "num_tokens": 537924322.0, "step": 14173 }, { "epoch": 1.8030784887418903, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4844610691070557, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8806999921798706, "num_tokens": 537962196.0, "step": 14174 }, { "epoch": 1.8032056990204808, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3331120014190674, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8795580267906189, "num_tokens": 537999916.0, "step": 14175 }, { "epoch": 1.8033329092990713, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5151376724243164, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8823937177658081, "num_tokens": 538035823.0, "step": 14176 }, { "epoch": 1.8034601195776618, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2710578441619873, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8614321351051331, "num_tokens": 538080664.0, "step": 14177 }, { "epoch": 1.8035873298562524, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.593168258666992, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.864922046661377, "num_tokens": 538123178.0, "step": 14178 }, { "epoch": 1.803714540134843, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1963512897491455, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8750390410423279, "num_tokens": 538158757.0, "step": 14179 }, { "epoch": 1.8038417504134334, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.738558053970337, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8668892979621887, "num_tokens": 538199019.0, "step": 14180 }, { "epoch": 1.803968960692024, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.114997148513794, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8710033297538757, "num_tokens": 538239839.0, "step": 14181 }, { "epoch": 1.8040961709706145, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9576005935668945, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8735744953155518, "num_tokens": 538287677.0, "step": 14182 }, { "epoch": 1.804223381249205, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1272456645965576, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8721258640289307, "num_tokens": 538322304.0, "step": 14183 }, { "epoch": 1.8043505915277955, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.222908973693848, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8772396445274353, "num_tokens": 538364087.0, "step": 14184 }, { "epoch": 1.804477801806386, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3146872520446777, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8692017793655396, "num_tokens": 538401507.0, "step": 14185 }, { "epoch": 1.8046050120849766, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.372797012329102, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8756144642829895, "num_tokens": 538440666.0, "step": 14186 }, { "epoch": 1.8047322223635671, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6453394889831543, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8631706833839417, "num_tokens": 538478915.0, "step": 14187 }, { "epoch": 1.8048594326421576, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.231074810028076, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8736329078674316, "num_tokens": 538517770.0, "step": 14188 }, { "epoch": 1.804986642920748, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7761034965515137, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8593120574951172, "num_tokens": 538555812.0, "step": 14189 }, { "epoch": 1.8051138531993385, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9947948455810547, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8857223987579346, "num_tokens": 538596023.0, "step": 14190 }, { "epoch": 1.805241063477929, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.711869716644287, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.8854766488075256, "num_tokens": 538630840.0, "step": 14191 }, { "epoch": 1.8053682737565195, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8324360847473145, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8632432222366333, "num_tokens": 538668515.0, "step": 14192 }, { "epoch": 1.80549548403511, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6210615634918213, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8618077039718628, "num_tokens": 538703996.0, "step": 14193 }, { "epoch": 1.8056226943137006, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6944308280944824, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8719435930252075, "num_tokens": 538742819.0, "step": 14194 }, { "epoch": 1.805749904592291, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5149800777435303, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8808048963546753, "num_tokens": 538776218.0, "step": 14195 }, { "epoch": 1.8058771148708814, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.637181520462036, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.871180534362793, "num_tokens": 538813377.0, "step": 14196 }, { "epoch": 1.806004325149472, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.41331148147583, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8682712316513062, "num_tokens": 538856755.0, "step": 14197 }, { "epoch": 1.8061315354280625, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6346864700317383, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.870471179485321, "num_tokens": 538898175.0, "step": 14198 }, { "epoch": 1.806258745706653, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.332418441772461, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8633512854576111, "num_tokens": 538935539.0, "step": 14199 }, { "epoch": 1.8063859559852435, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6969950199127197, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.8548493385314941, "num_tokens": 538978999.0, "step": 14200 }, { "epoch": 1.806513166263834, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.247967481613159, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8793767690658569, "num_tokens": 539013480.0, "step": 14201 }, { "epoch": 1.8066403765424246, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8366641998291016, "learning_rate": 1e-05, "loss": 0.4942, "mean_token_accuracy": 0.863003134727478, "num_tokens": 539047120.0, "step": 14202 }, { "epoch": 1.8067675868210151, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.195279836654663, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8814632892608643, "num_tokens": 539088243.0, "step": 14203 }, { "epoch": 1.8068947970996057, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.011856555938721, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8619642853736877, "num_tokens": 539128240.0, "step": 14204 }, { "epoch": 1.8070220073781962, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2616937160491943, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.870009183883667, "num_tokens": 539168792.0, "step": 14205 }, { "epoch": 1.8071492176567867, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.155388832092285, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8753100633621216, "num_tokens": 539210510.0, "step": 14206 }, { "epoch": 1.8072764279353772, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4532291889190674, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8738967180252075, "num_tokens": 539248926.0, "step": 14207 }, { "epoch": 1.8074036382139678, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.123163223266602, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8669427037239075, "num_tokens": 539285795.0, "step": 14208 }, { "epoch": 1.8075308484925583, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6099042892456055, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.877625584602356, "num_tokens": 539324891.0, "step": 14209 }, { "epoch": 1.8076580587711488, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.982454776763916, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8626186847686768, "num_tokens": 539364684.0, "step": 14210 }, { "epoch": 1.8077852690497394, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.652067184448242, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8791760206222534, "num_tokens": 539403312.0, "step": 14211 }, { "epoch": 1.8079124793283299, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.730764627456665, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8834340572357178, "num_tokens": 539443567.0, "step": 14212 }, { "epoch": 1.8080396896069202, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6306114196777344, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8651253581047058, "num_tokens": 539480969.0, "step": 14213 }, { "epoch": 1.8081668998855107, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5450446605682373, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8709478974342346, "num_tokens": 539518105.0, "step": 14214 }, { "epoch": 1.8082941101641012, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.604508399963379, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8570941686630249, "num_tokens": 539561364.0, "step": 14215 }, { "epoch": 1.8084213204426918, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4246163368225098, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8658605813980103, "num_tokens": 539594602.0, "step": 14216 }, { "epoch": 1.8085485307212823, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.593221664428711, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8579492568969727, "num_tokens": 539635985.0, "step": 14217 }, { "epoch": 1.8086757409998728, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3317530155181885, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8731420040130615, "num_tokens": 539675201.0, "step": 14218 }, { "epoch": 1.8088029512784631, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6114776134490967, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8796154260635376, "num_tokens": 539712661.0, "step": 14219 }, { "epoch": 1.8089301615570537, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.276782751083374, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8581238389015198, "num_tokens": 539749941.0, "step": 14220 }, { "epoch": 1.8090573718356442, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.668463945388794, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8669625520706177, "num_tokens": 539786906.0, "step": 14221 }, { "epoch": 1.8091845821142347, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2171528339385986, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8750522136688232, "num_tokens": 539824235.0, "step": 14222 }, { "epoch": 1.8093117923928252, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7688300609588623, "learning_rate": 1e-05, "loss": 0.4991, "mean_token_accuracy": 0.8604059219360352, "num_tokens": 539864326.0, "step": 14223 }, { "epoch": 1.8094390026714158, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1632866859436035, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8573305010795593, "num_tokens": 539909105.0, "step": 14224 }, { "epoch": 1.8095662129500063, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9413132667541504, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8714070320129395, "num_tokens": 539944846.0, "step": 14225 }, { "epoch": 1.8096934232285968, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.174473524093628, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.874521017074585, "num_tokens": 539984720.0, "step": 14226 }, { "epoch": 1.8098206335071874, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.131938934326172, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8640026450157166, "num_tokens": 540025002.0, "step": 14227 }, { "epoch": 1.8099478437857779, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3302509784698486, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8618699312210083, "num_tokens": 540060256.0, "step": 14228 }, { "epoch": 1.8100750540643684, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2222747802734375, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8696667551994324, "num_tokens": 540098073.0, "step": 14229 }, { "epoch": 1.810202264342959, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.547053575515747, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8713780045509338, "num_tokens": 540128704.0, "step": 14230 }, { "epoch": 1.8103294746215495, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1157426834106445, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8826957941055298, "num_tokens": 540161293.0, "step": 14231 }, { "epoch": 1.81045668490014, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.643650770187378, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8821446895599365, "num_tokens": 540193685.0, "step": 14232 }, { "epoch": 1.8105838951787305, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9221954345703125, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8822391629219055, "num_tokens": 540232217.0, "step": 14233 }, { "epoch": 1.810711105457321, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6290924549102783, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.86818528175354, "num_tokens": 540267561.0, "step": 14234 }, { "epoch": 1.8108383157359116, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7642822265625, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8786301016807556, "num_tokens": 540305788.0, "step": 14235 }, { "epoch": 1.8109655260145021, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5790202617645264, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8665601015090942, "num_tokens": 540341105.0, "step": 14236 }, { "epoch": 1.8110927362930926, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6287283897399902, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8668113350868225, "num_tokens": 540380502.0, "step": 14237 }, { "epoch": 1.811219946571683, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.516343593597412, "learning_rate": 1e-05, "loss": 0.5136, "mean_token_accuracy": 0.8517816066741943, "num_tokens": 540420561.0, "step": 14238 }, { "epoch": 1.8113471568502735, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5394093990325928, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8650290369987488, "num_tokens": 540460561.0, "step": 14239 }, { "epoch": 1.811474367128864, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4492783546447754, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8678945302963257, "num_tokens": 540498939.0, "step": 14240 }, { "epoch": 1.8116015774074545, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5114476680755615, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8627274036407471, "num_tokens": 540536127.0, "step": 14241 }, { "epoch": 1.811728787686045, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.393796920776367, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8824170827865601, "num_tokens": 540567812.0, "step": 14242 }, { "epoch": 1.8118559979646356, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5084452629089355, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8823057413101196, "num_tokens": 540604176.0, "step": 14243 }, { "epoch": 1.811983208243226, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.350294828414917, "learning_rate": 1e-05, "loss": 0.4943, "mean_token_accuracy": 0.85410475730896, "num_tokens": 540640989.0, "step": 14244 }, { "epoch": 1.8121104185218164, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.529752731323242, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8762425780296326, "num_tokens": 540680590.0, "step": 14245 }, { "epoch": 1.812237628800407, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.281559467315674, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8742372989654541, "num_tokens": 540723739.0, "step": 14246 }, { "epoch": 1.8123648390789975, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6049067974090576, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8720945119857788, "num_tokens": 540759686.0, "step": 14247 }, { "epoch": 1.812492049357588, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.212848424911499, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8847135305404663, "num_tokens": 540795710.0, "step": 14248 }, { "epoch": 1.8126192596361785, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7272074222564697, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8706673979759216, "num_tokens": 540838558.0, "step": 14249 }, { "epoch": 1.812746469914769, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.146688461303711, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8834344148635864, "num_tokens": 540874543.0, "step": 14250 }, { "epoch": 1.8128736801933596, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9283275604248047, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8725783228874207, "num_tokens": 540909903.0, "step": 14251 }, { "epoch": 1.8130008904719501, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.134172201156616, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8706156611442566, "num_tokens": 540947977.0, "step": 14252 }, { "epoch": 1.8131281007505406, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.177855968475342, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8608724474906921, "num_tokens": 540986288.0, "step": 14253 }, { "epoch": 1.8132553110291312, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.2915656566619873, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8724396228790283, "num_tokens": 541027607.0, "step": 14254 }, { "epoch": 1.8133825213077217, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.325899124145508, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8756661415100098, "num_tokens": 541064424.0, "step": 14255 }, { "epoch": 1.8135097315863122, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.5938167572021484, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8647326827049255, "num_tokens": 541097815.0, "step": 14256 }, { "epoch": 1.8136369418649028, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.20695686340332, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8635928630828857, "num_tokens": 541136567.0, "step": 14257 }, { "epoch": 1.8137641521434933, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7306439876556396, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8748529553413391, "num_tokens": 541174305.0, "step": 14258 }, { "epoch": 1.8138913624220838, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9938998222351074, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8717124462127686, "num_tokens": 541207118.0, "step": 14259 }, { "epoch": 1.8140185727006743, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7023019790649414, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8828697204589844, "num_tokens": 541242208.0, "step": 14260 }, { "epoch": 1.8141457829792649, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7935738563537598, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8747938871383667, "num_tokens": 541278553.0, "step": 14261 }, { "epoch": 1.8142729932578552, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.618927001953125, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8660411834716797, "num_tokens": 541318329.0, "step": 14262 }, { "epoch": 1.8144002035364457, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.660324811935425, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8782276511192322, "num_tokens": 541353406.0, "step": 14263 }, { "epoch": 1.8145274138150362, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.552067279815674, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8684048652648926, "num_tokens": 541385504.0, "step": 14264 }, { "epoch": 1.8146546240936268, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5608339309692383, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8641403913497925, "num_tokens": 541426172.0, "step": 14265 }, { "epoch": 1.8147818343722173, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4606707096099854, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8738656640052795, "num_tokens": 541464785.0, "step": 14266 }, { "epoch": 1.8149090446508078, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5343563556671143, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.863997220993042, "num_tokens": 541500271.0, "step": 14267 }, { "epoch": 1.8150362549293981, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.387464761734009, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8671796917915344, "num_tokens": 541537482.0, "step": 14268 }, { "epoch": 1.8151634652079887, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5406253337860107, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8746159672737122, "num_tokens": 541576669.0, "step": 14269 }, { "epoch": 1.8152906754865792, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3148581981658936, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8708672523498535, "num_tokens": 541616040.0, "step": 14270 }, { "epoch": 1.8154178857651697, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.600027322769165, "learning_rate": 1e-05, "loss": 0.4973, "mean_token_accuracy": 0.8549097180366516, "num_tokens": 541649994.0, "step": 14271 }, { "epoch": 1.8155450960437602, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2488396167755127, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8579949140548706, "num_tokens": 541688468.0, "step": 14272 }, { "epoch": 1.8156723063223508, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6971418857574463, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8677270412445068, "num_tokens": 541725173.0, "step": 14273 }, { "epoch": 1.8157995166009413, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.173597574234009, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8685851693153381, "num_tokens": 541762523.0, "step": 14274 }, { "epoch": 1.8159267268795318, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8639791011810303, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8692948818206787, "num_tokens": 541799731.0, "step": 14275 }, { "epoch": 1.8160539371581224, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.123898506164551, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8800767660140991, "num_tokens": 541844172.0, "step": 14276 }, { "epoch": 1.8161811474367129, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.095184326171875, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8716335296630859, "num_tokens": 541887526.0, "step": 14277 }, { "epoch": 1.8163083577153034, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2283639907836914, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8717168569564819, "num_tokens": 541932712.0, "step": 14278 }, { "epoch": 1.816435567993894, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.28986930847168, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8713592290878296, "num_tokens": 541968915.0, "step": 14279 }, { "epoch": 1.8165627782724845, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.520902395248413, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8650742769241333, "num_tokens": 542010296.0, "step": 14280 }, { "epoch": 1.816689988551075, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.223331451416016, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8869354724884033, "num_tokens": 542050096.0, "step": 14281 }, { "epoch": 1.8168171988296655, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7127487659454346, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8656649589538574, "num_tokens": 542086746.0, "step": 14282 }, { "epoch": 1.816944409108256, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.012554168701172, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8761143684387207, "num_tokens": 542123095.0, "step": 14283 }, { "epoch": 1.8170716193868466, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.702270746231079, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8756297826766968, "num_tokens": 542157797.0, "step": 14284 }, { "epoch": 1.817198829665437, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.806422233581543, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8637341260910034, "num_tokens": 542196091.0, "step": 14285 }, { "epoch": 1.8173260399440276, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6386280059814453, "learning_rate": 1e-05, "loss": 0.4975, "mean_token_accuracy": 0.8566768169403076, "num_tokens": 542234242.0, "step": 14286 }, { "epoch": 1.817453250222618, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.650110960006714, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8765041828155518, "num_tokens": 542273187.0, "step": 14287 }, { "epoch": 1.8175804605012085, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.55900239944458, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8774492740631104, "num_tokens": 542308116.0, "step": 14288 }, { "epoch": 1.817707670779799, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5622706413269043, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8768929243087769, "num_tokens": 542339964.0, "step": 14289 }, { "epoch": 1.8178348810583895, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4827606678009033, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8660117983818054, "num_tokens": 542381923.0, "step": 14290 }, { "epoch": 1.81796209133698, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.502685308456421, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8805787563323975, "num_tokens": 542418468.0, "step": 14291 }, { "epoch": 1.8180893016155706, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4512829780578613, "learning_rate": 1e-05, "loss": 0.5343, "mean_token_accuracy": 0.8515104651451111, "num_tokens": 542448806.0, "step": 14292 }, { "epoch": 1.818216511894161, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4703919887542725, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8727912902832031, "num_tokens": 542483991.0, "step": 14293 }, { "epoch": 1.8183437221727514, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3833980560302734, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8752153515815735, "num_tokens": 542528334.0, "step": 14294 }, { "epoch": 1.818470932451342, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4818639755249023, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.883864164352417, "num_tokens": 542562903.0, "step": 14295 }, { "epoch": 1.8185981427299325, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3298237323760986, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8743327856063843, "num_tokens": 542600150.0, "step": 14296 }, { "epoch": 1.818725353008523, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5256364345550537, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8788037300109863, "num_tokens": 542637894.0, "step": 14297 }, { "epoch": 1.8188525632871135, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.301281690597534, "learning_rate": 1e-05, "loss": 0.4896, "mean_token_accuracy": 0.8569966554641724, "num_tokens": 542667729.0, "step": 14298 }, { "epoch": 1.818979773565704, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5814268589019775, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.874350368976593, "num_tokens": 542699994.0, "step": 14299 }, { "epoch": 1.8191069838442946, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.206481695175171, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8777244091033936, "num_tokens": 542743849.0, "step": 14300 }, { "epoch": 1.8192341941228851, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7028441429138184, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8608914017677307, "num_tokens": 542784788.0, "step": 14301 }, { "epoch": 1.8193614044014756, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1435980796813965, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8696250319480896, "num_tokens": 542820942.0, "step": 14302 }, { "epoch": 1.8194886146800662, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.910381317138672, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8817850351333618, "num_tokens": 542857345.0, "step": 14303 }, { "epoch": 1.8196158249586567, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1078341007232666, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8804000020027161, "num_tokens": 542890952.0, "step": 14304 }, { "epoch": 1.8197430352372472, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.189855098724365, "learning_rate": 1e-05, "loss": 0.5134, "mean_token_accuracy": 0.857715368270874, "num_tokens": 542927487.0, "step": 14305 }, { "epoch": 1.8198702455158378, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.251105308532715, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.870129406452179, "num_tokens": 542967554.0, "step": 14306 }, { "epoch": 1.8199974557944283, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.397981643676758, "learning_rate": 1e-05, "loss": 0.5078, "mean_token_accuracy": 0.8582346439361572, "num_tokens": 543003768.0, "step": 14307 }, { "epoch": 1.8201246660730188, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.590200424194336, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8753262758255005, "num_tokens": 543041937.0, "step": 14308 }, { "epoch": 1.8202518763516093, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.301722526550293, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.88780677318573, "num_tokens": 543084358.0, "step": 14309 }, { "epoch": 1.8203790866301999, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7880656719207764, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.875678539276123, "num_tokens": 543121749.0, "step": 14310 }, { "epoch": 1.8205062969087902, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.09376335144043, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8593867421150208, "num_tokens": 543161160.0, "step": 14311 }, { "epoch": 1.8206335071873807, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7603602409362793, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8698197603225708, "num_tokens": 543197364.0, "step": 14312 }, { "epoch": 1.8207607174659712, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8969802856445312, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8858845233917236, "num_tokens": 543233381.0, "step": 14313 }, { "epoch": 1.8208879277445618, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6684787273406982, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8630434274673462, "num_tokens": 543266139.0, "step": 14314 }, { "epoch": 1.8210151380231523, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7530171871185303, "learning_rate": 1e-05, "loss": 0.5073, "mean_token_accuracy": 0.856765627861023, "num_tokens": 543307681.0, "step": 14315 }, { "epoch": 1.8211423483017428, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.543015480041504, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.8620404601097107, "num_tokens": 543347916.0, "step": 14316 }, { "epoch": 1.8212695585803331, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.675985336303711, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8807497024536133, "num_tokens": 543386326.0, "step": 14317 }, { "epoch": 1.8213967688589237, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.442643404006958, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8656258583068848, "num_tokens": 543423220.0, "step": 14318 }, { "epoch": 1.8215239791375142, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.660270929336548, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8767478466033936, "num_tokens": 543461672.0, "step": 14319 }, { "epoch": 1.8216511894161047, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.345609426498413, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8736553192138672, "num_tokens": 543498775.0, "step": 14320 }, { "epoch": 1.8217783996946952, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.709176778793335, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8827232718467712, "num_tokens": 543539966.0, "step": 14321 }, { "epoch": 1.8219056099732858, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2622315883636475, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8727971315383911, "num_tokens": 543577176.0, "step": 14322 }, { "epoch": 1.8220328202518763, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8294570446014404, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.8608101010322571, "num_tokens": 543621328.0, "step": 14323 }, { "epoch": 1.8221600305304668, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.220860719680786, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8592497706413269, "num_tokens": 543662446.0, "step": 14324 }, { "epoch": 1.8222872408090574, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 4.011619567871094, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8804757595062256, "num_tokens": 543695739.0, "step": 14325 }, { "epoch": 1.8224144510876479, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2724931240081787, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8769996166229248, "num_tokens": 543735823.0, "step": 14326 }, { "epoch": 1.8225416613662384, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.141687870025635, "learning_rate": 1e-05, "loss": 0.5156, "mean_token_accuracy": 0.8535252809524536, "num_tokens": 543771800.0, "step": 14327 }, { "epoch": 1.822668871644829, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.457482099533081, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8625169992446899, "num_tokens": 543812425.0, "step": 14328 }, { "epoch": 1.8227960819234195, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.093194007873535, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8749894499778748, "num_tokens": 543849867.0, "step": 14329 }, { "epoch": 1.82292329220201, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6224350929260254, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8826003074645996, "num_tokens": 543887289.0, "step": 14330 }, { "epoch": 1.8230505024806005, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8886024951934814, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8772865533828735, "num_tokens": 543923432.0, "step": 14331 }, { "epoch": 1.823177712759191, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.63950777053833, "learning_rate": 1e-05, "loss": 0.4966, "mean_token_accuracy": 0.85727858543396, "num_tokens": 543966655.0, "step": 14332 }, { "epoch": 1.8233049230377816, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6691315174102783, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.877622127532959, "num_tokens": 544006920.0, "step": 14333 }, { "epoch": 1.823432133316372, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.619529962539673, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8718217611312866, "num_tokens": 544044066.0, "step": 14334 }, { "epoch": 1.8235593435949626, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.50894832611084, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8652506470680237, "num_tokens": 544073795.0, "step": 14335 }, { "epoch": 1.823686553873553, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.604746103286743, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8774844408035278, "num_tokens": 544107378.0, "step": 14336 }, { "epoch": 1.8238137641521435, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3857548236846924, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8741728663444519, "num_tokens": 544144377.0, "step": 14337 }, { "epoch": 1.823940974430734, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6180636882781982, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8678208589553833, "num_tokens": 544176484.0, "step": 14338 }, { "epoch": 1.8240681847093245, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2983381748199463, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8775113821029663, "num_tokens": 544209152.0, "step": 14339 }, { "epoch": 1.824195394987915, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.666839838027954, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.866328239440918, "num_tokens": 544249197.0, "step": 14340 }, { "epoch": 1.8243226052665056, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2068660259246826, "learning_rate": 1e-05, "loss": 0.4042, "mean_token_accuracy": 0.8860564231872559, "num_tokens": 544290225.0, "step": 14341 }, { "epoch": 1.8244498155450959, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7748427391052246, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8676259517669678, "num_tokens": 544333015.0, "step": 14342 }, { "epoch": 1.8245770258236864, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.171830654144287, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8655768632888794, "num_tokens": 544370165.0, "step": 14343 }, { "epoch": 1.824704236102277, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9427223205566406, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8729060292243958, "num_tokens": 544407284.0, "step": 14344 }, { "epoch": 1.8248314463808675, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1990966796875, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8596621155738831, "num_tokens": 544441558.0, "step": 14345 }, { "epoch": 1.824958656659458, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.112616539001465, "learning_rate": 1e-05, "loss": 0.5362, "mean_token_accuracy": 0.8463451862335205, "num_tokens": 544478468.0, "step": 14346 }, { "epoch": 1.8250858669380485, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.305419445037842, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8878150582313538, "num_tokens": 544516271.0, "step": 14347 }, { "epoch": 1.825213077216639, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2119526863098145, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8813101053237915, "num_tokens": 544553198.0, "step": 14348 }, { "epoch": 1.8253402874952296, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5263609886169434, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.877719521522522, "num_tokens": 544586381.0, "step": 14349 }, { "epoch": 1.8254674977738201, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.122115135192871, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8694757223129272, "num_tokens": 544628342.0, "step": 14350 }, { "epoch": 1.8255947080524106, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.636794328689575, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8663873672485352, "num_tokens": 544667096.0, "step": 14351 }, { "epoch": 1.8257219183310012, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.934298038482666, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8765017986297607, "num_tokens": 544711467.0, "step": 14352 }, { "epoch": 1.8258491286095917, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6433193683624268, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8726640939712524, "num_tokens": 544751672.0, "step": 14353 }, { "epoch": 1.8259763388881822, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7546749114990234, "learning_rate": 1e-05, "loss": 0.3892, "mean_token_accuracy": 0.8938525915145874, "num_tokens": 544785479.0, "step": 14354 }, { "epoch": 1.8261035491667728, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.5877397060394287, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8740819692611694, "num_tokens": 544823122.0, "step": 14355 }, { "epoch": 1.8262307594453633, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.616779327392578, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8803970813751221, "num_tokens": 544862544.0, "step": 14356 }, { "epoch": 1.8263579697239538, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5373799800872803, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8645947575569153, "num_tokens": 544899538.0, "step": 14357 }, { "epoch": 1.8264851800025443, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5260937213897705, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8796830773353577, "num_tokens": 544934544.0, "step": 14358 }, { "epoch": 1.8266123902811349, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4740681648254395, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.863272488117218, "num_tokens": 544973304.0, "step": 14359 }, { "epoch": 1.8267396005597252, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.479893445968628, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.879003643989563, "num_tokens": 545010748.0, "step": 14360 }, { "epoch": 1.8268668108383157, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4197466373443604, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.865653932094574, "num_tokens": 545050742.0, "step": 14361 }, { "epoch": 1.8269940211169062, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4843811988830566, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8684105277061462, "num_tokens": 545083904.0, "step": 14362 }, { "epoch": 1.8271212313954968, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.36266827583313, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8809189796447754, "num_tokens": 545123885.0, "step": 14363 }, { "epoch": 1.8272484416740873, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.502901077270508, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8655244708061218, "num_tokens": 545160295.0, "step": 14364 }, { "epoch": 1.8273756519526778, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.304845094680786, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8595640063285828, "num_tokens": 545203469.0, "step": 14365 }, { "epoch": 1.8275028622312681, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5503647327423096, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8773323893547058, "num_tokens": 545242173.0, "step": 14366 }, { "epoch": 1.8276300725098586, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2371976375579834, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8651636242866516, "num_tokens": 545284968.0, "step": 14367 }, { "epoch": 1.8277572827884492, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.669079542160034, "learning_rate": 1e-05, "loss": 0.5071, "mean_token_accuracy": 0.8559830784797668, "num_tokens": 545324764.0, "step": 14368 }, { "epoch": 1.8278844930670397, "ewc_loss": 0.0712890625, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.4345855712890625e-05, "grad_norm": 3.1590797901153564, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8662277460098267, "num_tokens": 545362783.0, "step": 14369 }, { "epoch": 1.8280117033456302, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8481597900390625, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8806377649307251, "num_tokens": 545395757.0, "step": 14370 }, { "epoch": 1.8281389136242208, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1467881202697754, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8858823180198669, "num_tokens": 545432194.0, "step": 14371 }, { "epoch": 1.8282661239028113, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.051685810089111, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8759077787399292, "num_tokens": 545469743.0, "step": 14372 }, { "epoch": 1.8283933341814018, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1885101795196533, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8657392263412476, "num_tokens": 545508976.0, "step": 14373 }, { "epoch": 1.8285205444599923, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002689361572265625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.272583484649658, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8725329637527466, "num_tokens": 545548223.0, "step": 14374 }, { "epoch": 1.8286477547385829, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.445251226425171, "learning_rate": 1e-05, "loss": 0.5065, "mean_token_accuracy": 0.8556677103042603, "num_tokens": 545586110.0, "step": 14375 }, { "epoch": 1.8287749650171734, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.289501190185547, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8684782981872559, "num_tokens": 545625779.0, "step": 14376 }, { "epoch": 1.828902175295764, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.668689012527466, "learning_rate": 1e-05, "loss": 0.5222, "mean_token_accuracy": 0.8512915372848511, "num_tokens": 545669464.0, "step": 14377 }, { "epoch": 1.8290293855743545, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.120030879974365, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8788328170776367, "num_tokens": 545702674.0, "step": 14378 }, { "epoch": 1.829156595852945, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7138588428497314, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8764169216156006, "num_tokens": 545742410.0, "step": 14379 }, { "epoch": 1.8292838061315355, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.917964458465576, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.864800214767456, "num_tokens": 545779748.0, "step": 14380 }, { "epoch": 1.829411016410126, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6724555492401123, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8642424941062927, "num_tokens": 545811401.0, "step": 14381 }, { "epoch": 1.8295382266887166, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.736292600631714, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.879126787185669, "num_tokens": 545848565.0, "step": 14382 }, { "epoch": 1.829665436967307, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5705161094665527, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8676811456680298, "num_tokens": 545891245.0, "step": 14383 }, { "epoch": 1.8297926472458976, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6282310485839844, "learning_rate": 1e-05, "loss": 0.3905, "mean_token_accuracy": 0.8896620869636536, "num_tokens": 545924860.0, "step": 14384 }, { "epoch": 1.829919857524488, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.496175765991211, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8819382190704346, "num_tokens": 545958119.0, "step": 14385 }, { "epoch": 1.8300470678030785, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.583540678024292, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8650837540626526, "num_tokens": 545994195.0, "step": 14386 }, { "epoch": 1.830174278081669, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4140214920043945, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8717402219772339, "num_tokens": 546031998.0, "step": 14387 }, { "epoch": 1.8303014883602595, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.57794189453125, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.871955156326294, "num_tokens": 546067152.0, "step": 14388 }, { "epoch": 1.83042869863885, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3225269317626953, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.884962260723114, "num_tokens": 546104207.0, "step": 14389 }, { "epoch": 1.8305559089174406, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.624107837677002, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8586830496788025, "num_tokens": 546150659.0, "step": 14390 }, { "epoch": 1.8306831191960309, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2511825561523438, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8775022625923157, "num_tokens": 546188427.0, "step": 14391 }, { "epoch": 1.8308103294746214, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.73862361907959, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.864271879196167, "num_tokens": 546227249.0, "step": 14392 }, { "epoch": 1.830937539753212, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1952106952667236, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8623083829879761, "num_tokens": 546263469.0, "step": 14393 }, { "epoch": 1.8310647500318025, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.9020426273345947, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8693196773529053, "num_tokens": 546300820.0, "step": 14394 }, { "epoch": 1.831191960310393, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.182863712310791, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8772412538528442, "num_tokens": 546336639.0, "step": 14395 }, { "epoch": 1.8313191705889835, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.088945388793945, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.865107536315918, "num_tokens": 546377769.0, "step": 14396 }, { "epoch": 1.831446380867574, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.29246187210083, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8793591856956482, "num_tokens": 546419286.0, "step": 14397 }, { "epoch": 1.8315735911461646, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.215250015258789, "learning_rate": 1e-05, "loss": 0.5058, "mean_token_accuracy": 0.859065055847168, "num_tokens": 546458907.0, "step": 14398 }, { "epoch": 1.831700801424755, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.537421941757202, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.8898471593856812, "num_tokens": 546491682.0, "step": 14399 }, { "epoch": 1.8318280117033456, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.112166881561279, "learning_rate": 1e-05, "loss": 0.5295, "mean_token_accuracy": 0.8508461713790894, "num_tokens": 546537141.0, "step": 14400 }, { "epoch": 1.8319552219819362, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.656602621078491, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8678845167160034, "num_tokens": 546574318.0, "step": 14401 }, { "epoch": 1.8320824322605267, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9036455154418945, "learning_rate": 1e-05, "loss": 0.4878, "mean_token_accuracy": 0.861953854560852, "num_tokens": 546615938.0, "step": 14402 }, { "epoch": 1.8322096425391172, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.655797004699707, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8737565279006958, "num_tokens": 546652608.0, "step": 14403 }, { "epoch": 1.8323368528177078, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6903183460235596, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8837469816207886, "num_tokens": 546695363.0, "step": 14404 }, { "epoch": 1.8324640630962983, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6126816272735596, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8564606308937073, "num_tokens": 546741174.0, "step": 14405 }, { "epoch": 1.8325912733748888, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5304365158081055, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8788317441940308, "num_tokens": 546777827.0, "step": 14406 }, { "epoch": 1.8327184836534793, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5676867961883545, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8781880736351013, "num_tokens": 546821199.0, "step": 14407 }, { "epoch": 1.8328456939320699, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.437584400177002, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8713539242744446, "num_tokens": 546860151.0, "step": 14408 }, { "epoch": 1.8329729042106602, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5584254264831543, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.879430890083313, "num_tokens": 546892456.0, "step": 14409 }, { "epoch": 1.8331001144892507, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3698041439056396, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8635521531105042, "num_tokens": 546931611.0, "step": 14410 }, { "epoch": 1.8332273247678412, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.556875467300415, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8765536546707153, "num_tokens": 546968408.0, "step": 14411 }, { "epoch": 1.8333545350464318, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.305764675140381, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8589211106300354, "num_tokens": 547007025.0, "step": 14412 }, { "epoch": 1.8334817453250223, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.591888666152954, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8585926294326782, "num_tokens": 547051371.0, "step": 14413 }, { "epoch": 1.8336089556036128, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.264589309692383, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8641818761825562, "num_tokens": 547087476.0, "step": 14414 }, { "epoch": 1.8337361658822031, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6505017280578613, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8751118779182434, "num_tokens": 547126922.0, "step": 14415 }, { "epoch": 1.8338633761607936, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.196444272994995, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8763139247894287, "num_tokens": 547164678.0, "step": 14416 }, { "epoch": 1.8339905864393842, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7876839637756348, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8697061538696289, "num_tokens": 547200039.0, "step": 14417 }, { "epoch": 1.8341177967179747, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1382198333740234, "learning_rate": 1e-05, "loss": 0.3833, "mean_token_accuracy": 0.8914763927459717, "num_tokens": 547233325.0, "step": 14418 }, { "epoch": 1.8342450069965652, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.994016170501709, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8845571279525757, "num_tokens": 547266658.0, "step": 14419 }, { "epoch": 1.8343722172751558, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.1608402729034424, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8642421364784241, "num_tokens": 547305876.0, "step": 14420 }, { "epoch": 1.8344994275537463, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.220363616943359, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8773857355117798, "num_tokens": 547346847.0, "step": 14421 }, { "epoch": 1.8346266378323368, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.3597946166992188, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8831906318664551, "num_tokens": 547379896.0, "step": 14422 }, { "epoch": 1.8347538481109273, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.3023200035095215, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8657283782958984, "num_tokens": 547425666.0, "step": 14423 }, { "epoch": 1.8348810583895179, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.609623670578003, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8777159452438354, "num_tokens": 547467007.0, "step": 14424 }, { "epoch": 1.8350082686681084, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.182885646820068, "learning_rate": 1e-05, "loss": 0.5275, "mean_token_accuracy": 0.8517919778823853, "num_tokens": 547507718.0, "step": 14425 }, { "epoch": 1.835135478946699, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7251834869384766, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8766834139823914, "num_tokens": 547543445.0, "step": 14426 }, { "epoch": 1.8352626892252895, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9753453731536865, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8741428256034851, "num_tokens": 547576000.0, "step": 14427 }, { "epoch": 1.83538989950388, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.692905902862549, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.876198947429657, "num_tokens": 547609950.0, "step": 14428 }, { "epoch": 1.8355171097824705, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7782320976257324, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8773303031921387, "num_tokens": 547650133.0, "step": 14429 }, { "epoch": 1.835644320061061, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6176023483276367, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8674161434173584, "num_tokens": 547687754.0, "step": 14430 }, { "epoch": 1.8357715303396516, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.634596824645996, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8747915029525757, "num_tokens": 547732119.0, "step": 14431 }, { "epoch": 1.835898740618242, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.536893367767334, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8780694007873535, "num_tokens": 547770779.0, "step": 14432 }, { "epoch": 1.8360259508968326, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5557754039764404, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.879083514213562, "num_tokens": 547812524.0, "step": 14433 }, { "epoch": 1.836153161175423, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.474090576171875, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8651935458183289, "num_tokens": 547850955.0, "step": 14434 }, { "epoch": 1.8362803714540135, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.516124963760376, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8744592070579529, "num_tokens": 547891390.0, "step": 14435 }, { "epoch": 1.836407581732604, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.416860818862915, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8664882183074951, "num_tokens": 547930306.0, "step": 14436 }, { "epoch": 1.8365347920111945, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5088298320770264, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8633309602737427, "num_tokens": 547970940.0, "step": 14437 }, { "epoch": 1.836662002289785, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.347541093826294, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8791013360023499, "num_tokens": 548007908.0, "step": 14438 }, { "epoch": 1.8367892125683756, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.552382469177246, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8670616149902344, "num_tokens": 548045040.0, "step": 14439 }, { "epoch": 1.8369164228469659, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2782766819000244, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8722702264785767, "num_tokens": 548088501.0, "step": 14440 }, { "epoch": 1.8370436331255564, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6384692192077637, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8590120673179626, "num_tokens": 548121888.0, "step": 14441 }, { "epoch": 1.837170843404147, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1965785026550293, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.873517632484436, "num_tokens": 548161117.0, "step": 14442 }, { "epoch": 1.8372980536827375, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.784533977508545, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8820990920066833, "num_tokens": 548201709.0, "step": 14443 }, { "epoch": 1.837425263961328, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1304733753204346, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8708505630493164, "num_tokens": 548241843.0, "step": 14444 }, { "epoch": 1.8375524742399185, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.025304317474365, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8774817585945129, "num_tokens": 548278785.0, "step": 14445 }, { "epoch": 1.837679684518509, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.175830841064453, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8771528005599976, "num_tokens": 548315766.0, "step": 14446 }, { "epoch": 1.8378068947970996, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.267416477203369, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8679105043411255, "num_tokens": 548353971.0, "step": 14447 }, { "epoch": 1.83793410507569, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4153454303741455, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8722345232963562, "num_tokens": 548391261.0, "step": 14448 }, { "epoch": 1.8380613153542806, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.30903434753418, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8835973739624023, "num_tokens": 548424773.0, "step": 14449 }, { "epoch": 1.8381885256328712, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6640424728393555, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8872220516204834, "num_tokens": 548462404.0, "step": 14450 }, { "epoch": 1.8383157359114617, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.141494274139404, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8729902505874634, "num_tokens": 548504595.0, "step": 14451 }, { "epoch": 1.8384429461900522, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7235591411590576, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.87935471534729, "num_tokens": 548547945.0, "step": 14452 }, { "epoch": 1.8385701564686427, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9404823780059814, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8613336086273193, "num_tokens": 548590302.0, "step": 14453 }, { "epoch": 1.8386973667472333, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6715118885040283, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8828706741333008, "num_tokens": 548629688.0, "step": 14454 }, { "epoch": 1.8388245770258238, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7747185230255127, "learning_rate": 1e-05, "loss": 0.5168, "mean_token_accuracy": 0.8521978855133057, "num_tokens": 548666294.0, "step": 14455 }, { "epoch": 1.8389517873044143, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5800938606262207, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8596291542053223, "num_tokens": 548707536.0, "step": 14456 }, { "epoch": 1.8390789975830049, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.659778118133545, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8779138922691345, "num_tokens": 548751443.0, "step": 14457 }, { "epoch": 1.8392062078615952, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4884121417999268, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8697018623352051, "num_tokens": 548789259.0, "step": 14458 }, { "epoch": 1.8393334181401857, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.621765613555908, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8757918477058411, "num_tokens": 548822045.0, "step": 14459 }, { "epoch": 1.8394606284187762, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.410928249359131, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8599095940589905, "num_tokens": 548854722.0, "step": 14460 }, { "epoch": 1.8395878386973668, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6264657974243164, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8666030764579773, "num_tokens": 548888903.0, "step": 14461 }, { "epoch": 1.8397150489759573, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3184077739715576, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8766573667526245, "num_tokens": 548925611.0, "step": 14462 }, { "epoch": 1.8398422592545478, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6825602054595947, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8786705732345581, "num_tokens": 548967239.0, "step": 14463 }, { "epoch": 1.8399694695331381, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2406318187713623, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.876630961894989, "num_tokens": 549005453.0, "step": 14464 }, { "epoch": 1.8400966798117286, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8129262924194336, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8687922954559326, "num_tokens": 549041418.0, "step": 14465 }, { "epoch": 1.8402238900903192, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.203171730041504, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8799973130226135, "num_tokens": 549079223.0, "step": 14466 }, { "epoch": 1.8403511003689097, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9812874794006348, "learning_rate": 1e-05, "loss": 0.5368, "mean_token_accuracy": 0.8492185473442078, "num_tokens": 549124542.0, "step": 14467 }, { "epoch": 1.8404783106475002, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2763805389404297, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8613101243972778, "num_tokens": 549165374.0, "step": 14468 }, { "epoch": 1.8406055209260908, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.103765964508057, "learning_rate": 1e-05, "loss": 0.5108, "mean_token_accuracy": 0.8583177328109741, "num_tokens": 549203887.0, "step": 14469 }, { "epoch": 1.8407327312046813, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.392578363418579, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8702276945114136, "num_tokens": 549247583.0, "step": 14470 }, { "epoch": 1.8408599414832718, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.112685680389404, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8881796598434448, "num_tokens": 549282874.0, "step": 14471 }, { "epoch": 1.8409871517618623, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.5548696517944336, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8797655701637268, "num_tokens": 549323133.0, "step": 14472 }, { "epoch": 1.8411143620404529, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 8.305316925048828, "learning_rate": 1e-05, "loss": 0.5532, "mean_token_accuracy": 0.8559286594390869, "num_tokens": 549359794.0, "step": 14473 }, { "epoch": 1.8412415723190434, "ewc_loss": 0.150390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00012302398681640625, "grad_norm": 5.664900302886963, "learning_rate": 1e-05, "loss": 0.5867, "mean_token_accuracy": 0.8535533547401428, "num_tokens": 549393281.0, "step": 14474 }, { "epoch": 1.841368782597634, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.0021729469299316, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8665534257888794, "num_tokens": 549428181.0, "step": 14475 }, { "epoch": 1.8414959928762245, "ewc_loss": 0.126953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00010013580322265625, "grad_norm": 5.252411365509033, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8780663013458252, "num_tokens": 549465521.0, "step": 14476 }, { "epoch": 1.841623203154815, "ewc_loss": 0.1357421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 0.00010824203491210938, "grad_norm": 4.881802082061768, "learning_rate": 1e-05, "loss": 0.5153, "mean_token_accuracy": 0.8713650107383728, "num_tokens": 549503800.0, "step": 14477 }, { "epoch": 1.8417504134334055, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.608873128890991, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8824291229248047, "num_tokens": 549536725.0, "step": 14478 }, { "epoch": 1.841877623711996, "ewc_loss": 0.1083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 8.106231689453125e-05, "grad_norm": 4.517364025115967, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8729931116104126, "num_tokens": 549575774.0, "step": 14479 }, { "epoch": 1.8420048339905866, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 3.738236665725708, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8735268712043762, "num_tokens": 549616715.0, "step": 14480 }, { "epoch": 1.842132044269177, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.174984455108643, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8713419437408447, "num_tokens": 549652005.0, "step": 14481 }, { "epoch": 1.8422592545477676, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 3.9350550174713135, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8692245483398438, "num_tokens": 549683708.0, "step": 14482 }, { "epoch": 1.842386464826358, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.726536750793457, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8858808279037476, "num_tokens": 549722961.0, "step": 14483 }, { "epoch": 1.8425136751049485, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.890632390975952, "learning_rate": 1e-05, "loss": 0.49, "mean_token_accuracy": 0.8645033836364746, "num_tokens": 549758857.0, "step": 14484 }, { "epoch": 1.842640885383539, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.498811960220337, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8720062971115112, "num_tokens": 549793835.0, "step": 14485 }, { "epoch": 1.8427680956621295, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.837538480758667, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8736944794654846, "num_tokens": 549832691.0, "step": 14486 }, { "epoch": 1.84289530594072, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.54679799079895, "learning_rate": 1e-05, "loss": 0.5958, "mean_token_accuracy": 0.8346534371376038, "num_tokens": 549874981.0, "step": 14487 }, { "epoch": 1.8430225162193106, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.666835069656372, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8711299300193787, "num_tokens": 549917904.0, "step": 14488 }, { "epoch": 1.8431497264979009, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.469627857208252, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8694238066673279, "num_tokens": 549955421.0, "step": 14489 }, { "epoch": 1.8432769367764914, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.544646978378296, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8757543563842773, "num_tokens": 549986615.0, "step": 14490 }, { "epoch": 1.843404147055082, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.465651273727417, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8641494512557983, "num_tokens": 550022747.0, "step": 14491 }, { "epoch": 1.8435313573336725, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.4558002948760986, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8706099987030029, "num_tokens": 550060423.0, "step": 14492 }, { "epoch": 1.843658567612263, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.449901819229126, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8704442977905273, "num_tokens": 550102086.0, "step": 14493 }, { "epoch": 1.8437857778908535, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3983983993530273, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8867560029029846, "num_tokens": 550138281.0, "step": 14494 }, { "epoch": 1.843912988169444, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.462465286254883, "learning_rate": 1e-05, "loss": 0.5029, "mean_token_accuracy": 0.8567869663238525, "num_tokens": 550180219.0, "step": 14495 }, { "epoch": 1.8440401984480346, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3637077808380127, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8723974227905273, "num_tokens": 550218095.0, "step": 14496 }, { "epoch": 1.844167408726625, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.467193126678467, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8660014271736145, "num_tokens": 550256080.0, "step": 14497 }, { "epoch": 1.8442946190052156, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3300929069519043, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8565106391906738, "num_tokens": 550297471.0, "step": 14498 }, { "epoch": 1.8444218292838062, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.483844518661499, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.8889371156692505, "num_tokens": 550330495.0, "step": 14499 }, { "epoch": 1.8445490395623967, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.32057523727417, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.870178759098053, "num_tokens": 550362810.0, "step": 14500 }, { "epoch": 1.8446762498409872, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4764325618743896, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.869724690914154, "num_tokens": 550409408.0, "step": 14501 }, { "epoch": 1.8448034601195777, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.30250883102417, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8649188280105591, "num_tokens": 550441366.0, "step": 14502 }, { "epoch": 1.8449306703981683, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5137155055999756, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8772907257080078, "num_tokens": 550478247.0, "step": 14503 }, { "epoch": 1.8450578806767588, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.246523141860962, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.877560019493103, "num_tokens": 550511861.0, "step": 14504 }, { "epoch": 1.8451850909553493, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.578890085220337, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8655105829238892, "num_tokens": 550554959.0, "step": 14505 }, { "epoch": 1.8453123012339399, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1892852783203125, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.87525874376297, "num_tokens": 550589960.0, "step": 14506 }, { "epoch": 1.8454395115125302, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.693145513534546, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8710724711418152, "num_tokens": 550630481.0, "step": 14507 }, { "epoch": 1.8455667217911207, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1144673824310303, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8614956736564636, "num_tokens": 550671819.0, "step": 14508 }, { "epoch": 1.8456939320697112, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8932902812957764, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8794704079627991, "num_tokens": 550712230.0, "step": 14509 }, { "epoch": 1.8458211423483017, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.0676169395446777, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8694015741348267, "num_tokens": 550751052.0, "step": 14510 }, { "epoch": 1.8459483526268923, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.196887493133545, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8775121569633484, "num_tokens": 550787374.0, "step": 14511 }, { "epoch": 1.8460755629054828, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.1753406524658203, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8784105777740479, "num_tokens": 550829398.0, "step": 14512 }, { "epoch": 1.846202773184073, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.494452476501465, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8619015216827393, "num_tokens": 550868818.0, "step": 14513 }, { "epoch": 1.8463299834626636, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.584054946899414, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8762365579605103, "num_tokens": 550904136.0, "step": 14514 }, { "epoch": 1.8464571937412542, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 4.479323387145996, "learning_rate": 1e-05, "loss": 0.5842, "mean_token_accuracy": 0.8376760482788086, "num_tokens": 550943505.0, "step": 14515 }, { "epoch": 1.8465844040198447, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 3.8796145915985107, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.870390772819519, "num_tokens": 550983633.0, "step": 14516 }, { "epoch": 1.8467116142984352, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.223461627960205, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8811345100402832, "num_tokens": 551017683.0, "step": 14517 }, { "epoch": 1.8468388245770258, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.8691248893737793, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8693645000457764, "num_tokens": 551051731.0, "step": 14518 }, { "epoch": 1.8469660348556163, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0014448165893555, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8767799139022827, "num_tokens": 551090818.0, "step": 14519 }, { "epoch": 1.8470932451342068, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.7369585037231445, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8778970837593079, "num_tokens": 551122032.0, "step": 14520 }, { "epoch": 1.8472204554127973, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8594746589660645, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8635915517807007, "num_tokens": 551157708.0, "step": 14521 }, { "epoch": 1.8473476656913879, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5978283882141113, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8711512684822083, "num_tokens": 551194222.0, "step": 14522 }, { "epoch": 1.8474748759699784, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7893598079681396, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8702775835990906, "num_tokens": 551234387.0, "step": 14523 }, { "epoch": 1.847602086248569, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.469787836074829, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.875639796257019, "num_tokens": 551273330.0, "step": 14524 }, { "epoch": 1.8477292965271594, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7890963554382324, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8725528717041016, "num_tokens": 551309180.0, "step": 14525 }, { "epoch": 1.84785650680575, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.364222764968872, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8832955360412598, "num_tokens": 551344455.0, "step": 14526 }, { "epoch": 1.8479837170843405, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.862187623977661, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8621250987052917, "num_tokens": 551380795.0, "step": 14527 }, { "epoch": 1.848110927362931, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3009045124053955, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8590530753135681, "num_tokens": 551421314.0, "step": 14528 }, { "epoch": 1.8482381376415216, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9962384700775146, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8742489218711853, "num_tokens": 551459021.0, "step": 14529 }, { "epoch": 1.848365347920112, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3136966228485107, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8717861175537109, "num_tokens": 551498055.0, "step": 14530 }, { "epoch": 1.8484925581987026, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.135695934295654, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.865418553352356, "num_tokens": 551536746.0, "step": 14531 }, { "epoch": 1.848619768477293, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.4636037349700928, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8605322241783142, "num_tokens": 551576230.0, "step": 14532 }, { "epoch": 1.8487469787558835, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.126865386962891, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8709088563919067, "num_tokens": 551613476.0, "step": 14533 }, { "epoch": 1.848874189034474, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6176657676696777, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8732348084449768, "num_tokens": 551655177.0, "step": 14534 }, { "epoch": 1.8490013993130645, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.949723243713379, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8674217462539673, "num_tokens": 551691236.0, "step": 14535 }, { "epoch": 1.849128609591655, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.673262119293213, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8613578081130981, "num_tokens": 551732152.0, "step": 14536 }, { "epoch": 1.8492558198702456, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7227530479431152, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8765947222709656, "num_tokens": 551765221.0, "step": 14537 }, { "epoch": 1.8493830301488359, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6549952030181885, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8721789121627808, "num_tokens": 551805321.0, "step": 14538 }, { "epoch": 1.8495102404274264, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.506446599960327, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8638955354690552, "num_tokens": 551846752.0, "step": 14539 }, { "epoch": 1.849637450706017, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6381115913391113, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8682544827461243, "num_tokens": 551887877.0, "step": 14540 }, { "epoch": 1.8497646609846075, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3822836875915527, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8647441864013672, "num_tokens": 551919669.0, "step": 14541 }, { "epoch": 1.849891871263198, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.646106243133545, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8723486065864563, "num_tokens": 551955459.0, "step": 14542 }, { "epoch": 1.8500190815417885, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.275947093963623, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.868730902671814, "num_tokens": 551994889.0, "step": 14543 }, { "epoch": 1.850146291820379, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7129290103912354, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.873124361038208, "num_tokens": 552030176.0, "step": 14544 }, { "epoch": 1.8502735020989696, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.204890012741089, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8841094970703125, "num_tokens": 552063317.0, "step": 14545 }, { "epoch": 1.85040071237756, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8389270305633545, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8886488676071167, "num_tokens": 552100959.0, "step": 14546 }, { "epoch": 1.8505279226561506, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.173306941986084, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8767349720001221, "num_tokens": 552139181.0, "step": 14547 }, { "epoch": 1.8506551329347412, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 4.019184589385986, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8660517930984497, "num_tokens": 552172293.0, "step": 14548 }, { "epoch": 1.8507823432133317, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.217905282974243, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.868554949760437, "num_tokens": 552214655.0, "step": 14549 }, { "epoch": 1.8509095534919222, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.197287082672119, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.889983594417572, "num_tokens": 552254188.0, "step": 14550 }, { "epoch": 1.8510367637705127, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.4336259365081787, "learning_rate": 1e-05, "loss": 0.5171, "mean_token_accuracy": 0.8523023128509521, "num_tokens": 552286712.0, "step": 14551 }, { "epoch": 1.8511639740491033, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.213714599609375, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8703039288520813, "num_tokens": 552326846.0, "step": 14552 }, { "epoch": 1.8512911843276938, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6183204650878906, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8821172714233398, "num_tokens": 552356753.0, "step": 14553 }, { "epoch": 1.8514183946062843, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.068473815917969, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8609814047813416, "num_tokens": 552391312.0, "step": 14554 }, { "epoch": 1.8515456048848749, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.681159257888794, "learning_rate": 1e-05, "loss": 0.5327, "mean_token_accuracy": 0.847825288772583, "num_tokens": 552428782.0, "step": 14555 }, { "epoch": 1.8516728151634652, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.853971481323242, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8706121444702148, "num_tokens": 552471131.0, "step": 14556 }, { "epoch": 1.8518000254420557, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.625915765762329, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8753220438957214, "num_tokens": 552512242.0, "step": 14557 }, { "epoch": 1.8519272357206462, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.716913938522339, "learning_rate": 1e-05, "loss": 0.5247, "mean_token_accuracy": 0.8506462574005127, "num_tokens": 552545655.0, "step": 14558 }, { "epoch": 1.8520544459992367, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5552845001220703, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8705425262451172, "num_tokens": 552586340.0, "step": 14559 }, { "epoch": 1.8521816562778273, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5974912643432617, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8709401488304138, "num_tokens": 552624461.0, "step": 14560 }, { "epoch": 1.8523088665564178, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4908547401428223, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8824858069419861, "num_tokens": 552661566.0, "step": 14561 }, { "epoch": 1.852436076835008, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5487022399902344, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8827305436134338, "num_tokens": 552698184.0, "step": 14562 }, { "epoch": 1.8525632871135986, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4062705039978027, "learning_rate": 1e-05, "loss": 0.3956, "mean_token_accuracy": 0.8894164562225342, "num_tokens": 552741407.0, "step": 14563 }, { "epoch": 1.8526904973921892, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5562384128570557, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.863763689994812, "num_tokens": 552780038.0, "step": 14564 }, { "epoch": 1.8528177076707797, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.346919536590576, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.877865195274353, "num_tokens": 552818479.0, "step": 14565 }, { "epoch": 1.8529449179493702, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.599929094314575, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8768646717071533, "num_tokens": 552853762.0, "step": 14566 }, { "epoch": 1.8530721282279607, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2820217609405518, "learning_rate": 1e-05, "loss": 0.5156, "mean_token_accuracy": 0.8455076217651367, "num_tokens": 552892693.0, "step": 14567 }, { "epoch": 1.8531993385065513, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6743810176849365, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8780249953269958, "num_tokens": 552930127.0, "step": 14568 }, { "epoch": 1.8533265487851418, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.19789457321167, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8721814155578613, "num_tokens": 552969809.0, "step": 14569 }, { "epoch": 1.8534537590637323, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8316256999969482, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8778504729270935, "num_tokens": 553004747.0, "step": 14570 }, { "epoch": 1.8535809693423229, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1637606620788574, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.859104573726654, "num_tokens": 553049727.0, "step": 14571 }, { "epoch": 1.8537081796209134, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.046622276306152, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8710208535194397, "num_tokens": 553093263.0, "step": 14572 }, { "epoch": 1.853835389899504, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.250605583190918, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8717024326324463, "num_tokens": 553126198.0, "step": 14573 }, { "epoch": 1.8539626001780944, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.201611042022705, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8761717677116394, "num_tokens": 553156798.0, "step": 14574 }, { "epoch": 1.854089810456685, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.45066237449646, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8745724558830261, "num_tokens": 553193612.0, "step": 14575 }, { "epoch": 1.8542170207352755, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.172881126403809, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.876970112323761, "num_tokens": 553227376.0, "step": 14576 }, { "epoch": 1.854344231013866, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.627467155456543, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.868401050567627, "num_tokens": 553263045.0, "step": 14577 }, { "epoch": 1.8544714412924566, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.989243745803833, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8644717931747437, "num_tokens": 553301834.0, "step": 14578 }, { "epoch": 1.854598651571047, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.663959503173828, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8601457476615906, "num_tokens": 553345980.0, "step": 14579 }, { "epoch": 1.8547258618496376, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.785137176513672, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8680925965309143, "num_tokens": 553380645.0, "step": 14580 }, { "epoch": 1.854853072128228, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6277291774749756, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.863172709941864, "num_tokens": 553417952.0, "step": 14581 }, { "epoch": 1.8549802824068184, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5990912914276123, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8766888380050659, "num_tokens": 553454575.0, "step": 14582 }, { "epoch": 1.855107492685409, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.583294630050659, "learning_rate": 1e-05, "loss": 0.3854, "mean_token_accuracy": 0.8914741277694702, "num_tokens": 553489846.0, "step": 14583 }, { "epoch": 1.8552347029639995, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.482086420059204, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8777792453765869, "num_tokens": 553524444.0, "step": 14584 }, { "epoch": 1.85536191324259, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.557326555252075, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8766909241676331, "num_tokens": 553562295.0, "step": 14585 }, { "epoch": 1.8554891235211806, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4277570247650146, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.8560718297958374, "num_tokens": 553600240.0, "step": 14586 }, { "epoch": 1.8556163337997709, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.516941785812378, "learning_rate": 1e-05, "loss": 0.3978, "mean_token_accuracy": 0.8872700929641724, "num_tokens": 553641215.0, "step": 14587 }, { "epoch": 1.8557435440783614, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.374770402908325, "learning_rate": 1e-05, "loss": 0.4038, "mean_token_accuracy": 0.885371208190918, "num_tokens": 553675599.0, "step": 14588 }, { "epoch": 1.855870754356952, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5121326446533203, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8823906183242798, "num_tokens": 553706681.0, "step": 14589 }, { "epoch": 1.8559979646355425, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.340519905090332, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8736128807067871, "num_tokens": 553740919.0, "step": 14590 }, { "epoch": 1.856125174914133, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5318193435668945, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8755295872688293, "num_tokens": 553775509.0, "step": 14591 }, { "epoch": 1.8562523851927235, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.294795036315918, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8630092740058899, "num_tokens": 553809168.0, "step": 14592 }, { "epoch": 1.856379595471314, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5767059326171875, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8832489252090454, "num_tokens": 553848186.0, "step": 14593 }, { "epoch": 1.8565068057499046, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2231531143188477, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8753513097763062, "num_tokens": 553888928.0, "step": 14594 }, { "epoch": 1.856634016028495, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.688997507095337, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.859517514705658, "num_tokens": 553929095.0, "step": 14595 }, { "epoch": 1.8567612263070856, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.141146183013916, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8792211413383484, "num_tokens": 553969610.0, "step": 14596 }, { "epoch": 1.8568884365856761, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8759796619415283, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8728362321853638, "num_tokens": 554006707.0, "step": 14597 }, { "epoch": 1.8570156468642667, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.098919153213501, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8618743419647217, "num_tokens": 554049006.0, "step": 14598 }, { "epoch": 1.8571428571428572, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.150737762451172, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8743197917938232, "num_tokens": 554086682.0, "step": 14599 }, { "epoch": 1.8572700674214477, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.20200514793396, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8664306998252869, "num_tokens": 554122478.0, "step": 14600 }, { "epoch": 1.8573972777000383, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.396905899047852, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.878446638584137, "num_tokens": 554162640.0, "step": 14601 }, { "epoch": 1.8575244879786288, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.532360315322876, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8642171621322632, "num_tokens": 554202161.0, "step": 14602 }, { "epoch": 1.8576516982572193, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.398575782775879, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8756584525108337, "num_tokens": 554235401.0, "step": 14603 }, { "epoch": 1.8577789085358098, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.8079206943511963, "learning_rate": 1e-05, "loss": 0.5371, "mean_token_accuracy": 0.8512309789657593, "num_tokens": 554276887.0, "step": 14604 }, { "epoch": 1.8579061188144002, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.168181419372559, "learning_rate": 1e-05, "loss": 0.5193, "mean_token_accuracy": 0.8545878529548645, "num_tokens": 554310513.0, "step": 14605 }, { "epoch": 1.8580333290929907, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.801546573638916, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8737661838531494, "num_tokens": 554354225.0, "step": 14606 }, { "epoch": 1.8581605393715812, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.944765090942383, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8645946979522705, "num_tokens": 554395198.0, "step": 14607 }, { "epoch": 1.8582877496501717, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.7087481021881104, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8609053492546082, "num_tokens": 554430234.0, "step": 14608 }, { "epoch": 1.8584149599287623, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7889087200164795, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8708164691925049, "num_tokens": 554469381.0, "step": 14609 }, { "epoch": 1.8585421702073528, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5995328426361084, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8676323890686035, "num_tokens": 554500409.0, "step": 14610 }, { "epoch": 1.858669380485943, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6872127056121826, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8794883489608765, "num_tokens": 554537785.0, "step": 14611 }, { "epoch": 1.8587965907645336, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.479720115661621, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8618975877761841, "num_tokens": 554578420.0, "step": 14612 }, { "epoch": 1.8589238010431242, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6613123416900635, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8739995956420898, "num_tokens": 554615264.0, "step": 14613 }, { "epoch": 1.8590510113217147, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.389986515045166, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.870975911617279, "num_tokens": 554647327.0, "step": 14614 }, { "epoch": 1.8591782216003052, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6911184787750244, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8731873035430908, "num_tokens": 554682552.0, "step": 14615 }, { "epoch": 1.8593054318788957, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2908997535705566, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8630775213241577, "num_tokens": 554721348.0, "step": 14616 }, { "epoch": 1.8594326421574863, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.78517746925354, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8808014392852783, "num_tokens": 554760618.0, "step": 14617 }, { "epoch": 1.8595598524360768, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.220317840576172, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8649904727935791, "num_tokens": 554801831.0, "step": 14618 }, { "epoch": 1.8596870627146673, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9521265029907227, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8763014078140259, "num_tokens": 554838307.0, "step": 14619 }, { "epoch": 1.8598142729932579, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.232719898223877, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8668946027755737, "num_tokens": 554883833.0, "step": 14620 }, { "epoch": 1.8599414832718484, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.136477470397949, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8695340752601624, "num_tokens": 554920944.0, "step": 14621 }, { "epoch": 1.860068693550439, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.393428087234497, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8650704622268677, "num_tokens": 554958656.0, "step": 14622 }, { "epoch": 1.8601959038290294, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.180182933807373, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8637048602104187, "num_tokens": 555001864.0, "step": 14623 }, { "epoch": 1.86032311410762, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.604368209838867, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8763959407806396, "num_tokens": 555034268.0, "step": 14624 }, { "epoch": 1.8604503243862105, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.013256549835205, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.867315948009491, "num_tokens": 555075275.0, "step": 14625 }, { "epoch": 1.860577534664801, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6904654502868652, "learning_rate": 1e-05, "loss": 0.5633, "mean_token_accuracy": 0.8413835167884827, "num_tokens": 555110961.0, "step": 14626 }, { "epoch": 1.8607047449433916, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.771030902862549, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8570613861083984, "num_tokens": 555150369.0, "step": 14627 }, { "epoch": 1.860831955221982, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6660892963409424, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8614192008972168, "num_tokens": 555182555.0, "step": 14628 }, { "epoch": 1.8609591655005726, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5414867401123047, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8844990134239197, "num_tokens": 555223267.0, "step": 14629 }, { "epoch": 1.861086375779163, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6473116874694824, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8689572811126709, "num_tokens": 555263461.0, "step": 14630 }, { "epoch": 1.8612135860577534, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3928470611572266, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8702086210250854, "num_tokens": 555306201.0, "step": 14631 }, { "epoch": 1.861340796336344, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.662447452545166, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8719968795776367, "num_tokens": 555344418.0, "step": 14632 }, { "epoch": 1.8614680066149345, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2948484420776367, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8689236640930176, "num_tokens": 555386096.0, "step": 14633 }, { "epoch": 1.861595216893525, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.723780632019043, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8746665716171265, "num_tokens": 555423368.0, "step": 14634 }, { "epoch": 1.8617224271721156, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.226839780807495, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8652445673942566, "num_tokens": 555458624.0, "step": 14635 }, { "epoch": 1.8618496374507059, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.84334135055542, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8780374526977539, "num_tokens": 555491433.0, "step": 14636 }, { "epoch": 1.8619768477292964, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.18896484375, "learning_rate": 1e-05, "loss": 0.3919, "mean_token_accuracy": 0.8874249458312988, "num_tokens": 555527198.0, "step": 14637 }, { "epoch": 1.862104058007887, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.008365154266357, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8788491487503052, "num_tokens": 555563232.0, "step": 14638 }, { "epoch": 1.8622312682864774, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2504873275756836, "learning_rate": 1e-05, "loss": 0.4888, "mean_token_accuracy": 0.8586249351501465, "num_tokens": 555599805.0, "step": 14639 }, { "epoch": 1.862358478565068, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.170379638671875, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8659241199493408, "num_tokens": 555628835.0, "step": 14640 }, { "epoch": 1.8624856888436585, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.399285316467285, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8787267208099365, "num_tokens": 555668120.0, "step": 14641 }, { "epoch": 1.862612899122249, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.199830055236816, "learning_rate": 1e-05, "loss": 0.5115, "mean_token_accuracy": 0.8562005758285522, "num_tokens": 555704662.0, "step": 14642 }, { "epoch": 1.8627401094008396, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5843522548675537, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8815650939941406, "num_tokens": 555738867.0, "step": 14643 }, { "epoch": 1.86286731967943, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.076445579528809, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8669966459274292, "num_tokens": 555773500.0, "step": 14644 }, { "epoch": 1.8629945299580206, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.64589786529541, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8852607607841492, "num_tokens": 555812139.0, "step": 14645 }, { "epoch": 1.8631217402366111, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8914217948913574, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8609635233879089, "num_tokens": 555850823.0, "step": 14646 }, { "epoch": 1.8632489505152017, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6219489574432373, "learning_rate": 1e-05, "loss": 0.3972, "mean_token_accuracy": 0.8893197774887085, "num_tokens": 555888554.0, "step": 14647 }, { "epoch": 1.8633761607937922, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7218103408813477, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8691515922546387, "num_tokens": 555932629.0, "step": 14648 }, { "epoch": 1.8635033710723827, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5669033527374268, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8632398843765259, "num_tokens": 555970722.0, "step": 14649 }, { "epoch": 1.8636305813509733, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.613194465637207, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8796586990356445, "num_tokens": 556005784.0, "step": 14650 }, { "epoch": 1.8637577916295638, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4884440898895264, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8793182373046875, "num_tokens": 556046053.0, "step": 14651 }, { "epoch": 1.8638850019081543, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.559400796890259, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8646317720413208, "num_tokens": 556084265.0, "step": 14652 }, { "epoch": 1.8640122121867448, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.418959856033325, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8760256767272949, "num_tokens": 556125062.0, "step": 14653 }, { "epoch": 1.8641394224653351, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.552029609680176, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.85635906457901, "num_tokens": 556162590.0, "step": 14654 }, { "epoch": 1.8642666327439257, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3566558361053467, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8685662150382996, "num_tokens": 556200560.0, "step": 14655 }, { "epoch": 1.8643938430225162, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.588698387145996, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.878128707408905, "num_tokens": 556237199.0, "step": 14656 }, { "epoch": 1.8645210533011067, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2783432006835938, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8689490556716919, "num_tokens": 556278620.0, "step": 14657 }, { "epoch": 1.8646482635796973, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6799545288085938, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8756940364837646, "num_tokens": 556312844.0, "step": 14658 }, { "epoch": 1.8647754738582878, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.211453437805176, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8776580095291138, "num_tokens": 556351218.0, "step": 14659 }, { "epoch": 1.864902684136878, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.805933952331543, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8818402290344238, "num_tokens": 556388176.0, "step": 14660 }, { "epoch": 1.8650298944154686, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1825759410858154, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8659748435020447, "num_tokens": 556424996.0, "step": 14661 }, { "epoch": 1.8651571046940592, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9853360652923584, "learning_rate": 1e-05, "loss": 0.5297, "mean_token_accuracy": 0.852226734161377, "num_tokens": 556465961.0, "step": 14662 }, { "epoch": 1.8652843149726497, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2180933952331543, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8723844289779663, "num_tokens": 556504777.0, "step": 14663 }, { "epoch": 1.8654115252512402, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.160897254943848, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8777408599853516, "num_tokens": 556539737.0, "step": 14664 }, { "epoch": 1.8655387355298307, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3973686695098877, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8848788738250732, "num_tokens": 556578028.0, "step": 14665 }, { "epoch": 1.8656659458084213, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.185614585876465, "learning_rate": 1e-05, "loss": 0.5046, "mean_token_accuracy": 0.8611346483230591, "num_tokens": 556611362.0, "step": 14666 }, { "epoch": 1.8657931560870118, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.600405216217041, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.869925856590271, "num_tokens": 556652329.0, "step": 14667 }, { "epoch": 1.8659203663656023, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.000392436981201, "learning_rate": 1e-05, "loss": 0.3895, "mean_token_accuracy": 0.8948177099227905, "num_tokens": 556689531.0, "step": 14668 }, { "epoch": 1.8660475766441929, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6681935787200928, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8797610998153687, "num_tokens": 556722309.0, "step": 14669 }, { "epoch": 1.8661747869227834, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7901687622070312, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.881171703338623, "num_tokens": 556755682.0, "step": 14670 }, { "epoch": 1.866301997201374, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6327707767486572, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8807293772697449, "num_tokens": 556792591.0, "step": 14671 }, { "epoch": 1.8664292074799644, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.588712692260742, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8772631883621216, "num_tokens": 556835168.0, "step": 14672 }, { "epoch": 1.866556417758555, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.594118118286133, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8740552663803101, "num_tokens": 556876106.0, "step": 14673 }, { "epoch": 1.8666836280371455, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.475005865097046, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8681406378746033, "num_tokens": 556917154.0, "step": 14674 }, { "epoch": 1.866810838315736, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5738985538482666, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8668181300163269, "num_tokens": 556953706.0, "step": 14675 }, { "epoch": 1.8669380485943265, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3843770027160645, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8734087944030762, "num_tokens": 556992035.0, "step": 14676 }, { "epoch": 1.867065258872917, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5618631839752197, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8735493421554565, "num_tokens": 557034785.0, "step": 14677 }, { "epoch": 1.8671924691515076, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3290507793426514, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8626104593276978, "num_tokens": 557074910.0, "step": 14678 }, { "epoch": 1.867319679430098, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.600614070892334, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8733524680137634, "num_tokens": 557105327.0, "step": 14679 }, { "epoch": 1.8674468897086884, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2552130222320557, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8689660429954529, "num_tokens": 557149261.0, "step": 14680 }, { "epoch": 1.867574099987279, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.676666259765625, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8682971596717834, "num_tokens": 557190550.0, "step": 14681 }, { "epoch": 1.8677013102658695, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2132129669189453, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8567084074020386, "num_tokens": 557222828.0, "step": 14682 }, { "epoch": 1.86782852054446, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.78554368019104, "learning_rate": 1e-05, "loss": 0.4791, "mean_token_accuracy": 0.8663632869720459, "num_tokens": 557264228.0, "step": 14683 }, { "epoch": 1.8679557308230506, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.153284788131714, "learning_rate": 1e-05, "loss": 0.3975, "mean_token_accuracy": 0.8896981477737427, "num_tokens": 557301310.0, "step": 14684 }, { "epoch": 1.8680829411016409, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.972278118133545, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.875684380531311, "num_tokens": 557338596.0, "step": 14685 }, { "epoch": 1.8682101513802314, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1702919006347656, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8698028326034546, "num_tokens": 557374939.0, "step": 14686 }, { "epoch": 1.868337361658822, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.198882579803467, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8705464601516724, "num_tokens": 557411711.0, "step": 14687 }, { "epoch": 1.8684645719374124, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.345505475997925, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8733048439025879, "num_tokens": 557448892.0, "step": 14688 }, { "epoch": 1.868591782216003, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.285147666931152, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8649184107780457, "num_tokens": 557490911.0, "step": 14689 }, { "epoch": 1.8687189924945935, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.595381736755371, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.873239278793335, "num_tokens": 557529721.0, "step": 14690 }, { "epoch": 1.868846202773184, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.167332649230957, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.876639723777771, "num_tokens": 557567591.0, "step": 14691 }, { "epoch": 1.8689734130517746, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7005436420440674, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.8674008250236511, "num_tokens": 557609105.0, "step": 14692 }, { "epoch": 1.869100623330365, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9679219722747803, "learning_rate": 1e-05, "loss": 0.5617, "mean_token_accuracy": 0.8402602076530457, "num_tokens": 557658798.0, "step": 14693 }, { "epoch": 1.8692278336089556, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6737465858459473, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8865693807601929, "num_tokens": 557695572.0, "step": 14694 }, { "epoch": 1.8693550438875461, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7936623096466064, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8738277554512024, "num_tokens": 557728796.0, "step": 14695 }, { "epoch": 1.8694822541661367, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6005215644836426, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.8829221725463867, "num_tokens": 557761786.0, "step": 14696 }, { "epoch": 1.8696094644447272, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6539409160614014, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8633835911750793, "num_tokens": 557809454.0, "step": 14697 }, { "epoch": 1.8697366747233177, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.520385980606079, "learning_rate": 1e-05, "loss": 0.3791, "mean_token_accuracy": 0.8922587037086487, "num_tokens": 557846959.0, "step": 14698 }, { "epoch": 1.8698638850019083, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5862674713134766, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8824456930160522, "num_tokens": 557879873.0, "step": 14699 }, { "epoch": 1.8699910952804988, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.451920509338379, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8667248487472534, "num_tokens": 557918020.0, "step": 14700 }, { "epoch": 1.8701183055590893, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5600602626800537, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8673290014266968, "num_tokens": 557960633.0, "step": 14701 }, { "epoch": 1.8702455158376798, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3725249767303467, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8699609041213989, "num_tokens": 557998430.0, "step": 14702 }, { "epoch": 1.8703727261162701, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5752551555633545, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8629377484321594, "num_tokens": 558038283.0, "step": 14703 }, { "epoch": 1.8704999363948607, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2996511459350586, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8714488744735718, "num_tokens": 558079636.0, "step": 14704 }, { "epoch": 1.8706271466734512, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.643099308013916, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8799498677253723, "num_tokens": 558117670.0, "step": 14705 }, { "epoch": 1.8707543569520417, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2412710189819336, "learning_rate": 1e-05, "loss": 0.4941, "mean_token_accuracy": 0.856552004814148, "num_tokens": 558155732.0, "step": 14706 }, { "epoch": 1.8708815672306323, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7589452266693115, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8676775097846985, "num_tokens": 558195389.0, "step": 14707 }, { "epoch": 1.8710087775092228, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1611406803131104, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8833616971969604, "num_tokens": 558239535.0, "step": 14708 }, { "epoch": 1.871135987787813, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9580814838409424, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8702990412712097, "num_tokens": 558277441.0, "step": 14709 }, { "epoch": 1.8712631980664036, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1845028400421143, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.867577314376831, "num_tokens": 558319331.0, "step": 14710 }, { "epoch": 1.8713904083449941, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.165157794952393, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8740177154541016, "num_tokens": 558365494.0, "step": 14711 }, { "epoch": 1.8715176186235847, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.352651834487915, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8649700880050659, "num_tokens": 558407608.0, "step": 14712 }, { "epoch": 1.8716448289021752, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.263453006744385, "learning_rate": 1e-05, "loss": 0.538, "mean_token_accuracy": 0.8551027774810791, "num_tokens": 558444848.0, "step": 14713 }, { "epoch": 1.8717720391807657, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.606828451156616, "learning_rate": 1e-05, "loss": 0.5283, "mean_token_accuracy": 0.8499763011932373, "num_tokens": 558486630.0, "step": 14714 }, { "epoch": 1.8718992494593563, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.101434707641602, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.882278323173523, "num_tokens": 558523609.0, "step": 14715 }, { "epoch": 1.8720264597379468, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6957573890686035, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8780577182769775, "num_tokens": 558561065.0, "step": 14716 }, { "epoch": 1.8721536700165373, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8735408782958984, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8791697025299072, "num_tokens": 558602007.0, "step": 14717 }, { "epoch": 1.8722808802951278, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6923913955688477, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8709191083908081, "num_tokens": 558637712.0, "step": 14718 }, { "epoch": 1.8724080905737184, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6720471382141113, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8754579424858093, "num_tokens": 558671939.0, "step": 14719 }, { "epoch": 1.872535300852309, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6260387897491455, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8673858642578125, "num_tokens": 558708413.0, "step": 14720 }, { "epoch": 1.8726625111308994, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5310301780700684, "learning_rate": 1e-05, "loss": 0.5326, "mean_token_accuracy": 0.8437755107879639, "num_tokens": 558748993.0, "step": 14721 }, { "epoch": 1.87278972140949, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.578523874282837, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8664443492889404, "num_tokens": 558788058.0, "step": 14722 }, { "epoch": 1.8729169316880805, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.437769889831543, "learning_rate": 1e-05, "loss": 0.5245, "mean_token_accuracy": 0.8479009866714478, "num_tokens": 558828046.0, "step": 14723 }, { "epoch": 1.873044141966671, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.551323413848877, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8728708028793335, "num_tokens": 558862445.0, "step": 14724 }, { "epoch": 1.8731713522452615, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.352027177810669, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8820157051086426, "num_tokens": 558907645.0, "step": 14725 }, { "epoch": 1.873298562523852, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5669803619384766, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8733817338943481, "num_tokens": 558943253.0, "step": 14726 }, { "epoch": 1.8734257728024426, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3007941246032715, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8757152557373047, "num_tokens": 558980341.0, "step": 14727 }, { "epoch": 1.873552983081033, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.621288299560547, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8603994846343994, "num_tokens": 559018659.0, "step": 14728 }, { "epoch": 1.8736801933596234, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.235140800476074, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.873528778553009, "num_tokens": 559060254.0, "step": 14729 }, { "epoch": 1.873807403638214, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7077457904815674, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8733139038085938, "num_tokens": 559099166.0, "step": 14730 }, { "epoch": 1.8739346139168045, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.16926646232605, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8795228004455566, "num_tokens": 559135189.0, "step": 14731 }, { "epoch": 1.874061824195395, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8765041828155518, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8733147382736206, "num_tokens": 559178491.0, "step": 14732 }, { "epoch": 1.8741890344739855, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.14068603515625, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8635968565940857, "num_tokens": 559213977.0, "step": 14733 }, { "epoch": 1.8743162447525759, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.094480037689209, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.870265007019043, "num_tokens": 559254309.0, "step": 14734 }, { "epoch": 1.8744434550311664, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.226461172103882, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8839412331581116, "num_tokens": 559289756.0, "step": 14735 }, { "epoch": 1.874570665309757, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2833781242370605, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8721487522125244, "num_tokens": 559331143.0, "step": 14736 }, { "epoch": 1.8746978755883474, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4849886894226074, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8694584965705872, "num_tokens": 559368256.0, "step": 14737 }, { "epoch": 1.874825085866938, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.276388645172119, "learning_rate": 1e-05, "loss": 0.5252, "mean_token_accuracy": 0.8520389199256897, "num_tokens": 559408853.0, "step": 14738 }, { "epoch": 1.8749522961455285, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.713773012161255, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8616669178009033, "num_tokens": 559445606.0, "step": 14739 }, { "epoch": 1.875079506424119, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.071511745452881, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8737502694129944, "num_tokens": 559482482.0, "step": 14740 }, { "epoch": 1.8752067167027096, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.700589418411255, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8710712194442749, "num_tokens": 559522757.0, "step": 14741 }, { "epoch": 1.8753339269813, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8783349990844727, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8788334131240845, "num_tokens": 559556684.0, "step": 14742 }, { "epoch": 1.8754611372598906, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6401937007904053, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8705962300300598, "num_tokens": 559597548.0, "step": 14743 }, { "epoch": 1.8755883475384811, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7185819149017334, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8651604652404785, "num_tokens": 559638440.0, "step": 14744 }, { "epoch": 1.8757155578170717, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.559720277786255, "learning_rate": 1e-05, "loss": 0.5241, "mean_token_accuracy": 0.8491545915603638, "num_tokens": 559680819.0, "step": 14745 }, { "epoch": 1.8758427680956622, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.621502637863159, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8682798147201538, "num_tokens": 559721861.0, "step": 14746 }, { "epoch": 1.8759699783742527, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.486521005630493, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8698217868804932, "num_tokens": 559757370.0, "step": 14747 }, { "epoch": 1.8760971886528433, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5789692401885986, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.864578127861023, "num_tokens": 559795570.0, "step": 14748 }, { "epoch": 1.8762243989314338, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3975024223327637, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.875613808631897, "num_tokens": 559835115.0, "step": 14749 }, { "epoch": 1.8763516092100243, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5831103324890137, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8655192255973816, "num_tokens": 559876119.0, "step": 14750 }, { "epoch": 1.8764788194886148, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.33231782913208, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8755916357040405, "num_tokens": 559908349.0, "step": 14751 }, { "epoch": 1.8766060297672051, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6483092308044434, "learning_rate": 1e-05, "loss": 0.4824, "mean_token_accuracy": 0.8568437099456787, "num_tokens": 559947336.0, "step": 14752 }, { "epoch": 1.8767332400457957, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.254124402999878, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8622431755065918, "num_tokens": 559989571.0, "step": 14753 }, { "epoch": 1.8768604503243862, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.758620023727417, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8519921898841858, "num_tokens": 560022840.0, "step": 14754 }, { "epoch": 1.8769876606029767, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1925129890441895, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8625876903533936, "num_tokens": 560061939.0, "step": 14755 }, { "epoch": 1.8771148708815673, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9289205074310303, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8672776222229004, "num_tokens": 560097480.0, "step": 14756 }, { "epoch": 1.8772420811601578, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1883323192596436, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8691307902336121, "num_tokens": 560137957.0, "step": 14757 }, { "epoch": 1.877369291438748, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.12901496887207, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8656061887741089, "num_tokens": 560173717.0, "step": 14758 }, { "epoch": 1.8774965017173386, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000270843505859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3404531478881836, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8691373467445374, "num_tokens": 560209833.0, "step": 14759 }, { "epoch": 1.8776237119959291, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.21285343170166, "learning_rate": 1e-05, "loss": 0.5533, "mean_token_accuracy": 0.844931423664093, "num_tokens": 560250406.0, "step": 14760 }, { "epoch": 1.8777509222745197, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5530896186828613, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8696591854095459, "num_tokens": 560288828.0, "step": 14761 }, { "epoch": 1.8778781325531102, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.0918803215026855, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8789557218551636, "num_tokens": 560327431.0, "step": 14762 }, { "epoch": 1.8780053428317007, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6576833724975586, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8881810307502747, "num_tokens": 560367767.0, "step": 14763 }, { "epoch": 1.8781325531102913, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.893113613128662, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8826262354850769, "num_tokens": 560401179.0, "step": 14764 }, { "epoch": 1.8782597633888818, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.672264814376831, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8647151589393616, "num_tokens": 560431198.0, "step": 14765 }, { "epoch": 1.8783869736674723, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.681391716003418, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8708551526069641, "num_tokens": 560470244.0, "step": 14766 }, { "epoch": 1.8785141839460628, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6125600337982178, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8781267404556274, "num_tokens": 560511696.0, "step": 14767 }, { "epoch": 1.8786413942246534, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.519362211227417, "learning_rate": 1e-05, "loss": 0.3845, "mean_token_accuracy": 0.8945870995521545, "num_tokens": 560547555.0, "step": 14768 }, { "epoch": 1.878768604503244, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.584798812866211, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8687591552734375, "num_tokens": 560584349.0, "step": 14769 }, { "epoch": 1.8788958147818344, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4151880741119385, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8783036470413208, "num_tokens": 560628403.0, "step": 14770 }, { "epoch": 1.879023025060425, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.581641674041748, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8744653463363647, "num_tokens": 560663454.0, "step": 14771 }, { "epoch": 1.8791502353390155, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3430707454681396, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8687394857406616, "num_tokens": 560700506.0, "step": 14772 }, { "epoch": 1.879277445617606, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5962674617767334, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8768803477287292, "num_tokens": 560739690.0, "step": 14773 }, { "epoch": 1.8794046558961965, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.270087480545044, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.872753918170929, "num_tokens": 560786364.0, "step": 14774 }, { "epoch": 1.879531866174787, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.661334276199341, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8648302555084229, "num_tokens": 560828850.0, "step": 14775 }, { "epoch": 1.8796590764533776, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.223910093307495, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.878415048122406, "num_tokens": 560863401.0, "step": 14776 }, { "epoch": 1.879786286731968, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.778977632522583, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8705942630767822, "num_tokens": 560894347.0, "step": 14777 }, { "epoch": 1.8799134970105584, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.16413950920105, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8560972809791565, "num_tokens": 560935302.0, "step": 14778 }, { "epoch": 1.880040707289149, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.947427749633789, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8803266882896423, "num_tokens": 560971960.0, "step": 14779 }, { "epoch": 1.8801679175677395, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1660032272338867, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8641324043273926, "num_tokens": 561012267.0, "step": 14780 }, { "epoch": 1.88029512784633, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.1717400550842285, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8692819476127625, "num_tokens": 561048861.0, "step": 14781 }, { "epoch": 1.8804223381249205, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3159971237182617, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8736618161201477, "num_tokens": 561087832.0, "step": 14782 }, { "epoch": 1.8805495484035109, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.283447265625, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8800185918807983, "num_tokens": 561120657.0, "step": 14783 }, { "epoch": 1.8806767586821014, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.566030502319336, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8809137940406799, "num_tokens": 561161762.0, "step": 14784 }, { "epoch": 1.880803968960692, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.184633255004883, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8729840517044067, "num_tokens": 561199459.0, "step": 14785 }, { "epoch": 1.8809311792392824, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7088711261749268, "learning_rate": 1e-05, "loss": 0.5182, "mean_token_accuracy": 0.8564932346343994, "num_tokens": 561237958.0, "step": 14786 }, { "epoch": 1.881058389517873, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.976853847503662, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8650221824645996, "num_tokens": 561272086.0, "step": 14787 }, { "epoch": 1.8811855997964635, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6794421672821045, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8788166046142578, "num_tokens": 561309821.0, "step": 14788 }, { "epoch": 1.881312810075054, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.778376817703247, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8688086271286011, "num_tokens": 561351565.0, "step": 14789 }, { "epoch": 1.8814400203536445, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6245853900909424, "learning_rate": 1e-05, "loss": 0.5246, "mean_token_accuracy": 0.8493518233299255, "num_tokens": 561387399.0, "step": 14790 }, { "epoch": 1.881567230632235, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.632114887237549, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8658446073532104, "num_tokens": 561427917.0, "step": 14791 }, { "epoch": 1.8816944409108256, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5520858764648438, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8657265305519104, "num_tokens": 561465398.0, "step": 14792 }, { "epoch": 1.8818216511894161, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.531764507293701, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8731237649917603, "num_tokens": 561510264.0, "step": 14793 }, { "epoch": 1.8819488614680067, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.484785556793213, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8655869364738464, "num_tokens": 561550942.0, "step": 14794 }, { "epoch": 1.8820760717465972, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.506099224090576, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8767893314361572, "num_tokens": 561580571.0, "step": 14795 }, { "epoch": 1.8822032820251877, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4218549728393555, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8739941120147705, "num_tokens": 561618993.0, "step": 14796 }, { "epoch": 1.8823304923037782, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4978322982788086, "learning_rate": 1e-05, "loss": 0.51, "mean_token_accuracy": 0.8527790307998657, "num_tokens": 561659389.0, "step": 14797 }, { "epoch": 1.8824577025823688, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.366481065750122, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8705017566680908, "num_tokens": 561707703.0, "step": 14798 }, { "epoch": 1.8825849128609593, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.535374164581299, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8692685961723328, "num_tokens": 561749367.0, "step": 14799 }, { "epoch": 1.8827121231395498, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3090126514434814, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8607379198074341, "num_tokens": 561785234.0, "step": 14800 }, { "epoch": 1.8828393334181401, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5833256244659424, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8785979151725769, "num_tokens": 561825801.0, "step": 14801 }, { "epoch": 1.8829665436967307, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2330007553100586, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8614687919616699, "num_tokens": 561865030.0, "step": 14802 }, { "epoch": 1.8830937539753212, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6955840587615967, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8810908198356628, "num_tokens": 561899767.0, "step": 14803 }, { "epoch": 1.8832209642539117, "ewc_loss": 0.07177734375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.1625113487243652, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8640202283859253, "num_tokens": 561936794.0, "step": 14804 }, { "epoch": 1.8833481745325023, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.881551504135132, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.868998646736145, "num_tokens": 561971770.0, "step": 14805 }, { "epoch": 1.8834753848110928, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1157407760620117, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8666381239891052, "num_tokens": 562016400.0, "step": 14806 }, { "epoch": 1.883602595089683, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.144037246704102, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8770954608917236, "num_tokens": 562050732.0, "step": 14807 }, { "epoch": 1.8837298053682736, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.2286458015441895, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8715484738349915, "num_tokens": 562093619.0, "step": 14808 }, { "epoch": 1.8838570156468641, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.383391380310059, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8715317249298096, "num_tokens": 562131434.0, "step": 14809 }, { "epoch": 1.8839842259254547, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5812997817993164, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.85933518409729, "num_tokens": 562169914.0, "step": 14810 }, { "epoch": 1.8841114362040452, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.313453674316406, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8750012516975403, "num_tokens": 562207190.0, "step": 14811 }, { "epoch": 1.8842386464826357, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.765144109725952, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.8641552329063416, "num_tokens": 562246236.0, "step": 14812 }, { "epoch": 1.8843658567612263, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.094475746154785, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8818113803863525, "num_tokens": 562286875.0, "step": 14813 }, { "epoch": 1.8844930670398168, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7620999813079834, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.875606894493103, "num_tokens": 562325807.0, "step": 14814 }, { "epoch": 1.8846202773184073, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8893160820007324, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8726938962936401, "num_tokens": 562363663.0, "step": 14815 }, { "epoch": 1.8847474875969978, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6703295707702637, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.871306300163269, "num_tokens": 562399700.0, "step": 14816 }, { "epoch": 1.8848746978755884, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.743561267852783, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.8608522415161133, "num_tokens": 562434395.0, "step": 14817 }, { "epoch": 1.885001908154179, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5702195167541504, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8544452786445618, "num_tokens": 562472140.0, "step": 14818 }, { "epoch": 1.8851291184327694, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6567459106445312, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8691703081130981, "num_tokens": 562510556.0, "step": 14819 }, { "epoch": 1.88525632871136, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.457869529724121, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8668084144592285, "num_tokens": 562546679.0, "step": 14820 }, { "epoch": 1.8853835389899505, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.644021511077881, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.8663690686225891, "num_tokens": 562580707.0, "step": 14821 }, { "epoch": 1.885510749268541, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3596816062927246, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8697295188903809, "num_tokens": 562621753.0, "step": 14822 }, { "epoch": 1.8856379595471315, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.686692953109741, "learning_rate": 1e-05, "loss": 0.3941, "mean_token_accuracy": 0.8898236155509949, "num_tokens": 562654064.0, "step": 14823 }, { "epoch": 1.885765169825722, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2843286991119385, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.860365629196167, "num_tokens": 562692575.0, "step": 14824 }, { "epoch": 1.8858923801043126, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7836480140686035, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.881667971611023, "num_tokens": 562732655.0, "step": 14825 }, { "epoch": 1.886019590382903, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.225216865539551, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8682237863540649, "num_tokens": 562768618.0, "step": 14826 }, { "epoch": 1.8861468006614934, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9477951526641846, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8565062284469604, "num_tokens": 562805357.0, "step": 14827 }, { "epoch": 1.886274010940084, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2349565029144287, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.864986777305603, "num_tokens": 562848798.0, "step": 14828 }, { "epoch": 1.8864012212186745, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.117016792297363, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8788790106773376, "num_tokens": 562890108.0, "step": 14829 }, { "epoch": 1.886528431497265, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3892812728881836, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8766089677810669, "num_tokens": 562921089.0, "step": 14830 }, { "epoch": 1.8866556417758555, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.156223773956299, "learning_rate": 1e-05, "loss": 0.5651, "mean_token_accuracy": 0.8367283940315247, "num_tokens": 562961705.0, "step": 14831 }, { "epoch": 1.8867828520544458, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5496506690979004, "learning_rate": 1e-05, "loss": 0.4121, "mean_token_accuracy": 0.8859548568725586, "num_tokens": 563000011.0, "step": 14832 }, { "epoch": 1.8869100623330364, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.022653102874756, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8744910955429077, "num_tokens": 563042568.0, "step": 14833 }, { "epoch": 1.887037272611627, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.640691041946411, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8780509829521179, "num_tokens": 563075892.0, "step": 14834 }, { "epoch": 1.8871644828902174, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8193647861480713, "learning_rate": 1e-05, "loss": 0.485, "mean_token_accuracy": 0.8691827058792114, "num_tokens": 563117276.0, "step": 14835 }, { "epoch": 1.887291693168808, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.643968105316162, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8711075782775879, "num_tokens": 563152557.0, "step": 14836 }, { "epoch": 1.8874189034473985, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.607855796813965, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8730288743972778, "num_tokens": 563188620.0, "step": 14837 }, { "epoch": 1.887546113725989, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.616037607192993, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8649299144744873, "num_tokens": 563227582.0, "step": 14838 }, { "epoch": 1.8876733240045795, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4467573165893555, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8823367953300476, "num_tokens": 563260057.0, "step": 14839 }, { "epoch": 1.88780053428317, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.608167886734009, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8698954582214355, "num_tokens": 563299612.0, "step": 14840 }, { "epoch": 1.8879277445617606, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3504128456115723, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8727184534072876, "num_tokens": 563331659.0, "step": 14841 }, { "epoch": 1.8880549548403511, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6322436332702637, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8627743721008301, "num_tokens": 563371411.0, "step": 14842 }, { "epoch": 1.8881821651189417, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2775163650512695, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8705815672874451, "num_tokens": 563401498.0, "step": 14843 }, { "epoch": 1.8883093753975322, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6896815299987793, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8854538202285767, "num_tokens": 563441383.0, "step": 14844 }, { "epoch": 1.8884365856761227, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.221550464630127, "learning_rate": 1e-05, "loss": 0.4726, "mean_token_accuracy": 0.8609952926635742, "num_tokens": 563478547.0, "step": 14845 }, { "epoch": 1.8885637959547132, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.796144485473633, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8697608709335327, "num_tokens": 563521366.0, "step": 14846 }, { "epoch": 1.8886910062333038, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1751842498779297, "learning_rate": 1e-05, "loss": 0.4962, "mean_token_accuracy": 0.8535906672477722, "num_tokens": 563560062.0, "step": 14847 }, { "epoch": 1.8888182165118943, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9738032817840576, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8630357980728149, "num_tokens": 563598448.0, "step": 14848 }, { "epoch": 1.8889454267904848, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.1883671283721924, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8780637979507446, "num_tokens": 563639274.0, "step": 14849 }, { "epoch": 1.8890726370690751, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.162655353546143, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8682825565338135, "num_tokens": 563679309.0, "step": 14850 }, { "epoch": 1.8891998473476657, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3505046367645264, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8693780303001404, "num_tokens": 563717594.0, "step": 14851 }, { "epoch": 1.8893270576262562, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.246382236480713, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8800992369651794, "num_tokens": 563749442.0, "step": 14852 }, { "epoch": 1.8894542679048467, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.573920249938965, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8628308773040771, "num_tokens": 563788519.0, "step": 14853 }, { "epoch": 1.8895814781834372, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.138612747192383, "learning_rate": 1e-05, "loss": 0.499, "mean_token_accuracy": 0.8613256216049194, "num_tokens": 563827233.0, "step": 14854 }, { "epoch": 1.8897086884620278, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.678048849105835, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8693954348564148, "num_tokens": 563863534.0, "step": 14855 }, { "epoch": 1.889835898740618, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9488627910614014, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8531970977783203, "num_tokens": 563903237.0, "step": 14856 }, { "epoch": 1.8899631090192086, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.646575689315796, "learning_rate": 1e-05, "loss": 0.3846, "mean_token_accuracy": 0.8926003575325012, "num_tokens": 563944334.0, "step": 14857 }, { "epoch": 1.8900903192977991, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.760025978088379, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.866237998008728, "num_tokens": 563979801.0, "step": 14858 }, { "epoch": 1.8902175295763897, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5889487266540527, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8638396263122559, "num_tokens": 564022348.0, "step": 14859 }, { "epoch": 1.8903447398549802, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.632115125656128, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.883442223072052, "num_tokens": 564063285.0, "step": 14860 }, { "epoch": 1.8904719501335707, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5192081928253174, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8697943687438965, "num_tokens": 564105323.0, "step": 14861 }, { "epoch": 1.8905991604121613, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5815351009368896, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8604846000671387, "num_tokens": 564140238.0, "step": 14862 }, { "epoch": 1.8907263706907518, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4582321643829346, "learning_rate": 1e-05, "loss": 0.5356, "mean_token_accuracy": 0.842415452003479, "num_tokens": 564178441.0, "step": 14863 }, { "epoch": 1.8908535809693423, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.533601760864258, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8715834617614746, "num_tokens": 564217586.0, "step": 14864 }, { "epoch": 1.8909807912479328, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3881561756134033, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8597229719161987, "num_tokens": 564259198.0, "step": 14865 }, { "epoch": 1.8911080015265234, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.538416624069214, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.884027361869812, "num_tokens": 564298622.0, "step": 14866 }, { "epoch": 1.891235211805114, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.328824758529663, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8761234283447266, "num_tokens": 564331573.0, "step": 14867 }, { "epoch": 1.8913624220837044, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.592855930328369, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8639258146286011, "num_tokens": 564371461.0, "step": 14868 }, { "epoch": 1.891489632362295, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2512755393981934, "learning_rate": 1e-05, "loss": 0.3865, "mean_token_accuracy": 0.8890671730041504, "num_tokens": 564410160.0, "step": 14869 }, { "epoch": 1.8916168426408855, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.695251703262329, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.8588577508926392, "num_tokens": 564455954.0, "step": 14870 }, { "epoch": 1.891744052919476, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1911871433258057, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8704497814178467, "num_tokens": 564490196.0, "step": 14871 }, { "epoch": 1.8918712631980665, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8583462238311768, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8759944438934326, "num_tokens": 564530362.0, "step": 14872 }, { "epoch": 1.891998473476657, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.159517526626587, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8675404191017151, "num_tokens": 564563528.0, "step": 14873 }, { "epoch": 1.8921256837552476, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.064442157745361, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.872124195098877, "num_tokens": 564596865.0, "step": 14874 }, { "epoch": 1.892252894033838, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2390267848968506, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8634557127952576, "num_tokens": 564635740.0, "step": 14875 }, { "epoch": 1.8923801043124284, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.233551979064941, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.8595647811889648, "num_tokens": 564679007.0, "step": 14876 }, { "epoch": 1.892507314591019, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.453948974609375, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8757169842720032, "num_tokens": 564719211.0, "step": 14877 }, { "epoch": 1.8926345248696095, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.219386100769043, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.88444983959198, "num_tokens": 564756482.0, "step": 14878 }, { "epoch": 1.8927617351482, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.653780937194824, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.871174156665802, "num_tokens": 564796304.0, "step": 14879 }, { "epoch": 1.8928889454267905, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.043490886688232, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8667029142379761, "num_tokens": 564835444.0, "step": 14880 }, { "epoch": 1.8930161557053808, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6989641189575195, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.8558414578437805, "num_tokens": 564873637.0, "step": 14881 }, { "epoch": 1.8931433659839714, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8133108615875244, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8725214004516602, "num_tokens": 564914742.0, "step": 14882 }, { "epoch": 1.893270576262562, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.647334098815918, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8711984753608704, "num_tokens": 564953519.0, "step": 14883 }, { "epoch": 1.8933977865411524, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6481986045837402, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8638453483581543, "num_tokens": 564984994.0, "step": 14884 }, { "epoch": 1.893524996819743, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.589472770690918, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8598793745040894, "num_tokens": 565028584.0, "step": 14885 }, { "epoch": 1.8936522070983335, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5112416744232178, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8693653345108032, "num_tokens": 565068726.0, "step": 14886 }, { "epoch": 1.893779417376924, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5416572093963623, "learning_rate": 1e-05, "loss": 0.4946, "mean_token_accuracy": 0.8614600300788879, "num_tokens": 565112156.0, "step": 14887 }, { "epoch": 1.8939066276555145, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4455578327178955, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8730826377868652, "num_tokens": 565146158.0, "step": 14888 }, { "epoch": 1.894033837934105, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5170629024505615, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8826156258583069, "num_tokens": 565176571.0, "step": 14889 }, { "epoch": 1.8941610482126956, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.390291213989258, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8677785992622375, "num_tokens": 565216162.0, "step": 14890 }, { "epoch": 1.8942882584912861, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.504101276397705, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8838321566581726, "num_tokens": 565252010.0, "step": 14891 }, { "epoch": 1.8944154687698767, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3441977500915527, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8538362383842468, "num_tokens": 565297159.0, "step": 14892 }, { "epoch": 1.8945426790484672, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.510347604751587, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8730452060699463, "num_tokens": 565337074.0, "step": 14893 }, { "epoch": 1.8946698893270577, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.308091640472412, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8631080389022827, "num_tokens": 565379551.0, "step": 14894 }, { "epoch": 1.8947970996056482, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5541532039642334, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8737043142318726, "num_tokens": 565418611.0, "step": 14895 }, { "epoch": 1.8949243098842388, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.257622003555298, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8722341060638428, "num_tokens": 565452804.0, "step": 14896 }, { "epoch": 1.8950515201628293, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6334497928619385, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.865403413772583, "num_tokens": 565490628.0, "step": 14897 }, { "epoch": 1.8951787304414198, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.1937153339385986, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8639522194862366, "num_tokens": 565523990.0, "step": 14898 }, { "epoch": 1.8953059407200101, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7670745849609375, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8768247365951538, "num_tokens": 565562684.0, "step": 14899 }, { "epoch": 1.8954331509986007, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1166558265686035, "learning_rate": 1e-05, "loss": 0.3774, "mean_token_accuracy": 0.8912963271141052, "num_tokens": 565596479.0, "step": 14900 }, { "epoch": 1.8955603612771912, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.0083909034729, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8681321144104004, "num_tokens": 565639112.0, "step": 14901 }, { "epoch": 1.8956875715557817, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1304545402526855, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8779407739639282, "num_tokens": 565671379.0, "step": 14902 }, { "epoch": 1.8958147818343722, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.292248249053955, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8770519495010376, "num_tokens": 565708115.0, "step": 14903 }, { "epoch": 1.8959419921129628, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.3580923080444336, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.86346834897995, "num_tokens": 565744767.0, "step": 14904 }, { "epoch": 1.896069202391553, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.426061153411865, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8699220418930054, "num_tokens": 565783709.0, "step": 14905 }, { "epoch": 1.8961964126701436, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.688628673553467, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8742475509643555, "num_tokens": 565818503.0, "step": 14906 }, { "epoch": 1.8963236229487341, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.289461612701416, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8885913491249084, "num_tokens": 565850188.0, "step": 14907 }, { "epoch": 1.8964508332273247, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.821420669555664, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8630989789962769, "num_tokens": 565889177.0, "step": 14908 }, { "epoch": 1.8965780435059152, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.067200183868408, "learning_rate": 1e-05, "loss": 0.5352, "mean_token_accuracy": 0.8496137857437134, "num_tokens": 565929316.0, "step": 14909 }, { "epoch": 1.8967052537845057, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7665693759918213, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8869831562042236, "num_tokens": 565961874.0, "step": 14910 }, { "epoch": 1.8968324640630962, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8836236000061035, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8552323579788208, "num_tokens": 565995143.0, "step": 14911 }, { "epoch": 1.8969596743416868, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.654099941253662, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8692067861557007, "num_tokens": 566032297.0, "step": 14912 }, { "epoch": 1.8970868846202773, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.746772289276123, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8749065399169922, "num_tokens": 566073138.0, "step": 14913 }, { "epoch": 1.8972140948988678, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.536763906478882, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8748565912246704, "num_tokens": 566115471.0, "step": 14914 }, { "epoch": 1.8973413051774584, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6903469562530518, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8670492172241211, "num_tokens": 566152644.0, "step": 14915 }, { "epoch": 1.8974685154560489, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4263880252838135, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8679367303848267, "num_tokens": 566197411.0, "step": 14916 }, { "epoch": 1.8975957257346394, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6894402503967285, "learning_rate": 1e-05, "loss": 0.3774, "mean_token_accuracy": 0.8936575651168823, "num_tokens": 566234738.0, "step": 14917 }, { "epoch": 1.89772293601323, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.342478036880493, "learning_rate": 1e-05, "loss": 0.5054, "mean_token_accuracy": 0.8526219129562378, "num_tokens": 566274445.0, "step": 14918 }, { "epoch": 1.8978501462918205, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.756256341934204, "learning_rate": 1e-05, "loss": 0.4024, "mean_token_accuracy": 0.8892513513565063, "num_tokens": 566312710.0, "step": 14919 }, { "epoch": 1.897977356570411, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.270225763320923, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8680789470672607, "num_tokens": 566349229.0, "step": 14920 }, { "epoch": 1.8981045668490015, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.891979455947876, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8639509677886963, "num_tokens": 566387208.0, "step": 14921 }, { "epoch": 1.898231777127592, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2398533821105957, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8707607388496399, "num_tokens": 566423616.0, "step": 14922 }, { "epoch": 1.8983589874061826, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.06416130065918, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8734788298606873, "num_tokens": 566457984.0, "step": 14923 }, { "epoch": 1.898486197684773, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.339459180831909, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8659074902534485, "num_tokens": 566495664.0, "step": 14924 }, { "epoch": 1.8986134079633634, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.147520065307617, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8627562522888184, "num_tokens": 566533750.0, "step": 14925 }, { "epoch": 1.898740618241954, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5194003582000732, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.8615837693214417, "num_tokens": 566576247.0, "step": 14926 }, { "epoch": 1.8988678285205445, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.051534175872803, "learning_rate": 1e-05, "loss": 0.5171, "mean_token_accuracy": 0.8547433018684387, "num_tokens": 566618185.0, "step": 14927 }, { "epoch": 1.898995038799135, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.636169672012329, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8585833311080933, "num_tokens": 566659219.0, "step": 14928 }, { "epoch": 1.8991222490777255, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8351330757141113, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8776158690452576, "num_tokens": 566695101.0, "step": 14929 }, { "epoch": 1.8992494593563158, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6507456302642822, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8687742948532104, "num_tokens": 566735286.0, "step": 14930 }, { "epoch": 1.8993766696349064, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.61448335647583, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8715483546257019, "num_tokens": 566773534.0, "step": 14931 }, { "epoch": 1.899503879913497, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6325016021728516, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8759819269180298, "num_tokens": 566819141.0, "step": 14932 }, { "epoch": 1.8996310901920874, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.445561170578003, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8859095573425293, "num_tokens": 566855095.0, "step": 14933 }, { "epoch": 1.899758300470678, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.646718740463257, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8626682162284851, "num_tokens": 566888989.0, "step": 14934 }, { "epoch": 1.8998855107492685, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.3204903602600098, "learning_rate": 1e-05, "loss": 0.4097, "mean_token_accuracy": 0.8824418783187866, "num_tokens": 566923049.0, "step": 14935 }, { "epoch": 1.900012721027859, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.70121169090271, "learning_rate": 1e-05, "loss": 0.5219, "mean_token_accuracy": 0.8509776592254639, "num_tokens": 566962049.0, "step": 14936 }, { "epoch": 1.9001399313064495, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.221329689025879, "learning_rate": 1e-05, "loss": 0.4009, "mean_token_accuracy": 0.8854779005050659, "num_tokens": 567003538.0, "step": 14937 }, { "epoch": 1.90026714158504, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8151257038116455, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8812335133552551, "num_tokens": 567035821.0, "step": 14938 }, { "epoch": 1.9003943518636306, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.192654848098755, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8790438771247864, "num_tokens": 567068095.0, "step": 14939 }, { "epoch": 1.9005215621422211, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9791502952575684, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8741493821144104, "num_tokens": 567104741.0, "step": 14940 }, { "epoch": 1.9006487724208116, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.216872453689575, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8679906129837036, "num_tokens": 567142500.0, "step": 14941 }, { "epoch": 1.9007759826994022, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1466779708862305, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8681743741035461, "num_tokens": 567182001.0, "step": 14942 }, { "epoch": 1.9009031929779927, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3661930561065674, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8673029541969299, "num_tokens": 567219809.0, "step": 14943 }, { "epoch": 1.9010304032565832, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.221283435821533, "learning_rate": 1e-05, "loss": 0.5546, "mean_token_accuracy": 0.8423868417739868, "num_tokens": 567260309.0, "step": 14944 }, { "epoch": 1.9011576135351738, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.572197675704956, "learning_rate": 1e-05, "loss": 0.4956, "mean_token_accuracy": 0.8577848672866821, "num_tokens": 567295840.0, "step": 14945 }, { "epoch": 1.9012848238137643, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.101901054382324, "learning_rate": 1e-05, "loss": 0.5311, "mean_token_accuracy": 0.8499596118927002, "num_tokens": 567335546.0, "step": 14946 }, { "epoch": 1.9014120340923548, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.65010142326355, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8872087001800537, "num_tokens": 567372996.0, "step": 14947 }, { "epoch": 1.9015392443709451, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.923332691192627, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8784650564193726, "num_tokens": 567411758.0, "step": 14948 }, { "epoch": 1.9016664546495357, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6391172409057617, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8681021928787231, "num_tokens": 567446859.0, "step": 14949 }, { "epoch": 1.9017936649281262, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7573020458221436, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8799669146537781, "num_tokens": 567480890.0, "step": 14950 }, { "epoch": 1.9019208752067167, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5718190670013428, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8762363195419312, "num_tokens": 567523381.0, "step": 14951 }, { "epoch": 1.9020480854853072, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6364855766296387, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.869294285774231, "num_tokens": 567564926.0, "step": 14952 }, { "epoch": 1.9021752957638978, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5103487968444824, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8679362535476685, "num_tokens": 567603351.0, "step": 14953 }, { "epoch": 1.902302506042488, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.57302188873291, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8837735056877136, "num_tokens": 567636710.0, "step": 14954 }, { "epoch": 1.9024297163210786, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.435521125793457, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.8825151324272156, "num_tokens": 567676325.0, "step": 14955 }, { "epoch": 1.9025569265996691, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5501585006713867, "learning_rate": 1e-05, "loss": 0.3949, "mean_token_accuracy": 0.890101432800293, "num_tokens": 567720663.0, "step": 14956 }, { "epoch": 1.9026841368782597, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.37144136428833, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8618948459625244, "num_tokens": 567761589.0, "step": 14957 }, { "epoch": 1.9028113471568502, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5692193508148193, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8814587593078613, "num_tokens": 567801649.0, "step": 14958 }, { "epoch": 1.9029385574354407, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.313309669494629, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8794577121734619, "num_tokens": 567839953.0, "step": 14959 }, { "epoch": 1.9030657677140312, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6394617557525635, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.87772536277771, "num_tokens": 567877634.0, "step": 14960 }, { "epoch": 1.9031929779926218, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2557950019836426, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.8573249578475952, "num_tokens": 567911504.0, "step": 14961 }, { "epoch": 1.9033201882712123, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.734468698501587, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8753024339675903, "num_tokens": 567954904.0, "step": 14962 }, { "epoch": 1.9034473985498028, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2184736728668213, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8639324903488159, "num_tokens": 567992384.0, "step": 14963 }, { "epoch": 1.9035746088283934, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8687610626220703, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8678953051567078, "num_tokens": 568027692.0, "step": 14964 }, { "epoch": 1.9037018191069839, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.190119504928589, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8651894330978394, "num_tokens": 568063385.0, "step": 14965 }, { "epoch": 1.9038290293855744, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.042366981506348, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8809226751327515, "num_tokens": 568098315.0, "step": 14966 }, { "epoch": 1.903956239664165, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2653920650482178, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8611295819282532, "num_tokens": 568144032.0, "step": 14967 }, { "epoch": 1.9040834499427555, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1807146072387695, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8669877648353577, "num_tokens": 568174988.0, "step": 14968 }, { "epoch": 1.904210660221346, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.4578287601470947, "learning_rate": 1e-05, "loss": 0.4993, "mean_token_accuracy": 0.8580586910247803, "num_tokens": 568219794.0, "step": 14969 }, { "epoch": 1.9043378704999365, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.131594657897949, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8773030042648315, "num_tokens": 568264995.0, "step": 14970 }, { "epoch": 1.904465080778527, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6149356365203857, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8737779259681702, "num_tokens": 568300876.0, "step": 14971 }, { "epoch": 1.9045922910571176, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9610071182250977, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8671876192092896, "num_tokens": 568337341.0, "step": 14972 }, { "epoch": 1.9047195013357079, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6535120010375977, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8659157752990723, "num_tokens": 568373524.0, "step": 14973 }, { "epoch": 1.9048467116142984, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.732926845550537, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8845522403717041, "num_tokens": 568412575.0, "step": 14974 }, { "epoch": 1.904973921892889, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.620894193649292, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8741811513900757, "num_tokens": 568446794.0, "step": 14975 }, { "epoch": 1.9051011321714795, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5591020584106445, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8664939999580383, "num_tokens": 568490645.0, "step": 14976 }, { "epoch": 1.90522834245007, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5936429500579834, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8720616698265076, "num_tokens": 568529281.0, "step": 14977 }, { "epoch": 1.9053555527286605, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.437195062637329, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8802653551101685, "num_tokens": 568568722.0, "step": 14978 }, { "epoch": 1.9054827630072508, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5825443267822266, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8743035197257996, "num_tokens": 568604736.0, "step": 14979 }, { "epoch": 1.9056099732858414, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.352216958999634, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.884806752204895, "num_tokens": 568643720.0, "step": 14980 }, { "epoch": 1.905737183564432, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5961883068084717, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8820273876190186, "num_tokens": 568681805.0, "step": 14981 }, { "epoch": 1.9058643938430224, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2759592533111572, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8649474382400513, "num_tokens": 568723107.0, "step": 14982 }, { "epoch": 1.905991604121613, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6770834922790527, "learning_rate": 1e-05, "loss": 0.5186, "mean_token_accuracy": 0.8523565530776978, "num_tokens": 568760229.0, "step": 14983 }, { "epoch": 1.9061188144002035, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.2301526069641113, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.859321117401123, "num_tokens": 568794539.0, "step": 14984 }, { "epoch": 1.906246024678794, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7789013385772705, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8649013042449951, "num_tokens": 568834808.0, "step": 14985 }, { "epoch": 1.9063732349573845, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1697580814361572, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8704288005828857, "num_tokens": 568875004.0, "step": 14986 }, { "epoch": 1.906500445235975, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9449210166931152, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8740134239196777, "num_tokens": 568917372.0, "step": 14987 }, { "epoch": 1.9066276555145656, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.183401346206665, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8643280267715454, "num_tokens": 568953567.0, "step": 14988 }, { "epoch": 1.9067548657931561, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.157748699188232, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8603557348251343, "num_tokens": 568988634.0, "step": 14989 }, { "epoch": 1.9068820760717466, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.317432165145874, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8757175803184509, "num_tokens": 569023929.0, "step": 14990 }, { "epoch": 1.9070092863503372, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.26487398147583, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.878884494304657, "num_tokens": 569054483.0, "step": 14991 }, { "epoch": 1.9071364966289277, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5511646270751953, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8611239790916443, "num_tokens": 569096919.0, "step": 14992 }, { "epoch": 1.9072637069075182, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1767401695251465, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8669981956481934, "num_tokens": 569134161.0, "step": 14993 }, { "epoch": 1.9073909171861088, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6801462173461914, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8710254430770874, "num_tokens": 569173391.0, "step": 14994 }, { "epoch": 1.9075181274646993, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.013251781463623, "learning_rate": 1e-05, "loss": 0.5279, "mean_token_accuracy": 0.8518435955047607, "num_tokens": 569206173.0, "step": 14995 }, { "epoch": 1.9076453377432898, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6883678436279297, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8639155626296997, "num_tokens": 569241692.0, "step": 14996 }, { "epoch": 1.9077725480218801, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.803023338317871, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8796181678771973, "num_tokens": 569276361.0, "step": 14997 }, { "epoch": 1.9078997583004706, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.64451003074646, "learning_rate": 1e-05, "loss": 0.5339, "mean_token_accuracy": 0.844352662563324, "num_tokens": 569314478.0, "step": 14998 }, { "epoch": 1.9080269685790612, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.642876386642456, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8734911680221558, "num_tokens": 569354867.0, "step": 14999 }, { "epoch": 1.9081541788576517, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5555756092071533, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8772984743118286, "num_tokens": 569394031.0, "step": 15000 }, { "epoch": 1.9082813891362422, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5377492904663086, "learning_rate": 1e-05, "loss": 0.3754, "mean_token_accuracy": 0.8962262272834778, "num_tokens": 569430899.0, "step": 15001 }, { "epoch": 1.9084085994148328, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4973666667938232, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8505605459213257, "num_tokens": 569469871.0, "step": 15002 }, { "epoch": 1.908535809693423, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4945733547210693, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8695492744445801, "num_tokens": 569508385.0, "step": 15003 }, { "epoch": 1.9086630199720136, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4319634437561035, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.872268795967102, "num_tokens": 569549705.0, "step": 15004 }, { "epoch": 1.9087902302506041, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.498781442642212, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8598315715789795, "num_tokens": 569584092.0, "step": 15005 }, { "epoch": 1.9089174405291947, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3839571475982666, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.869910478591919, "num_tokens": 569619683.0, "step": 15006 }, { "epoch": 1.9090446508077852, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5034162998199463, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8780566453933716, "num_tokens": 569656347.0, "step": 15007 }, { "epoch": 1.9091718610863757, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3266360759735107, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8768240809440613, "num_tokens": 569689633.0, "step": 15008 }, { "epoch": 1.9092990713649662, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5496301651000977, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8586413860321045, "num_tokens": 569730685.0, "step": 15009 }, { "epoch": 1.9094262816435568, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2560949325561523, "learning_rate": 1e-05, "loss": 0.4891, "mean_token_accuracy": 0.8588650226593018, "num_tokens": 569773343.0, "step": 15010 }, { "epoch": 1.9095534919221473, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.642806053161621, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8697078227996826, "num_tokens": 569814417.0, "step": 15011 }, { "epoch": 1.9096807022007378, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.173166513442993, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8773060441017151, "num_tokens": 569854436.0, "step": 15012 }, { "epoch": 1.9098079124793284, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.816368341445923, "learning_rate": 1e-05, "loss": 0.5088, "mean_token_accuracy": 0.8541391491889954, "num_tokens": 569897530.0, "step": 15013 }, { "epoch": 1.9099351227579189, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.125349521636963, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8747486472129822, "num_tokens": 569931927.0, "step": 15014 }, { "epoch": 1.9100623330365094, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.062654495239258, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8728052973747253, "num_tokens": 569968283.0, "step": 15015 }, { "epoch": 1.9101895433151, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.185121536254883, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8672193288803101, "num_tokens": 570009073.0, "step": 15016 }, { "epoch": 1.9103167535936905, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.303991317749023, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8701466917991638, "num_tokens": 570046421.0, "step": 15017 }, { "epoch": 1.910443963872281, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4146530628204346, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8652594089508057, "num_tokens": 570093679.0, "step": 15018 }, { "epoch": 1.9105711741508715, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.370275020599365, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8788161277770996, "num_tokens": 570132575.0, "step": 15019 }, { "epoch": 1.910698384429462, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.725543737411499, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8679509162902832, "num_tokens": 570169782.0, "step": 15020 }, { "epoch": 1.9108255947080524, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.187671184539795, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8752182126045227, "num_tokens": 570205034.0, "step": 15021 }, { "epoch": 1.9109528049866429, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.794560670852661, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8717054128646851, "num_tokens": 570236804.0, "step": 15022 }, { "epoch": 1.9110800152652334, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9641036987304688, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8716849088668823, "num_tokens": 570266893.0, "step": 15023 }, { "epoch": 1.911207225543824, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.719581127166748, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8805525898933411, "num_tokens": 570297779.0, "step": 15024 }, { "epoch": 1.9113344358224145, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7716338634490967, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8751029968261719, "num_tokens": 570334774.0, "step": 15025 }, { "epoch": 1.911461646101005, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6129002571105957, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.864896833896637, "num_tokens": 570373788.0, "step": 15026 }, { "epoch": 1.9115888563795955, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6648309230804443, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8751041889190674, "num_tokens": 570412156.0, "step": 15027 }, { "epoch": 1.9117160666581858, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.523106813430786, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8757165670394897, "num_tokens": 570442702.0, "step": 15028 }, { "epoch": 1.9118432769367764, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.604809045791626, "learning_rate": 1e-05, "loss": 0.5014, "mean_token_accuracy": 0.8584468364715576, "num_tokens": 570485302.0, "step": 15029 }, { "epoch": 1.9119704872153669, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.422231674194336, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8747732043266296, "num_tokens": 570524820.0, "step": 15030 }, { "epoch": 1.9120976974939574, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.608283758163452, "learning_rate": 1e-05, "loss": 0.4961, "mean_token_accuracy": 0.8585975766181946, "num_tokens": 570567973.0, "step": 15031 }, { "epoch": 1.912224907772548, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.334930658340454, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8760783672332764, "num_tokens": 570610369.0, "step": 15032 }, { "epoch": 1.9123521180511385, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6515941619873047, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8655874729156494, "num_tokens": 570653226.0, "step": 15033 }, { "epoch": 1.912479328329729, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.257510185241699, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8662166595458984, "num_tokens": 570691505.0, "step": 15034 }, { "epoch": 1.9126065386083195, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7663333415985107, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8625833988189697, "num_tokens": 570733781.0, "step": 15035 }, { "epoch": 1.91273374888691, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.1941277980804443, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8753703236579895, "num_tokens": 570772675.0, "step": 15036 }, { "epoch": 1.9128609591655006, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.95190691947937, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8838032484054565, "num_tokens": 570813967.0, "step": 15037 }, { "epoch": 1.9129881694440911, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.201586961746216, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8798316717147827, "num_tokens": 570850315.0, "step": 15038 }, { "epoch": 1.9131153797226816, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.146332740783691, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8642653822898865, "num_tokens": 570890191.0, "step": 15039 }, { "epoch": 1.9132425900012722, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3600382804870605, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8739780783653259, "num_tokens": 570926564.0, "step": 15040 }, { "epoch": 1.9133698002798627, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.20986795425415, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8728055357933044, "num_tokens": 570966334.0, "step": 15041 }, { "epoch": 1.9134970105584532, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.581451654434204, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8646777272224426, "num_tokens": 571008008.0, "step": 15042 }, { "epoch": 1.9136242208370438, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.08435583114624, "learning_rate": 1e-05, "loss": 0.5068, "mean_token_accuracy": 0.8599998950958252, "num_tokens": 571046915.0, "step": 15043 }, { "epoch": 1.9137514311156343, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6697096824645996, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8704560995101929, "num_tokens": 571085945.0, "step": 15044 }, { "epoch": 1.9138786413942248, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.868757724761963, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8846551775932312, "num_tokens": 571119839.0, "step": 15045 }, { "epoch": 1.9140058516728151, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6551389694213867, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.881149172782898, "num_tokens": 571156888.0, "step": 15046 }, { "epoch": 1.9141330619514056, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.658799409866333, "learning_rate": 1e-05, "loss": 0.3975, "mean_token_accuracy": 0.890283465385437, "num_tokens": 571197022.0, "step": 15047 }, { "epoch": 1.9142602722299962, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.615241527557373, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8635458946228027, "num_tokens": 571234780.0, "step": 15048 }, { "epoch": 1.9143874825085867, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5196242332458496, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8744251132011414, "num_tokens": 571272804.0, "step": 15049 }, { "epoch": 1.9145146927871772, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.572049617767334, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8683217763900757, "num_tokens": 571315455.0, "step": 15050 }, { "epoch": 1.9146419030657678, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4310286045074463, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8770776391029358, "num_tokens": 571350522.0, "step": 15051 }, { "epoch": 1.914769113344358, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.549508810043335, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8771109580993652, "num_tokens": 571384452.0, "step": 15052 }, { "epoch": 1.9148963236229486, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.355783462524414, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8880070447921753, "num_tokens": 571423041.0, "step": 15053 }, { "epoch": 1.9150235339015391, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5722506046295166, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8853327035903931, "num_tokens": 571454744.0, "step": 15054 }, { "epoch": 1.9151507441801296, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2969281673431396, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8812888860702515, "num_tokens": 571492937.0, "step": 15055 }, { "epoch": 1.9152779544587202, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.61586856842041, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8855153322219849, "num_tokens": 571531756.0, "step": 15056 }, { "epoch": 1.9154051647373107, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.23928165435791, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8616685271263123, "num_tokens": 571566095.0, "step": 15057 }, { "epoch": 1.9155323750159012, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.722587823867798, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8745743632316589, "num_tokens": 571599169.0, "step": 15058 }, { "epoch": 1.9156595852944918, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1755948066711426, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8797703385353088, "num_tokens": 571632063.0, "step": 15059 }, { "epoch": 1.9157867955730823, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.878532886505127, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8739326000213623, "num_tokens": 571670867.0, "step": 15060 }, { "epoch": 1.9159140058516728, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.133761167526245, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8714307546615601, "num_tokens": 571710445.0, "step": 15061 }, { "epoch": 1.9160412161302633, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.109562873840332, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8892300128936768, "num_tokens": 571748410.0, "step": 15062 }, { "epoch": 1.9161684264088539, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2387335300445557, "learning_rate": 1e-05, "loss": 0.4952, "mean_token_accuracy": 0.8573489189147949, "num_tokens": 571786386.0, "step": 15063 }, { "epoch": 1.9162956366874444, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.302423000335693, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8648300766944885, "num_tokens": 571824414.0, "step": 15064 }, { "epoch": 1.916422846966035, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5040316581726074, "learning_rate": 1e-05, "loss": 0.5223, "mean_token_accuracy": 0.850195050239563, "num_tokens": 571863085.0, "step": 15065 }, { "epoch": 1.9165500572446255, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.277356147766113, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8689748048782349, "num_tokens": 571894749.0, "step": 15066 }, { "epoch": 1.916677267523216, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.71726131439209, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8610368967056274, "num_tokens": 571925962.0, "step": 15067 }, { "epoch": 1.9168044778018065, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0774970054626465, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8616675138473511, "num_tokens": 571966021.0, "step": 15068 }, { "epoch": 1.916931688080397, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.707455635070801, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8814268112182617, "num_tokens": 572002684.0, "step": 15069 }, { "epoch": 1.9170588983589874, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.893256664276123, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8727936744689941, "num_tokens": 572038054.0, "step": 15070 }, { "epoch": 1.9171861086375779, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.645294666290283, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8677771687507629, "num_tokens": 572076851.0, "step": 15071 }, { "epoch": 1.9173133189161684, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.739377975463867, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8758552074432373, "num_tokens": 572115642.0, "step": 15072 }, { "epoch": 1.917440529194759, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5562281608581543, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8801718950271606, "num_tokens": 572154803.0, "step": 15073 }, { "epoch": 1.9175677394733495, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6527013778686523, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8666363954544067, "num_tokens": 572189241.0, "step": 15074 }, { "epoch": 1.91769494975194, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002727508544921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.467693328857422, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8852605819702148, "num_tokens": 572230464.0, "step": 15075 }, { "epoch": 1.9178221600305305, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.622131109237671, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8818486928939819, "num_tokens": 572268863.0, "step": 15076 }, { "epoch": 1.9179493703091208, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.393343925476074, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8675108551979065, "num_tokens": 572307484.0, "step": 15077 }, { "epoch": 1.9180765805877114, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.628830909729004, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8698663711547852, "num_tokens": 572347653.0, "step": 15078 }, { "epoch": 1.9182037908663019, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.321560859680176, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8588533401489258, "num_tokens": 572386850.0, "step": 15079 }, { "epoch": 1.9183310011448924, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.703113079071045, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8616730570793152, "num_tokens": 572422416.0, "step": 15080 }, { "epoch": 1.918458211423483, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.252453327178955, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8601537346839905, "num_tokens": 572458845.0, "step": 15081 }, { "epoch": 1.9185854217020735, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.809293031692505, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8742938041687012, "num_tokens": 572500553.0, "step": 15082 }, { "epoch": 1.918712631980664, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2051970958709717, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8639536499977112, "num_tokens": 572541315.0, "step": 15083 }, { "epoch": 1.9188398422592545, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9843361377716064, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.878607988357544, "num_tokens": 572574461.0, "step": 15084 }, { "epoch": 1.918967052537845, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2478854656219482, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8634196519851685, "num_tokens": 572616860.0, "step": 15085 }, { "epoch": 1.9190942628164356, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1362433433532715, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8761510252952576, "num_tokens": 572654228.0, "step": 15086 }, { "epoch": 1.919221473095026, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.535128116607666, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8700631856918335, "num_tokens": 572693603.0, "step": 15087 }, { "epoch": 1.9193486833736166, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.029775619506836, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8627814650535583, "num_tokens": 572730292.0, "step": 15088 }, { "epoch": 1.9194758936522072, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.528707981109619, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8659306764602661, "num_tokens": 572764348.0, "step": 15089 }, { "epoch": 1.9196031039307977, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.891683578491211, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8781538605690002, "num_tokens": 572798410.0, "step": 15090 }, { "epoch": 1.9197303142093882, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5624663829803467, "learning_rate": 1e-05, "loss": 0.5137, "mean_token_accuracy": 0.8585768342018127, "num_tokens": 572832067.0, "step": 15091 }, { "epoch": 1.9198575244879788, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7315409183502197, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.875502347946167, "num_tokens": 572866676.0, "step": 15092 }, { "epoch": 1.9199847347665693, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5360851287841797, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8612821102142334, "num_tokens": 572913631.0, "step": 15093 }, { "epoch": 1.9201119450451598, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.586300849914551, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8758893013000488, "num_tokens": 572953154.0, "step": 15094 }, { "epoch": 1.9202391553237501, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5197250843048096, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8643932938575745, "num_tokens": 572992259.0, "step": 15095 }, { "epoch": 1.9203663656023406, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.497495412826538, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8723838329315186, "num_tokens": 573024464.0, "step": 15096 }, { "epoch": 1.9204935758809312, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4960391521453857, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8766381740570068, "num_tokens": 573060620.0, "step": 15097 }, { "epoch": 1.9206207861595217, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.425657033920288, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8661102056503296, "num_tokens": 573098284.0, "step": 15098 }, { "epoch": 1.9207479964381122, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4954447746276855, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8715540170669556, "num_tokens": 573136253.0, "step": 15099 }, { "epoch": 1.9208752067167028, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3707361221313477, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8706233501434326, "num_tokens": 573174843.0, "step": 15100 }, { "epoch": 1.921002416995293, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5127813816070557, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8588821291923523, "num_tokens": 573215759.0, "step": 15101 }, { "epoch": 1.9211296272738836, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3146796226501465, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8730320334434509, "num_tokens": 573255297.0, "step": 15102 }, { "epoch": 1.9212568375524741, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.546571969985962, "learning_rate": 1e-05, "loss": 0.3696, "mean_token_accuracy": 0.8986234664916992, "num_tokens": 573293152.0, "step": 15103 }, { "epoch": 1.9213840478310646, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.25665020942688, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8700392246246338, "num_tokens": 573330456.0, "step": 15104 }, { "epoch": 1.9215112581096552, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6342997550964355, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8633944988250732, "num_tokens": 573371595.0, "step": 15105 }, { "epoch": 1.9216384683882457, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.200169801712036, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8668068051338196, "num_tokens": 573417414.0, "step": 15106 }, { "epoch": 1.9217656786668362, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.770798683166504, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8649097084999084, "num_tokens": 573452146.0, "step": 15107 }, { "epoch": 1.9218928889454268, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1253676414489746, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8641045093536377, "num_tokens": 573485469.0, "step": 15108 }, { "epoch": 1.9220200992240173, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9775383472442627, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8752851486206055, "num_tokens": 573524763.0, "step": 15109 }, { "epoch": 1.9221473095026078, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.197085380554199, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.865470826625824, "num_tokens": 573559326.0, "step": 15110 }, { "epoch": 1.9222745197811983, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.15419864654541, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8662692904472351, "num_tokens": 573604700.0, "step": 15111 }, { "epoch": 1.9224017300597889, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.247818946838379, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8839451670646667, "num_tokens": 573639570.0, "step": 15112 }, { "epoch": 1.9225289403383794, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.370031833648682, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8669112324714661, "num_tokens": 573669849.0, "step": 15113 }, { "epoch": 1.92265615061697, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.657484292984009, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8671690225601196, "num_tokens": 573702402.0, "step": 15114 }, { "epoch": 1.9227833608955605, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.213167667388916, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8675408363342285, "num_tokens": 573737428.0, "step": 15115 }, { "epoch": 1.922910571174151, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.676380157470703, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8767861127853394, "num_tokens": 573775111.0, "step": 15116 }, { "epoch": 1.9230377814527415, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.05181884765625, "learning_rate": 1e-05, "loss": 0.4891, "mean_token_accuracy": 0.8612724542617798, "num_tokens": 573814550.0, "step": 15117 }, { "epoch": 1.923164991731332, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.688357353210449, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8596721291542053, "num_tokens": 573852932.0, "step": 15118 }, { "epoch": 1.9232922020099223, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.871544122695923, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8782470226287842, "num_tokens": 573887769.0, "step": 15119 }, { "epoch": 1.9234194122885129, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6190171241760254, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8650444746017456, "num_tokens": 573925258.0, "step": 15120 }, { "epoch": 1.9235466225671034, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7426748275756836, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8817468285560608, "num_tokens": 573956690.0, "step": 15121 }, { "epoch": 1.923673832845694, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.532994508743286, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8582073450088501, "num_tokens": 573993822.0, "step": 15122 }, { "epoch": 1.9238010431242845, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.670764684677124, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8682769536972046, "num_tokens": 574030087.0, "step": 15123 }, { "epoch": 1.923928253402875, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.446725368499756, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8779732584953308, "num_tokens": 574071964.0, "step": 15124 }, { "epoch": 1.9240554636814655, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6538875102996826, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8662207722663879, "num_tokens": 574105706.0, "step": 15125 }, { "epoch": 1.9241826739600558, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.36315655708313, "learning_rate": 1e-05, "loss": 0.5051, "mean_token_accuracy": 0.8538649082183838, "num_tokens": 574145798.0, "step": 15126 }, { "epoch": 1.9243098842386464, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6729142665863037, "learning_rate": 1e-05, "loss": 0.4084, "mean_token_accuracy": 0.8857563734054565, "num_tokens": 574192631.0, "step": 15127 }, { "epoch": 1.9244370945172369, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3050220012664795, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8518412709236145, "num_tokens": 574228601.0, "step": 15128 }, { "epoch": 1.9245643047958274, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7523112297058105, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8772981762886047, "num_tokens": 574266416.0, "step": 15129 }, { "epoch": 1.924691515074418, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.255596160888672, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8600239157676697, "num_tokens": 574297414.0, "step": 15130 }, { "epoch": 1.9248187253530085, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.859076976776123, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8878262042999268, "num_tokens": 574333749.0, "step": 15131 }, { "epoch": 1.924945935631599, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2283143997192383, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8697113394737244, "num_tokens": 574377186.0, "step": 15132 }, { "epoch": 1.9250731459101895, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.009219646453857, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8754433989524841, "num_tokens": 574414814.0, "step": 15133 }, { "epoch": 1.92520035618878, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3137576580047607, "learning_rate": 1e-05, "loss": 0.4701, "mean_token_accuracy": 0.8635976314544678, "num_tokens": 574453070.0, "step": 15134 }, { "epoch": 1.9253275664673706, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.095279693603516, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8775467872619629, "num_tokens": 574487564.0, "step": 15135 }, { "epoch": 1.925454776745961, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4676313400268555, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8680322766304016, "num_tokens": 574528456.0, "step": 15136 }, { "epoch": 1.9255819870245516, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0218281745910645, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8726679682731628, "num_tokens": 574567724.0, "step": 15137 }, { "epoch": 1.9257091973031422, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.587723970413208, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8711349964141846, "num_tokens": 574599702.0, "step": 15138 }, { "epoch": 1.9258364075817327, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8293778896331787, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.876898467540741, "num_tokens": 574633709.0, "step": 15139 }, { "epoch": 1.9259636178603232, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6135377883911133, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8693034648895264, "num_tokens": 574670490.0, "step": 15140 }, { "epoch": 1.9260908281389137, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.627189874649048, "learning_rate": 1e-05, "loss": 0.4954, "mean_token_accuracy": 0.8574119210243225, "num_tokens": 574706563.0, "step": 15141 }, { "epoch": 1.9262180384175043, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6148195266723633, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.874207615852356, "num_tokens": 574744743.0, "step": 15142 }, { "epoch": 1.9263452486960948, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.445181369781494, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8692736625671387, "num_tokens": 574782236.0, "step": 15143 }, { "epoch": 1.926472458974685, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6214747428894043, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8827470541000366, "num_tokens": 574826457.0, "step": 15144 }, { "epoch": 1.9265996692532756, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3233084678649902, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8624539971351624, "num_tokens": 574870362.0, "step": 15145 }, { "epoch": 1.9267268795318662, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.673938035964966, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8765305280685425, "num_tokens": 574907984.0, "step": 15146 }, { "epoch": 1.9268540898104567, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.245048761367798, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8636171817779541, "num_tokens": 574950813.0, "step": 15147 }, { "epoch": 1.9269813000890472, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7769362926483154, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8753872513771057, "num_tokens": 574985210.0, "step": 15148 }, { "epoch": 1.9271085103676378, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1956214904785156, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8782989382743835, "num_tokens": 575020263.0, "step": 15149 }, { "epoch": 1.927235720646228, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9205946922302246, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8683567047119141, "num_tokens": 575060632.0, "step": 15150 }, { "epoch": 1.9273629309248186, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.182129383087158, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.867682933807373, "num_tokens": 575102725.0, "step": 15151 }, { "epoch": 1.9274901412034091, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.100250720977783, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8853443264961243, "num_tokens": 575143069.0, "step": 15152 }, { "epoch": 1.9276173514819996, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2889697551727295, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8734701871871948, "num_tokens": 575182618.0, "step": 15153 }, { "epoch": 1.9277445617605902, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2326812744140625, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8700328469276428, "num_tokens": 575224576.0, "step": 15154 }, { "epoch": 1.9278717720391807, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.505157470703125, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8818120956420898, "num_tokens": 575260152.0, "step": 15155 }, { "epoch": 1.9279989823177712, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.18862247467041, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8682534694671631, "num_tokens": 575292308.0, "step": 15156 }, { "epoch": 1.9281261925963618, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6560287475585938, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8662166595458984, "num_tokens": 575326305.0, "step": 15157 }, { "epoch": 1.9282534028749523, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0151047706604, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8666403293609619, "num_tokens": 575361775.0, "step": 15158 }, { "epoch": 1.9283806131535428, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6648502349853516, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8797706961631775, "num_tokens": 575397508.0, "step": 15159 }, { "epoch": 1.9285078234321333, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8200297355651855, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8683773279190063, "num_tokens": 575437770.0, "step": 15160 }, { "epoch": 1.9286350337107239, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6242425441741943, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8755033016204834, "num_tokens": 575478322.0, "step": 15161 }, { "epoch": 1.9287622439893144, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6694583892822266, "learning_rate": 1e-05, "loss": 0.5023, "mean_token_accuracy": 0.8587920665740967, "num_tokens": 575518816.0, "step": 15162 }, { "epoch": 1.928889454267905, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.551518201828003, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8624719381332397, "num_tokens": 575554276.0, "step": 15163 }, { "epoch": 1.9290166645464955, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.582871198654175, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8655412197113037, "num_tokens": 575591331.0, "step": 15164 }, { "epoch": 1.929143874825086, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4790356159210205, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8688375949859619, "num_tokens": 575629449.0, "step": 15165 }, { "epoch": 1.9292710851036765, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5388875007629395, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8575471639633179, "num_tokens": 575668169.0, "step": 15166 }, { "epoch": 1.929398295382267, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4902408123016357, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8606809377670288, "num_tokens": 575708481.0, "step": 15167 }, { "epoch": 1.9295255056608573, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.497911214828491, "learning_rate": 1e-05, "loss": 0.5167, "mean_token_accuracy": 0.849785566329956, "num_tokens": 575744879.0, "step": 15168 }, { "epoch": 1.9296527159394479, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3943030834198, "learning_rate": 1e-05, "loss": 0.3866, "mean_token_accuracy": 0.891829252243042, "num_tokens": 575781587.0, "step": 15169 }, { "epoch": 1.9297799262180384, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4992921352386475, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8690191507339478, "num_tokens": 575816662.0, "step": 15170 }, { "epoch": 1.929907136496629, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3548858165740967, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8660507202148438, "num_tokens": 575855169.0, "step": 15171 }, { "epoch": 1.9300343467752195, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.51713490486145, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8825781941413879, "num_tokens": 575896939.0, "step": 15172 }, { "epoch": 1.93016155705381, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3063735961914062, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8750820159912109, "num_tokens": 575930587.0, "step": 15173 }, { "epoch": 1.9302887673324005, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5870447158813477, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8563340902328491, "num_tokens": 575964432.0, "step": 15174 }, { "epoch": 1.9304159776109908, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2308037281036377, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8734755516052246, "num_tokens": 576004916.0, "step": 15175 }, { "epoch": 1.9305431878895813, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6996331214904785, "learning_rate": 1e-05, "loss": 0.5221, "mean_token_accuracy": 0.8507449626922607, "num_tokens": 576047857.0, "step": 15176 }, { "epoch": 1.9306703981681719, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.15956449508667, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8616365790367126, "num_tokens": 576089087.0, "step": 15177 }, { "epoch": 1.9307976084467624, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8732335567474365, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8736382126808167, "num_tokens": 576134106.0, "step": 15178 }, { "epoch": 1.930924818725353, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1147806644439697, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8669992685317993, "num_tokens": 576172566.0, "step": 15179 }, { "epoch": 1.9310520290039435, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.144935607910156, "learning_rate": 1e-05, "loss": 0.559, "mean_token_accuracy": 0.8458261489868164, "num_tokens": 576214469.0, "step": 15180 }, { "epoch": 1.931179239282534, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2253692150115967, "learning_rate": 1e-05, "loss": 0.3841, "mean_token_accuracy": 0.8926684856414795, "num_tokens": 576249316.0, "step": 15181 }, { "epoch": 1.9313064495611245, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.372249603271484, "learning_rate": 1e-05, "loss": 0.5299, "mean_token_accuracy": 0.8502540588378906, "num_tokens": 576290748.0, "step": 15182 }, { "epoch": 1.931433659839715, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5482919216156006, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.872115969657898, "num_tokens": 576329137.0, "step": 15183 }, { "epoch": 1.9315608701183056, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.325797080993652, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8834707736968994, "num_tokens": 576367147.0, "step": 15184 }, { "epoch": 1.931688080396896, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.7615833282470703, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.8614290356636047, "num_tokens": 576406068.0, "step": 15185 }, { "epoch": 1.9318152906754866, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.132540702819824, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8652685880661011, "num_tokens": 576438150.0, "step": 15186 }, { "epoch": 1.9319425009540772, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.768616199493408, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.874582052230835, "num_tokens": 576470630.0, "step": 15187 }, { "epoch": 1.9320697112326677, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9212424755096436, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8770050406455994, "num_tokens": 576505896.0, "step": 15188 }, { "epoch": 1.9321969215112582, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6694679260253906, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8799343109130859, "num_tokens": 576545204.0, "step": 15189 }, { "epoch": 1.9323241317898487, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7802064418792725, "learning_rate": 1e-05, "loss": 0.5216, "mean_token_accuracy": 0.8597378730773926, "num_tokens": 576590746.0, "step": 15190 }, { "epoch": 1.9324513420684393, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5645477771759033, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8691667318344116, "num_tokens": 576631185.0, "step": 15191 }, { "epoch": 1.9325785523470298, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6989800930023193, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8692905902862549, "num_tokens": 576668945.0, "step": 15192 }, { "epoch": 1.93270576262562, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4656240940093994, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8714330792427063, "num_tokens": 576706052.0, "step": 15193 }, { "epoch": 1.9328329729042106, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.684438467025757, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8708978295326233, "num_tokens": 576740668.0, "step": 15194 }, { "epoch": 1.9329601831828012, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3650002479553223, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8611570596694946, "num_tokens": 576780312.0, "step": 15195 }, { "epoch": 1.9330873934613917, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.723154067993164, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8821648955345154, "num_tokens": 576817130.0, "step": 15196 }, { "epoch": 1.9332146037399822, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2872233390808105, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.8861693143844604, "num_tokens": 576849823.0, "step": 15197 }, { "epoch": 1.9333418140185727, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.829026222229004, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8625940084457397, "num_tokens": 576890890.0, "step": 15198 }, { "epoch": 1.933469024297163, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2477669715881348, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8558872938156128, "num_tokens": 576927926.0, "step": 15199 }, { "epoch": 1.9335962345757536, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9793813228607178, "learning_rate": 1e-05, "loss": 0.4996, "mean_token_accuracy": 0.8591612577438354, "num_tokens": 576966194.0, "step": 15200 }, { "epoch": 1.933723444854344, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2898218631744385, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.8592509627342224, "num_tokens": 577007218.0, "step": 15201 }, { "epoch": 1.9338506551329346, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.097306728363037, "learning_rate": 1e-05, "loss": 0.4102, "mean_token_accuracy": 0.8893820643424988, "num_tokens": 577040551.0, "step": 15202 }, { "epoch": 1.9339778654115252, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.422840118408203, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8715888857841492, "num_tokens": 577082676.0, "step": 15203 }, { "epoch": 1.9341050756901157, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.084787845611572, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8774381279945374, "num_tokens": 577124572.0, "step": 15204 }, { "epoch": 1.9342322859687062, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.559340476989746, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8836349248886108, "num_tokens": 577167571.0, "step": 15205 }, { "epoch": 1.9343594962472968, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9345171451568604, "learning_rate": 1e-05, "loss": 0.5159, "mean_token_accuracy": 0.8517199754714966, "num_tokens": 577211976.0, "step": 15206 }, { "epoch": 1.9344867065258873, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.622755289077759, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.876888632774353, "num_tokens": 577252572.0, "step": 15207 }, { "epoch": 1.9346139168044778, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.740396499633789, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8572096824645996, "num_tokens": 577286923.0, "step": 15208 }, { "epoch": 1.9347411270830683, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6110002994537354, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8673132061958313, "num_tokens": 577324520.0, "step": 15209 }, { "epoch": 1.9348683373616589, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.54434871673584, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8668040633201599, "num_tokens": 577361100.0, "step": 15210 }, { "epoch": 1.9349955476402494, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.598010540008545, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8685073256492615, "num_tokens": 577402714.0, "step": 15211 }, { "epoch": 1.93512275791884, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.4204251766204834, "learning_rate": 1e-05, "loss": 0.5013, "mean_token_accuracy": 0.8592731952667236, "num_tokens": 577437226.0, "step": 15212 }, { "epoch": 1.9352499681974304, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.601926803588867, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.873650312423706, "num_tokens": 577475564.0, "step": 15213 }, { "epoch": 1.935377178476021, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3362584114074707, "learning_rate": 1e-05, "loss": 0.5243, "mean_token_accuracy": 0.8509615659713745, "num_tokens": 577512623.0, "step": 15214 }, { "epoch": 1.9355043887546115, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6411292552948, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8700542449951172, "num_tokens": 577545515.0, "step": 15215 }, { "epoch": 1.935631599033202, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2570266723632812, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8713699579238892, "num_tokens": 577577043.0, "step": 15216 }, { "epoch": 1.9357588093117923, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.711207151412964, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8777351379394531, "num_tokens": 577616311.0, "step": 15217 }, { "epoch": 1.9358860195903829, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.187282085418701, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8741971254348755, "num_tokens": 577654963.0, "step": 15218 }, { "epoch": 1.9360132298689734, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.854527711868286, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8812587261199951, "num_tokens": 577688283.0, "step": 15219 }, { "epoch": 1.936140440147564, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1704890727996826, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.862899661064148, "num_tokens": 577726855.0, "step": 15220 }, { "epoch": 1.9362676504261545, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.04888916015625, "learning_rate": 1e-05, "loss": 0.5182, "mean_token_accuracy": 0.8587688207626343, "num_tokens": 577765699.0, "step": 15221 }, { "epoch": 1.936394860704745, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.238654851913452, "learning_rate": 1e-05, "loss": 0.3845, "mean_token_accuracy": 0.8935050964355469, "num_tokens": 577797393.0, "step": 15222 }, { "epoch": 1.9365220709833355, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2114434242248535, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8794721364974976, "num_tokens": 577838768.0, "step": 15223 }, { "epoch": 1.9366492812619258, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.421509027481079, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8732411861419678, "num_tokens": 577882455.0, "step": 15224 }, { "epoch": 1.9367764915405163, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2303361892700195, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8852258920669556, "num_tokens": 577922649.0, "step": 15225 }, { "epoch": 1.9369037018191069, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6248843669891357, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8720260262489319, "num_tokens": 577961679.0, "step": 15226 }, { "epoch": 1.9370309120976974, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.083248138427734, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8761146664619446, "num_tokens": 578001848.0, "step": 15227 }, { "epoch": 1.937158122376288, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6842305660247803, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.874169111251831, "num_tokens": 578039278.0, "step": 15228 }, { "epoch": 1.9372853326548785, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.887895345687866, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8694174289703369, "num_tokens": 578078152.0, "step": 15229 }, { "epoch": 1.937412542933469, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.641075372695923, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8645931482315063, "num_tokens": 578115927.0, "step": 15230 }, { "epoch": 1.9375397532120595, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.716959238052368, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8747819066047668, "num_tokens": 578157425.0, "step": 15231 }, { "epoch": 1.93766696349065, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5718629360198975, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.87701016664505, "num_tokens": 578201985.0, "step": 15232 }, { "epoch": 1.9377941737692406, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6094446182250977, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8693047761917114, "num_tokens": 578239471.0, "step": 15233 }, { "epoch": 1.937921384047831, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.510852813720703, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8795756101608276, "num_tokens": 578272302.0, "step": 15234 }, { "epoch": 1.9380485943264216, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5380022525787354, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8692764043807983, "num_tokens": 578314585.0, "step": 15235 }, { "epoch": 1.9381758046050122, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4501147270202637, "learning_rate": 1e-05, "loss": 0.5095, "mean_token_accuracy": 0.8507344722747803, "num_tokens": 578355561.0, "step": 15236 }, { "epoch": 1.9383030148836027, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5245742797851562, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8669523596763611, "num_tokens": 578398903.0, "step": 15237 }, { "epoch": 1.9384302251621932, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.391404151916504, "learning_rate": 1e-05, "loss": 0.5384, "mean_token_accuracy": 0.845140814781189, "num_tokens": 578439293.0, "step": 15238 }, { "epoch": 1.9385574354407837, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.541449546813965, "learning_rate": 1e-05, "loss": 0.4976, "mean_token_accuracy": 0.8539263010025024, "num_tokens": 578477965.0, "step": 15239 }, { "epoch": 1.9386846457193743, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.3220407962799072, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8756742477416992, "num_tokens": 578515577.0, "step": 15240 }, { "epoch": 1.9388118559979648, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5925984382629395, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8715620040893555, "num_tokens": 578557415.0, "step": 15241 }, { "epoch": 1.938939066276555, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.2493929862976074, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8763087391853333, "num_tokens": 578594556.0, "step": 15242 }, { "epoch": 1.9390662765551456, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7179970741271973, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8810997009277344, "num_tokens": 578627219.0, "step": 15243 }, { "epoch": 1.9391934868337362, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.279763698577881, "learning_rate": 1e-05, "loss": 0.4943, "mean_token_accuracy": 0.8563087582588196, "num_tokens": 578669135.0, "step": 15244 }, { "epoch": 1.9393206971123267, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.784583568572998, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8769724369049072, "num_tokens": 578706174.0, "step": 15245 }, { "epoch": 1.9394479073909172, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1678318977355957, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8683065176010132, "num_tokens": 578738942.0, "step": 15246 }, { "epoch": 1.9395751176695077, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.956021547317505, "learning_rate": 1e-05, "loss": 0.3942, "mean_token_accuracy": 0.892253041267395, "num_tokens": 578771553.0, "step": 15247 }, { "epoch": 1.939702327948098, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.1779346466064453, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8663270473480225, "num_tokens": 578811532.0, "step": 15248 }, { "epoch": 1.9398295382266886, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.163676738739014, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8685282468795776, "num_tokens": 578852858.0, "step": 15249 }, { "epoch": 1.939956748505279, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.342395067214966, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.867266058921814, "num_tokens": 578890340.0, "step": 15250 }, { "epoch": 1.9400839587838696, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.241196155548096, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8788983821868896, "num_tokens": 578928953.0, "step": 15251 }, { "epoch": 1.9402111690624602, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5682761669158936, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8808395862579346, "num_tokens": 578967978.0, "step": 15252 }, { "epoch": 1.9403383793410507, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.123264789581299, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8763014674186707, "num_tokens": 579005484.0, "step": 15253 }, { "epoch": 1.9404655896196412, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6880526542663574, "learning_rate": 1e-05, "loss": 0.5183, "mean_token_accuracy": 0.8537963628768921, "num_tokens": 579045172.0, "step": 15254 }, { "epoch": 1.9405927998982317, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.917696714401245, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8811293840408325, "num_tokens": 579075001.0, "step": 15255 }, { "epoch": 1.9407200101768223, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.658445119857788, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8703329563140869, "num_tokens": 579116029.0, "step": 15256 }, { "epoch": 1.9408472204554128, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7137985229492188, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8683693408966064, "num_tokens": 579157741.0, "step": 15257 }, { "epoch": 1.9409744307340033, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6075398921966553, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8715943694114685, "num_tokens": 579196033.0, "step": 15258 }, { "epoch": 1.9411016410125939, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5684173107147217, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.880281388759613, "num_tokens": 579235060.0, "step": 15259 }, { "epoch": 1.9412288512911844, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.559453010559082, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8727405667304993, "num_tokens": 579272184.0, "step": 15260 }, { "epoch": 1.941356061569775, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.482703924179077, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8724954724311829, "num_tokens": 579309461.0, "step": 15261 }, { "epoch": 1.9414832718483654, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.515226125717163, "learning_rate": 1e-05, "loss": 0.5017, "mean_token_accuracy": 0.8546313047409058, "num_tokens": 579351871.0, "step": 15262 }, { "epoch": 1.941610482126956, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.440976619720459, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.8558651804924011, "num_tokens": 579390123.0, "step": 15263 }, { "epoch": 1.9417376924055465, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4967153072357178, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8614822626113892, "num_tokens": 579426015.0, "step": 15264 }, { "epoch": 1.941864902684137, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3930444717407227, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.857155978679657, "num_tokens": 579465853.0, "step": 15265 }, { "epoch": 1.9419921129627273, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4799928665161133, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8667752742767334, "num_tokens": 579499740.0, "step": 15266 }, { "epoch": 1.9421193232413179, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.346416473388672, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8863176107406616, "num_tokens": 579538908.0, "step": 15267 }, { "epoch": 1.9422465335199084, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.499002456665039, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8728061318397522, "num_tokens": 579576026.0, "step": 15268 }, { "epoch": 1.942373743798499, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3009328842163086, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8764201402664185, "num_tokens": 579618980.0, "step": 15269 }, { "epoch": 1.9425009540770894, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.563253879547119, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8568177819252014, "num_tokens": 579659205.0, "step": 15270 }, { "epoch": 1.94262816435568, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.231775999069214, "learning_rate": 1e-05, "loss": 0.4028, "mean_token_accuracy": 0.8819212913513184, "num_tokens": 579695598.0, "step": 15271 }, { "epoch": 1.9427553746342705, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6755120754241943, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8717966079711914, "num_tokens": 579730871.0, "step": 15272 }, { "epoch": 1.9428825849128608, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1410586833953857, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8670659065246582, "num_tokens": 579772879.0, "step": 15273 }, { "epoch": 1.9430097951914513, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.863882064819336, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8663126230239868, "num_tokens": 579812854.0, "step": 15274 }, { "epoch": 1.9431370054700419, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.087554693222046, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8648182153701782, "num_tokens": 579857324.0, "step": 15275 }, { "epoch": 1.9432642157486324, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.157559871673584, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.871889591217041, "num_tokens": 579897103.0, "step": 15276 }, { "epoch": 1.943391426027223, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.1869728565216064, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8635071516036987, "num_tokens": 579937382.0, "step": 15277 }, { "epoch": 1.9435186363058135, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.446008205413818, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.87644362449646, "num_tokens": 579978591.0, "step": 15278 }, { "epoch": 1.943645846584404, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.550145149230957, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8751037120819092, "num_tokens": 580014354.0, "step": 15279 }, { "epoch": 1.9437730568629945, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.440397262573242, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8772610425949097, "num_tokens": 580055575.0, "step": 15280 }, { "epoch": 1.943900267141585, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.839268922805786, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8644843101501465, "num_tokens": 580105007.0, "step": 15281 }, { "epoch": 1.9440274774201756, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1960859298706055, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8619641661643982, "num_tokens": 580150820.0, "step": 15282 }, { "epoch": 1.944154687698766, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.845207691192627, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8812021017074585, "num_tokens": 580188303.0, "step": 15283 }, { "epoch": 1.9442818979773566, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9939210414886475, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8686292171478271, "num_tokens": 580221824.0, "step": 15284 }, { "epoch": 1.9444091082559471, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.732043504714966, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8676231503486633, "num_tokens": 580262615.0, "step": 15285 }, { "epoch": 1.9445363185345377, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.837214946746826, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8765957951545715, "num_tokens": 580293535.0, "step": 15286 }, { "epoch": 1.9446635288131282, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.614532470703125, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8700464963912964, "num_tokens": 580328645.0, "step": 15287 }, { "epoch": 1.9447907390917187, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.745028018951416, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8709912300109863, "num_tokens": 580365157.0, "step": 15288 }, { "epoch": 1.9449179493703093, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4977848529815674, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8772112727165222, "num_tokens": 580398195.0, "step": 15289 }, { "epoch": 1.9450451596488998, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.709651231765747, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8679388165473938, "num_tokens": 580440238.0, "step": 15290 }, { "epoch": 1.94517236992749, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3951523303985596, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8638153672218323, "num_tokens": 580477997.0, "step": 15291 }, { "epoch": 1.9452995802060806, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7505674362182617, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8721504211425781, "num_tokens": 580516335.0, "step": 15292 }, { "epoch": 1.9454267904846712, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.314272403717041, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8811873197555542, "num_tokens": 580546492.0, "step": 15293 }, { "epoch": 1.9455540007632617, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.849611759185791, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8835746049880981, "num_tokens": 580583804.0, "step": 15294 }, { "epoch": 1.9456812110418522, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2601704597473145, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8594657778739929, "num_tokens": 580618111.0, "step": 15295 }, { "epoch": 1.9458084213204427, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 4.007023811340332, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8683105707168579, "num_tokens": 580657058.0, "step": 15296 }, { "epoch": 1.945935631599033, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2873148918151855, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8612039089202881, "num_tokens": 580701628.0, "step": 15297 }, { "epoch": 1.9460628418776236, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.154024600982666, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8741905689239502, "num_tokens": 580738766.0, "step": 15298 }, { "epoch": 1.946190052156214, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.467583417892456, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8723652958869934, "num_tokens": 580780110.0, "step": 15299 }, { "epoch": 1.9463172624348046, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.120301723480225, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8785993456840515, "num_tokens": 580825208.0, "step": 15300 }, { "epoch": 1.9464444727133952, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6243367195129395, "learning_rate": 1e-05, "loss": 0.5164, "mean_token_accuracy": 0.8538516759872437, "num_tokens": 580870339.0, "step": 15301 }, { "epoch": 1.9465716829919857, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9443366527557373, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8819185495376587, "num_tokens": 580899509.0, "step": 15302 }, { "epoch": 1.9466988932705762, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.677168607711792, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8732483386993408, "num_tokens": 580941345.0, "step": 15303 }, { "epoch": 1.9468261035491667, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6992740631103516, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8770691156387329, "num_tokens": 580976819.0, "step": 15304 }, { "epoch": 1.9469533138277573, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.658325433731079, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8769855499267578, "num_tokens": 581011971.0, "step": 15305 }, { "epoch": 1.9470805241063478, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5079095363616943, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.855565071105957, "num_tokens": 581045871.0, "step": 15306 }, { "epoch": 1.9472077343849383, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6618218421936035, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.880752682685852, "num_tokens": 581077986.0, "step": 15307 }, { "epoch": 1.9473349446635289, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.36138916015625, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.873341977596283, "num_tokens": 581112988.0, "step": 15308 }, { "epoch": 1.9474621549421194, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 4.228129863739014, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8635907173156738, "num_tokens": 581155314.0, "step": 15309 }, { "epoch": 1.94758936522071, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.241626262664795, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.877411425113678, "num_tokens": 581203622.0, "step": 15310 }, { "epoch": 1.9477165754993004, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.110113143920898, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8747773170471191, "num_tokens": 581234306.0, "step": 15311 }, { "epoch": 1.947843785777891, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000274658203125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.288166046142578, "learning_rate": 1e-05, "loss": 0.5084, "mean_token_accuracy": 0.8537406921386719, "num_tokens": 581275884.0, "step": 15312 }, { "epoch": 1.9479709960564815, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.305945873260498, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8853104114532471, "num_tokens": 581312674.0, "step": 15313 }, { "epoch": 1.948098206335072, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5477821826934814, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8631178140640259, "num_tokens": 581352400.0, "step": 15314 }, { "epoch": 1.9482254166136623, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002765655517578125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.250295639038086, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8657552003860474, "num_tokens": 581390136.0, "step": 15315 }, { "epoch": 1.9483526268922529, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7272822856903076, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8828718662261963, "num_tokens": 581424864.0, "step": 15316 }, { "epoch": 1.9484798371708434, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.067691802978516, "learning_rate": 1e-05, "loss": 0.5261, "mean_token_accuracy": 0.8551334142684937, "num_tokens": 581467791.0, "step": 15317 }, { "epoch": 1.948607047449434, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7208123207092285, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8710993528366089, "num_tokens": 581505121.0, "step": 15318 }, { "epoch": 1.9487342577280244, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.879868507385254, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8744188547134399, "num_tokens": 581543306.0, "step": 15319 }, { "epoch": 1.948861468006615, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6402106285095215, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8677598834037781, "num_tokens": 581573476.0, "step": 15320 }, { "epoch": 1.9489886782852053, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.775921583175659, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8819573521614075, "num_tokens": 581606693.0, "step": 15321 }, { "epoch": 1.9491158885637958, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.536433458328247, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8703426122665405, "num_tokens": 581637826.0, "step": 15322 }, { "epoch": 1.9492430988423863, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.72509503364563, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8833268880844116, "num_tokens": 581670590.0, "step": 15323 }, { "epoch": 1.9493703091209769, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.437492847442627, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8709657192230225, "num_tokens": 581704091.0, "step": 15324 }, { "epoch": 1.9494975193995674, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7221486568450928, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8780299425125122, "num_tokens": 581747362.0, "step": 15325 }, { "epoch": 1.949624729678158, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3498473167419434, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8730287551879883, "num_tokens": 581781105.0, "step": 15326 }, { "epoch": 1.9497519399567484, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.783015489578247, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.875451922416687, "num_tokens": 581824800.0, "step": 15327 }, { "epoch": 1.949879150235339, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2948157787323, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8592316508293152, "num_tokens": 581865984.0, "step": 15328 }, { "epoch": 1.9500063605139295, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9019479751586914, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8713623285293579, "num_tokens": 581903098.0, "step": 15329 }, { "epoch": 1.95013357079252, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2853636741638184, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8856630325317383, "num_tokens": 581940704.0, "step": 15330 }, { "epoch": 1.9502607810711106, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0188374519348145, "learning_rate": 1e-05, "loss": 0.5016, "mean_token_accuracy": 0.86090487241745, "num_tokens": 581983598.0, "step": 15331 }, { "epoch": 1.950387991349701, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3746073246002197, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8751698732376099, "num_tokens": 582020073.0, "step": 15332 }, { "epoch": 1.9505152016282916, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.064303874969482, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8846877813339233, "num_tokens": 582056597.0, "step": 15333 }, { "epoch": 1.9506424119068821, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5062296390533447, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8864704966545105, "num_tokens": 582090164.0, "step": 15334 }, { "epoch": 1.9507696221854727, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.95515775680542, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8791483640670776, "num_tokens": 582130553.0, "step": 15335 }, { "epoch": 1.9508968324640632, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.591024160385132, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8772828578948975, "num_tokens": 582161175.0, "step": 15336 }, { "epoch": 1.9510240427426537, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7615394592285156, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8726770877838135, "num_tokens": 582194424.0, "step": 15337 }, { "epoch": 1.9511512530212443, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6095855236053467, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8679711818695068, "num_tokens": 582231446.0, "step": 15338 }, { "epoch": 1.9512784632998348, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.554138660430908, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8681756854057312, "num_tokens": 582270486.0, "step": 15339 }, { "epoch": 1.951405673578425, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6182854175567627, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8734437227249146, "num_tokens": 582303583.0, "step": 15340 }, { "epoch": 1.9515328838570156, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3861465454101562, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8703610897064209, "num_tokens": 582344982.0, "step": 15341 }, { "epoch": 1.9516600941356061, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.645775318145752, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8804866075515747, "num_tokens": 582383373.0, "step": 15342 }, { "epoch": 1.9517873044141967, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2660534381866455, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8704540729522705, "num_tokens": 582426671.0, "step": 15343 }, { "epoch": 1.9519145146927872, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.749481201171875, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8642871379852295, "num_tokens": 582464801.0, "step": 15344 }, { "epoch": 1.9520417249713777, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1824376583099365, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8693128228187561, "num_tokens": 582507627.0, "step": 15345 }, { "epoch": 1.952168935249968, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9085214138031006, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8812252283096313, "num_tokens": 582544072.0, "step": 15346 }, { "epoch": 1.9522961455285586, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1687912940979004, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.867384672164917, "num_tokens": 582587928.0, "step": 15347 }, { "epoch": 1.952423355807149, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.117830276489258, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.872139573097229, "num_tokens": 582622182.0, "step": 15348 }, { "epoch": 1.9525505660857396, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.298858404159546, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8894675374031067, "num_tokens": 582654890.0, "step": 15349 }, { "epoch": 1.9526777763643302, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.23171329498291, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8751192688941956, "num_tokens": 582690874.0, "step": 15350 }, { "epoch": 1.9528049866429207, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.50396990776062, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8760409951210022, "num_tokens": 582726543.0, "step": 15351 }, { "epoch": 1.9529321969215112, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.185890197753906, "learning_rate": 1e-05, "loss": 0.4998, "mean_token_accuracy": 0.861655056476593, "num_tokens": 582764552.0, "step": 15352 }, { "epoch": 1.9530594072001017, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.652773141860962, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8596674203872681, "num_tokens": 582804613.0, "step": 15353 }, { "epoch": 1.9531866174786923, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.009398460388184, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.875356912612915, "num_tokens": 582841510.0, "step": 15354 }, { "epoch": 1.9533138277572828, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.669029712677002, "learning_rate": 1e-05, "loss": 0.5414, "mean_token_accuracy": 0.8422091007232666, "num_tokens": 582879371.0, "step": 15355 }, { "epoch": 1.9534410380358733, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8329110145568848, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8639933466911316, "num_tokens": 582915354.0, "step": 15356 }, { "epoch": 1.9535682483144639, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.608168840408325, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8775817155838013, "num_tokens": 582949897.0, "step": 15357 }, { "epoch": 1.9536954585930544, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7049107551574707, "learning_rate": 1e-05, "loss": 0.4994, "mean_token_accuracy": 0.8589077591896057, "num_tokens": 582984383.0, "step": 15358 }, { "epoch": 1.953822668871645, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5372190475463867, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.866381824016571, "num_tokens": 583023146.0, "step": 15359 }, { "epoch": 1.9539498791502354, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6146373748779297, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8825773000717163, "num_tokens": 583061485.0, "step": 15360 }, { "epoch": 1.954077089428826, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.469562530517578, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8683012127876282, "num_tokens": 583097278.0, "step": 15361 }, { "epoch": 1.9542042997074165, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.6187100410461426, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8771498203277588, "num_tokens": 583140458.0, "step": 15362 }, { "epoch": 1.954331509986007, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.36857533454895, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8782601952552795, "num_tokens": 583184798.0, "step": 15363 }, { "epoch": 1.9544587202645973, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.636354446411133, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8794538378715515, "num_tokens": 583220517.0, "step": 15364 }, { "epoch": 1.9545859305431879, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3169195652008057, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8617637157440186, "num_tokens": 583256615.0, "step": 15365 }, { "epoch": 1.9547131408217784, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7089619636535645, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8668206930160522, "num_tokens": 583294533.0, "step": 15366 }, { "epoch": 1.954840351100369, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.238145589828491, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8816781044006348, "num_tokens": 583333610.0, "step": 15367 }, { "epoch": 1.9549675613789594, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.850795269012451, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8793635368347168, "num_tokens": 583366424.0, "step": 15368 }, { "epoch": 1.95509477165755, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.213404655456543, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8717958331108093, "num_tokens": 583403939.0, "step": 15369 }, { "epoch": 1.9552219819361403, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.011858940124512, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8709057569503784, "num_tokens": 583441053.0, "step": 15370 }, { "epoch": 1.9553491922147308, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2738351821899414, "learning_rate": 1e-05, "loss": 0.3752, "mean_token_accuracy": 0.8925272226333618, "num_tokens": 583471168.0, "step": 15371 }, { "epoch": 1.9554764024933213, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.137211799621582, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8669782876968384, "num_tokens": 583510604.0, "step": 15372 }, { "epoch": 1.9556036127719119, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.443648099899292, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8732262253761292, "num_tokens": 583546186.0, "step": 15373 }, { "epoch": 1.9557308230505024, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.169220924377441, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8577761650085449, "num_tokens": 583581301.0, "step": 15374 }, { "epoch": 1.955858033329093, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6128382682800293, "learning_rate": 1e-05, "loss": 0.5219, "mean_token_accuracy": 0.8557249307632446, "num_tokens": 583624005.0, "step": 15375 }, { "epoch": 1.9559852436076834, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.94266414642334, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8702985048294067, "num_tokens": 583662738.0, "step": 15376 }, { "epoch": 1.956112453886274, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6541967391967773, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.880268931388855, "num_tokens": 583700742.0, "step": 15377 }, { "epoch": 1.9562396641648645, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.716745138168335, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8782525062561035, "num_tokens": 583740529.0, "step": 15378 }, { "epoch": 1.956366874443455, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.637181282043457, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8842542171478271, "num_tokens": 583776866.0, "step": 15379 }, { "epoch": 1.9564940847220456, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5282232761383057, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8765108585357666, "num_tokens": 583815561.0, "step": 15380 }, { "epoch": 1.956621295000636, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6396336555480957, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.878175675868988, "num_tokens": 583846701.0, "step": 15381 }, { "epoch": 1.9567485052792266, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.389172315597534, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8838764429092407, "num_tokens": 583885561.0, "step": 15382 }, { "epoch": 1.9568757155578171, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.648033380508423, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8644402027130127, "num_tokens": 583928662.0, "step": 15383 }, { "epoch": 1.9570029258364077, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.302264451980591, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8782694339752197, "num_tokens": 583969097.0, "step": 15384 }, { "epoch": 1.9571301361149982, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.715883255004883, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8674925565719604, "num_tokens": 584008547.0, "step": 15385 }, { "epoch": 1.9572573463935887, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.25004506111145, "learning_rate": 1e-05, "loss": 0.5217, "mean_token_accuracy": 0.8534274101257324, "num_tokens": 584046982.0, "step": 15386 }, { "epoch": 1.9573845566721793, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8070945739746094, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8841772079467773, "num_tokens": 584081011.0, "step": 15387 }, { "epoch": 1.9575117669507698, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2087595462799072, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8622304201126099, "num_tokens": 584115550.0, "step": 15388 }, { "epoch": 1.95763897722936, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.947554588317871, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8800809383392334, "num_tokens": 584153733.0, "step": 15389 }, { "epoch": 1.9577661875079506, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2393529415130615, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8673766255378723, "num_tokens": 584182097.0, "step": 15390 }, { "epoch": 1.9578933977865411, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.074288368225098, "learning_rate": 1e-05, "loss": 0.3994, "mean_token_accuracy": 0.8931708335876465, "num_tokens": 584215089.0, "step": 15391 }, { "epoch": 1.9580206080651317, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3284616470336914, "learning_rate": 1e-05, "loss": 0.4046, "mean_token_accuracy": 0.8876973390579224, "num_tokens": 584244635.0, "step": 15392 }, { "epoch": 1.9581478183437222, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.153176784515381, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.878791332244873, "num_tokens": 584280267.0, "step": 15393 }, { "epoch": 1.9582750286223127, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.470395803451538, "learning_rate": 1e-05, "loss": 0.4063, "mean_token_accuracy": 0.8875993490219116, "num_tokens": 584311998.0, "step": 15394 }, { "epoch": 1.958402238900903, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1262736320495605, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8713777661323547, "num_tokens": 584342975.0, "step": 15395 }, { "epoch": 1.9585294491794936, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5915427207946777, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8711600303649902, "num_tokens": 584379087.0, "step": 15396 }, { "epoch": 1.958656659458084, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9710915088653564, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8650199770927429, "num_tokens": 584420126.0, "step": 15397 }, { "epoch": 1.9587838697366746, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6140847206115723, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8735010623931885, "num_tokens": 584457742.0, "step": 15398 }, { "epoch": 1.9589110800152651, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.799703359603882, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8763030767440796, "num_tokens": 584497439.0, "step": 15399 }, { "epoch": 1.9590382902938557, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5868008136749268, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8821327686309814, "num_tokens": 584532922.0, "step": 15400 }, { "epoch": 1.9591655005724462, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6496081352233887, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8827186226844788, "num_tokens": 584574834.0, "step": 15401 }, { "epoch": 1.9592927108510367, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5372016429901123, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8727221488952637, "num_tokens": 584615494.0, "step": 15402 }, { "epoch": 1.9594199211296273, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.565732479095459, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8628219366073608, "num_tokens": 584651494.0, "step": 15403 }, { "epoch": 1.9595471314082178, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4888761043548584, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8633208870887756, "num_tokens": 584691556.0, "step": 15404 }, { "epoch": 1.9596743416868083, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5051052570343018, "learning_rate": 1e-05, "loss": 0.5097, "mean_token_accuracy": 0.8513284921646118, "num_tokens": 584735960.0, "step": 15405 }, { "epoch": 1.9598015519653988, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.454596996307373, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8646572828292847, "num_tokens": 584770808.0, "step": 15406 }, { "epoch": 1.9599287622439894, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.478965997695923, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8748148679733276, "num_tokens": 584811451.0, "step": 15407 }, { "epoch": 1.96005597252258, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4154977798461914, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8704613447189331, "num_tokens": 584845050.0, "step": 15408 }, { "epoch": 1.9601831828011704, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.467543840408325, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8729891777038574, "num_tokens": 584883857.0, "step": 15409 }, { "epoch": 1.960310393079761, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3960094451904297, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8671188354492188, "num_tokens": 584918118.0, "step": 15410 }, { "epoch": 1.9604376033583515, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4610445499420166, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8699888586997986, "num_tokens": 584960789.0, "step": 15411 }, { "epoch": 1.960564813636942, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.359769582748413, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.8860123157501221, "num_tokens": 585000631.0, "step": 15412 }, { "epoch": 1.9606920239155323, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4876668453216553, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.874302327632904, "num_tokens": 585036615.0, "step": 15413 }, { "epoch": 1.9608192341941229, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3028292655944824, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8845033645629883, "num_tokens": 585075076.0, "step": 15414 }, { "epoch": 1.9609464444727134, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.560110569000244, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8779746294021606, "num_tokens": 585112215.0, "step": 15415 }, { "epoch": 1.961073654751304, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.242032527923584, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8773292303085327, "num_tokens": 585148920.0, "step": 15416 }, { "epoch": 1.9612008650298944, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6629021167755127, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8650221228599548, "num_tokens": 585191128.0, "step": 15417 }, { "epoch": 1.961328075308485, "ewc_loss": 0.072265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.458427429199219e-05, "grad_norm": 3.160442352294922, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8679506182670593, "num_tokens": 585234130.0, "step": 15418 }, { "epoch": 1.9614552855870753, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8532767295837402, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8686766624450684, "num_tokens": 585267089.0, "step": 15419 }, { "epoch": 1.9615824958656658, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.090167760848999, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8753923177719116, "num_tokens": 585310508.0, "step": 15420 }, { "epoch": 1.9617097061442563, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.143923282623291, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8767375349998474, "num_tokens": 585346579.0, "step": 15421 }, { "epoch": 1.9618369164228469, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.1831984519958496, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8658923506736755, "num_tokens": 585383914.0, "step": 15422 }, { "epoch": 1.9619641267014374, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.435149192810059, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8639909029006958, "num_tokens": 585419322.0, "step": 15423 }, { "epoch": 1.962091336980028, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5243301391601562, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.877754807472229, "num_tokens": 585459081.0, "step": 15424 }, { "epoch": 1.9622185472586184, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.445251941680908, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8661800026893616, "num_tokens": 585497116.0, "step": 15425 }, { "epoch": 1.962345757537209, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.834089994430542, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8631033301353455, "num_tokens": 585534704.0, "step": 15426 }, { "epoch": 1.9624729678157995, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.200794219970703, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.879051923751831, "num_tokens": 585575806.0, "step": 15427 }, { "epoch": 1.96260017809439, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.8369030952453613, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8789083957672119, "num_tokens": 585612585.0, "step": 15428 }, { "epoch": 1.9627273883729806, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9989941120147705, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8774985074996948, "num_tokens": 585647414.0, "step": 15429 }, { "epoch": 1.962854598651571, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.708292245864868, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8689276576042175, "num_tokens": 585693802.0, "step": 15430 }, { "epoch": 1.9629818089301616, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8521416187286377, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.873247504234314, "num_tokens": 585727955.0, "step": 15431 }, { "epoch": 1.9631090192087521, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.598112106323242, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8608608245849609, "num_tokens": 585769784.0, "step": 15432 }, { "epoch": 1.9632362294873427, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.763291835784912, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8778653740882874, "num_tokens": 585809989.0, "step": 15433 }, { "epoch": 1.9633634397659332, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.496685266494751, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8730832934379578, "num_tokens": 585851054.0, "step": 15434 }, { "epoch": 1.9634906500445237, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.748995542526245, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8615828156471252, "num_tokens": 585886638.0, "step": 15435 }, { "epoch": 1.9636178603231143, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.381124496459961, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8631636500358582, "num_tokens": 585925572.0, "step": 15436 }, { "epoch": 1.9637450706017048, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8099019527435303, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8718496561050415, "num_tokens": 585961939.0, "step": 15437 }, { "epoch": 1.963872280880295, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3099703788757324, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8690112829208374, "num_tokens": 585996804.0, "step": 15438 }, { "epoch": 1.9639994911588856, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.9157259464263916, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8808765411376953, "num_tokens": 586040347.0, "step": 15439 }, { "epoch": 1.9641267014374761, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.279235601425171, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8795222043991089, "num_tokens": 586078495.0, "step": 15440 }, { "epoch": 1.9642539117160667, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.071589946746826, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.879595160484314, "num_tokens": 586114743.0, "step": 15441 }, { "epoch": 1.9643811219946572, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.379249095916748, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8793637752532959, "num_tokens": 586150462.0, "step": 15442 }, { "epoch": 1.9645083322732477, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.144617080688477, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8711239099502563, "num_tokens": 586182860.0, "step": 15443 }, { "epoch": 1.964635542551838, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5399436950683594, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.868287205696106, "num_tokens": 586226438.0, "step": 15444 }, { "epoch": 1.9647627528304286, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.036965847015381, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.873521089553833, "num_tokens": 586267022.0, "step": 15445 }, { "epoch": 1.964889963109019, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6531193256378174, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.879623532295227, "num_tokens": 586299415.0, "step": 15446 }, { "epoch": 1.9650171733876096, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8203020095825195, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8738951683044434, "num_tokens": 586337351.0, "step": 15447 }, { "epoch": 1.9651443836662001, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6696646213531494, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8642578721046448, "num_tokens": 586378311.0, "step": 15448 }, { "epoch": 1.9652715939447907, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.604161024093628, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8731218576431274, "num_tokens": 586414137.0, "step": 15449 }, { "epoch": 1.9653988042233812, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6463260650634766, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8673657178878784, "num_tokens": 586451536.0, "step": 15450 }, { "epoch": 1.9655260145019717, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4391822814941406, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8685256838798523, "num_tokens": 586484643.0, "step": 15451 }, { "epoch": 1.9656532247805623, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.662898063659668, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8617839813232422, "num_tokens": 586523926.0, "step": 15452 }, { "epoch": 1.9657804350591528, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3032326698303223, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8754945397377014, "num_tokens": 586564999.0, "step": 15453 }, { "epoch": 1.9659076453377433, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7240943908691406, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8641883134841919, "num_tokens": 586603455.0, "step": 15454 }, { "epoch": 1.9660348556163338, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.22495174407959, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8684085607528687, "num_tokens": 586646345.0, "step": 15455 }, { "epoch": 1.9661620658949244, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8428525924682617, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8675309419631958, "num_tokens": 586688065.0, "step": 15456 }, { "epoch": 1.966289276173515, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.190095901489258, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8780428767204285, "num_tokens": 586722631.0, "step": 15457 }, { "epoch": 1.9664164864521054, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.005439281463623, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8897773623466492, "num_tokens": 586762630.0, "step": 15458 }, { "epoch": 1.966543696730696, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2378923892974854, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8722149133682251, "num_tokens": 586799514.0, "step": 15459 }, { "epoch": 1.9666709070092865, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.151878356933594, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8831560611724854, "num_tokens": 586839507.0, "step": 15460 }, { "epoch": 1.966798117287877, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.397099733352661, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8627896308898926, "num_tokens": 586879263.0, "step": 15461 }, { "epoch": 1.9669253275664673, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.179663181304932, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8729621171951294, "num_tokens": 586920297.0, "step": 15462 }, { "epoch": 1.9670525378450578, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.575786828994751, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.861556887626648, "num_tokens": 586961171.0, "step": 15463 }, { "epoch": 1.9671797481236484, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.058200359344482, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8736207485198975, "num_tokens": 587003808.0, "step": 15464 }, { "epoch": 1.967306958402239, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.63855242729187, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8609957098960876, "num_tokens": 587042056.0, "step": 15465 }, { "epoch": 1.9674341686808294, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8953776359558105, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8581238985061646, "num_tokens": 587083009.0, "step": 15466 }, { "epoch": 1.96756137895942, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6130197048187256, "learning_rate": 1e-05, "loss": 0.5191, "mean_token_accuracy": 0.8521027565002441, "num_tokens": 587125129.0, "step": 15467 }, { "epoch": 1.9676885892380103, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.73956561088562, "learning_rate": 1e-05, "loss": 0.5017, "mean_token_accuracy": 0.8561010360717773, "num_tokens": 587163314.0, "step": 15468 }, { "epoch": 1.9678157995166008, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.548497200012207, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8683362007141113, "num_tokens": 587208275.0, "step": 15469 }, { "epoch": 1.9679430097951913, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.637917995452881, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8699182868003845, "num_tokens": 587242885.0, "step": 15470 }, { "epoch": 1.9680702200737819, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4863343238830566, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.86397784948349, "num_tokens": 587287787.0, "step": 15471 }, { "epoch": 1.9681974303523724, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5813355445861816, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8847605586051941, "num_tokens": 587325445.0, "step": 15472 }, { "epoch": 1.968324640630963, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.417818546295166, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8723937273025513, "num_tokens": 587366691.0, "step": 15473 }, { "epoch": 1.9684518509095534, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5995376110076904, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.8623015284538269, "num_tokens": 587406310.0, "step": 15474 }, { "epoch": 1.968579061188144, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.358804941177368, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8685946464538574, "num_tokens": 587446654.0, "step": 15475 }, { "epoch": 1.9687062714667345, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6426925659179688, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8680387139320374, "num_tokens": 587481089.0, "step": 15476 }, { "epoch": 1.968833481745325, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2782795429229736, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8709839582443237, "num_tokens": 587521293.0, "step": 15477 }, { "epoch": 1.9689606920239155, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.735140323638916, "learning_rate": 1e-05, "loss": 0.4793, "mean_token_accuracy": 0.864205002784729, "num_tokens": 587564193.0, "step": 15478 }, { "epoch": 1.969087902302506, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.223013401031494, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8636566996574402, "num_tokens": 587603487.0, "step": 15479 }, { "epoch": 1.9692151125810966, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.881002426147461, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8611498475074768, "num_tokens": 587641131.0, "step": 15480 }, { "epoch": 1.9693423228596871, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2212600708007812, "learning_rate": 1e-05, "loss": 0.5075, "mean_token_accuracy": 0.852231502532959, "num_tokens": 587680012.0, "step": 15481 }, { "epoch": 1.9694695331382777, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.032006740570068, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8740652799606323, "num_tokens": 587721904.0, "step": 15482 }, { "epoch": 1.9695967434168682, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3100221157073975, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8751220107078552, "num_tokens": 587752080.0, "step": 15483 }, { "epoch": 1.9697239536954587, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.117790699005127, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.882010281085968, "num_tokens": 587787838.0, "step": 15484 }, { "epoch": 1.9698511639740492, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.4607436656951904, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8717248439788818, "num_tokens": 587826370.0, "step": 15485 }, { "epoch": 1.9699783742526398, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.046527862548828, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8876961469650269, "num_tokens": 587866716.0, "step": 15486 }, { "epoch": 1.97010558453123, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5734364986419678, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8747236728668213, "num_tokens": 587902766.0, "step": 15487 }, { "epoch": 1.9702327948098206, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.881539821624756, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.8597078323364258, "num_tokens": 587944584.0, "step": 15488 }, { "epoch": 1.9703600050884111, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.604868173599243, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.8602287173271179, "num_tokens": 587983482.0, "step": 15489 }, { "epoch": 1.9704872153670017, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6901583671569824, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8799479007720947, "num_tokens": 588019621.0, "step": 15490 }, { "epoch": 1.9706144256455922, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.594172239303589, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8780879974365234, "num_tokens": 588056845.0, "step": 15491 }, { "epoch": 1.9707416359241827, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.518984079360962, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8788303732872009, "num_tokens": 588093731.0, "step": 15492 }, { "epoch": 1.970868846202773, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5837550163269043, "learning_rate": 1e-05, "loss": 0.4004, "mean_token_accuracy": 0.886703372001648, "num_tokens": 588128908.0, "step": 15493 }, { "epoch": 1.9709960564813636, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4406893253326416, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8578277826309204, "num_tokens": 588167729.0, "step": 15494 }, { "epoch": 1.971123266759954, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5941271781921387, "learning_rate": 1e-05, "loss": 0.5026, "mean_token_accuracy": 0.8550336956977844, "num_tokens": 588200337.0, "step": 15495 }, { "epoch": 1.9712504770385446, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3351876735687256, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.871051013469696, "num_tokens": 588240736.0, "step": 15496 }, { "epoch": 1.9713776873171351, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6303811073303223, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8615185618400574, "num_tokens": 588275187.0, "step": 15497 }, { "epoch": 1.9715048975957257, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2877471446990967, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8645914196968079, "num_tokens": 588308800.0, "step": 15498 }, { "epoch": 1.9716321078743162, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6861276626586914, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8848252296447754, "num_tokens": 588347688.0, "step": 15499 }, { "epoch": 1.9717593181529067, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.229435920715332, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8728823661804199, "num_tokens": 588380791.0, "step": 15500 }, { "epoch": 1.9718865284314973, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.794473886489868, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.859748363494873, "num_tokens": 588413667.0, "step": 15501 }, { "epoch": 1.9720137387100878, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1705048084259033, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8730154037475586, "num_tokens": 588453884.0, "step": 15502 }, { "epoch": 1.9721409489886783, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9785616397857666, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8737735748291016, "num_tokens": 588488895.0, "step": 15503 }, { "epoch": 1.9722681592672688, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.181394100189209, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8758635520935059, "num_tokens": 588526807.0, "step": 15504 }, { "epoch": 1.9723953695458594, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.194436550140381, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.87602698802948, "num_tokens": 588562272.0, "step": 15505 }, { "epoch": 1.97252257982445, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3581435680389404, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.867980420589447, "num_tokens": 588598298.0, "step": 15506 }, { "epoch": 1.9726497901030404, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2825188636779785, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8722733855247498, "num_tokens": 588637833.0, "step": 15507 }, { "epoch": 1.972777000381631, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.571925640106201, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8701906204223633, "num_tokens": 588674933.0, "step": 15508 }, { "epoch": 1.9729042106602215, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.1880202293396, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8731350302696228, "num_tokens": 588711899.0, "step": 15509 }, { "epoch": 1.973031420938812, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6983895301818848, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8709263801574707, "num_tokens": 588749702.0, "step": 15510 }, { "epoch": 1.9731586312174023, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9863452911376953, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8784308433532715, "num_tokens": 588787066.0, "step": 15511 }, { "epoch": 1.9732858414959928, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.689631462097168, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8735088109970093, "num_tokens": 588824062.0, "step": 15512 }, { "epoch": 1.9734130517745834, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.791361093521118, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8704103827476501, "num_tokens": 588864858.0, "step": 15513 }, { "epoch": 1.973540262053174, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6238808631896973, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8680068254470825, "num_tokens": 588900219.0, "step": 15514 }, { "epoch": 1.9736674723317644, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6538636684417725, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8673489093780518, "num_tokens": 588937256.0, "step": 15515 }, { "epoch": 1.973794682610355, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5455493927001953, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.878696620464325, "num_tokens": 588971263.0, "step": 15516 }, { "epoch": 1.9739218928889453, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.576137065887451, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8676089644432068, "num_tokens": 589010710.0, "step": 15517 }, { "epoch": 1.9740491031675358, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4848873615264893, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8759625554084778, "num_tokens": 589043940.0, "step": 15518 }, { "epoch": 1.9741763134461263, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.539135694503784, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.868148922920227, "num_tokens": 589082187.0, "step": 15519 }, { "epoch": 1.9743035237247168, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3990612030029297, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8790622353553772, "num_tokens": 589126870.0, "step": 15520 }, { "epoch": 1.9744307340033074, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5456271171569824, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8664484024047852, "num_tokens": 589168378.0, "step": 15521 }, { "epoch": 1.974557944281898, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3396716117858887, "learning_rate": 1e-05, "loss": 0.5022, "mean_token_accuracy": 0.8552127480506897, "num_tokens": 589213070.0, "step": 15522 }, { "epoch": 1.9746851545604884, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.6048126220703125, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.868366003036499, "num_tokens": 589253008.0, "step": 15523 }, { "epoch": 1.974812364839079, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2745728492736816, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8654138445854187, "num_tokens": 589291786.0, "step": 15524 }, { "epoch": 1.9749395751176695, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6940784454345703, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.872808575630188, "num_tokens": 589332045.0, "step": 15525 }, { "epoch": 1.97506678539626, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.2020866870880127, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8507559299468994, "num_tokens": 589374983.0, "step": 15526 }, { "epoch": 1.9751939956748505, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8330936431884766, "learning_rate": 1e-05, "loss": 0.3912, "mean_token_accuracy": 0.8949153423309326, "num_tokens": 589419786.0, "step": 15527 }, { "epoch": 1.975321205953441, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.159740447998047, "learning_rate": 1e-05, "loss": 0.3923, "mean_token_accuracy": 0.8899561166763306, "num_tokens": 589454365.0, "step": 15528 }, { "epoch": 1.9754484162320316, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.066755294799805, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8758828639984131, "num_tokens": 589486829.0, "step": 15529 }, { "epoch": 1.9755756265106221, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.225527286529541, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8730854392051697, "num_tokens": 589527500.0, "step": 15530 }, { "epoch": 1.9757028367892127, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.260807991027832, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8839300870895386, "num_tokens": 589570064.0, "step": 15531 }, { "epoch": 1.9758300470678032, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4679980278015137, "learning_rate": 1e-05, "loss": 0.5079, "mean_token_accuracy": 0.855247974395752, "num_tokens": 589605327.0, "step": 15532 }, { "epoch": 1.9759572573463937, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.257892608642578, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8810796737670898, "num_tokens": 589640886.0, "step": 15533 }, { "epoch": 1.9760844676249842, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.663161039352417, "learning_rate": 1e-05, "loss": 0.5135, "mean_token_accuracy": 0.8570811152458191, "num_tokens": 589681178.0, "step": 15534 }, { "epoch": 1.9762116779035748, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.075865268707275, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.883507490158081, "num_tokens": 589719073.0, "step": 15535 }, { "epoch": 1.976338888182165, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7058868408203125, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8617959022521973, "num_tokens": 589760528.0, "step": 15536 }, { "epoch": 1.9764660984607556, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.88114595413208, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8712300062179565, "num_tokens": 589795946.0, "step": 15537 }, { "epoch": 1.9765933087393461, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.666799306869507, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8705658912658691, "num_tokens": 589829139.0, "step": 15538 }, { "epoch": 1.9767205190179367, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.699113130569458, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8840965032577515, "num_tokens": 589862916.0, "step": 15539 }, { "epoch": 1.9768477292965272, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5909101963043213, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8783186078071594, "num_tokens": 589900230.0, "step": 15540 }, { "epoch": 1.9769749395751177, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5802619457244873, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8714262247085571, "num_tokens": 589938737.0, "step": 15541 }, { "epoch": 1.977102149853708, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5234673023223877, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8777751326560974, "num_tokens": 589977989.0, "step": 15542 }, { "epoch": 1.9772293601322986, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5140485763549805, "learning_rate": 1e-05, "loss": 0.3848, "mean_token_accuracy": 0.8892102241516113, "num_tokens": 590015547.0, "step": 15543 }, { "epoch": 1.977356570410889, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.457153081893921, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8724902272224426, "num_tokens": 590055257.0, "step": 15544 }, { "epoch": 1.9774837806894796, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5110628604888916, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8750696182250977, "num_tokens": 590087911.0, "step": 15545 }, { "epoch": 1.9776109909680701, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.400219440460205, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8682708740234375, "num_tokens": 590128889.0, "step": 15546 }, { "epoch": 1.9777382012466607, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5187673568725586, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8596004247665405, "num_tokens": 590168078.0, "step": 15547 }, { "epoch": 1.9778654115252512, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.346773386001587, "learning_rate": 1e-05, "loss": 0.5037, "mean_token_accuracy": 0.8533404469490051, "num_tokens": 590211361.0, "step": 15548 }, { "epoch": 1.9779926218038417, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5683751106262207, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8616591691970825, "num_tokens": 590249922.0, "step": 15549 }, { "epoch": 1.9781198320824323, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.2774980068206787, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8729948997497559, "num_tokens": 590288284.0, "step": 15550 }, { "epoch": 1.9782470423610228, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.649247646331787, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8580197095870972, "num_tokens": 590324373.0, "step": 15551 }, { "epoch": 1.9783742526396133, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.1989309787750244, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8731166124343872, "num_tokens": 590365374.0, "step": 15552 }, { "epoch": 1.9785014629182038, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8045740127563477, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8699005246162415, "num_tokens": 590402712.0, "step": 15553 }, { "epoch": 1.9786286731967944, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.506111145019531e-05, "grad_norm": 3.162623167037964, "learning_rate": 1e-05, "loss": 0.5353, "mean_token_accuracy": 0.8535985946655273, "num_tokens": 590438434.0, "step": 15554 }, { "epoch": 1.978755883475385, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 4.002389907836914, "learning_rate": 1e-05, "loss": 0.4969, "mean_token_accuracy": 0.8627616167068481, "num_tokens": 590482619.0, "step": 15555 }, { "epoch": 1.9788830937539754, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1547303199768066, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8558567762374878, "num_tokens": 590527828.0, "step": 15556 }, { "epoch": 1.979010304032566, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.268837928771973, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8745183348655701, "num_tokens": 590561758.0, "step": 15557 }, { "epoch": 1.9791375143111565, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3722941875457764, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8833187818527222, "num_tokens": 590598323.0, "step": 15558 }, { "epoch": 1.979264724589747, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.36251974105835, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8660567402839661, "num_tokens": 590642469.0, "step": 15559 }, { "epoch": 1.9793919348683373, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.66794753074646, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8654091358184814, "num_tokens": 590683603.0, "step": 15560 }, { "epoch": 1.9795191451469278, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.238327980041504, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8805401921272278, "num_tokens": 590710136.0, "step": 15561 }, { "epoch": 1.9796463554255184, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.789742946624756, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8693243265151978, "num_tokens": 590743104.0, "step": 15562 }, { "epoch": 1.979773565704109, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9949123859405518, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8663833141326904, "num_tokens": 590781851.0, "step": 15563 }, { "epoch": 1.9799007759826994, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.7257540225982666, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8602733612060547, "num_tokens": 590819865.0, "step": 15564 }, { "epoch": 1.98002798626129, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8123586177825928, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8805932998657227, "num_tokens": 590856478.0, "step": 15565 }, { "epoch": 1.9801551965398803, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6317598819732666, "learning_rate": 1e-05, "loss": 0.4876, "mean_token_accuracy": 0.8606306314468384, "num_tokens": 590895765.0, "step": 15566 }, { "epoch": 1.9802824068184708, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6868772506713867, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8746311068534851, "num_tokens": 590931238.0, "step": 15567 }, { "epoch": 1.9804096170970613, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.544192314147949, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8748363256454468, "num_tokens": 590969875.0, "step": 15568 }, { "epoch": 1.9805368273756518, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.624680757522583, "learning_rate": 1e-05, "loss": 0.4906, "mean_token_accuracy": 0.8577901721000671, "num_tokens": 591007964.0, "step": 15569 }, { "epoch": 1.9806640376542424, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.439960479736328, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.875942587852478, "num_tokens": 591049339.0, "step": 15570 }, { "epoch": 1.980791247932833, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.65858793258667, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8803136348724365, "num_tokens": 591086429.0, "step": 15571 }, { "epoch": 1.9809184582114234, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.359577178955078, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.876056969165802, "num_tokens": 591128252.0, "step": 15572 }, { "epoch": 1.981045668490014, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6741750240325928, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.8840211629867554, "num_tokens": 591165049.0, "step": 15573 }, { "epoch": 1.9811728787686045, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2715327739715576, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.881986141204834, "num_tokens": 591196805.0, "step": 15574 }, { "epoch": 1.981300089047195, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7815680503845215, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8728259205818176, "num_tokens": 591242897.0, "step": 15575 }, { "epoch": 1.9814272993257855, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.204150438308716, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8705251216888428, "num_tokens": 591280684.0, "step": 15576 }, { "epoch": 1.981554509604376, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9636666774749756, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8751495480537415, "num_tokens": 591315264.0, "step": 15577 }, { "epoch": 1.9816817198829666, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.220560073852539, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.868701696395874, "num_tokens": 591358028.0, "step": 15578 }, { "epoch": 1.9818089301615571, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.156621932983398, "learning_rate": 1e-05, "loss": 0.5334, "mean_token_accuracy": 0.8525197505950928, "num_tokens": 591404272.0, "step": 15579 }, { "epoch": 1.9819361404401477, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.377464771270752, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8850139379501343, "num_tokens": 591439896.0, "step": 15580 }, { "epoch": 1.9820633507187382, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.220550537109375, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8780403733253479, "num_tokens": 591477318.0, "step": 15581 }, { "epoch": 1.9821905609973287, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.606468439102173, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8672801852226257, "num_tokens": 591518600.0, "step": 15582 }, { "epoch": 1.9823177712759192, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.096583366394043, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8679425716400146, "num_tokens": 591547414.0, "step": 15583 }, { "epoch": 1.9824449815545098, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6895101070404053, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8595285415649414, "num_tokens": 591594270.0, "step": 15584 }, { "epoch": 1.9825721918331, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8562042713165283, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8746012449264526, "num_tokens": 591638252.0, "step": 15585 }, { "epoch": 1.9826994021116906, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6737070083618164, "learning_rate": 1e-05, "loss": 0.529, "mean_token_accuracy": 0.8506314754486084, "num_tokens": 591680268.0, "step": 15586 }, { "epoch": 1.9828266123902811, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6456665992736816, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8754677176475525, "num_tokens": 591717724.0, "step": 15587 }, { "epoch": 1.9829538226688717, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6406118869781494, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.885536789894104, "num_tokens": 591754562.0, "step": 15588 }, { "epoch": 1.9830810329474622, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.487147331237793, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8761391639709473, "num_tokens": 591794838.0, "step": 15589 }, { "epoch": 1.9832082432260527, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6374752521514893, "learning_rate": 1e-05, "loss": 0.5148, "mean_token_accuracy": 0.8517651557922363, "num_tokens": 591834771.0, "step": 15590 }, { "epoch": 1.983335453504643, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.386998176574707, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8595188856124878, "num_tokens": 591872764.0, "step": 15591 }, { "epoch": 1.9834626637832335, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.653778314590454, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8670576810836792, "num_tokens": 591907035.0, "step": 15592 }, { "epoch": 1.983589874061824, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3111040592193604, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8765370845794678, "num_tokens": 591940797.0, "step": 15593 }, { "epoch": 1.9837170843404146, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.708176374435425, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8862359523773193, "num_tokens": 591976872.0, "step": 15594 }, { "epoch": 1.9838442946190051, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.235644817352295, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8566575050354004, "num_tokens": 592015933.0, "step": 15595 }, { "epoch": 1.9839715048975957, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8326356410980225, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8784027099609375, "num_tokens": 592046796.0, "step": 15596 }, { "epoch": 1.9840987151761862, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1802401542663574, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8709883689880371, "num_tokens": 592092841.0, "step": 15597 }, { "epoch": 1.9842259254547767, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 4.015299320220947, "learning_rate": 1e-05, "loss": 0.5152, "mean_token_accuracy": 0.854617714881897, "num_tokens": 592134118.0, "step": 15598 }, { "epoch": 1.9843531357333672, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2253525257110596, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8649893999099731, "num_tokens": 592167292.0, "step": 15599 }, { "epoch": 1.9844803460119578, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.176862716674805, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8835126161575317, "num_tokens": 592205616.0, "step": 15600 }, { "epoch": 1.9846075562905483, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3820176124572754, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8785144090652466, "num_tokens": 592246650.0, "step": 15601 }, { "epoch": 1.9847347665691388, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.235087871551514, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8757325410842896, "num_tokens": 592284729.0, "step": 15602 }, { "epoch": 1.9848619768477294, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.5928776264190674, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8749872446060181, "num_tokens": 592318423.0, "step": 15603 }, { "epoch": 1.9849891871263199, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.118076324462891, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.870025634765625, "num_tokens": 592351304.0, "step": 15604 }, { "epoch": 1.9851163974049104, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6734886169433594, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.873572826385498, "num_tokens": 592388353.0, "step": 15605 }, { "epoch": 1.985243607683501, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9238078594207764, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8778631687164307, "num_tokens": 592427651.0, "step": 15606 }, { "epoch": 1.9853708179620915, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6483066082000732, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.872683048248291, "num_tokens": 592463638.0, "step": 15607 }, { "epoch": 1.985498028240682, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7590746879577637, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8616011738777161, "num_tokens": 592501508.0, "step": 15608 }, { "epoch": 1.9856252385192723, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.584404230117798, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8795858025550842, "num_tokens": 592538418.0, "step": 15609 }, { "epoch": 1.9857524487978628, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6275131702423096, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8855559229850769, "num_tokens": 592573552.0, "step": 15610 }, { "epoch": 1.9858796590764534, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5219099521636963, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8743250370025635, "num_tokens": 592607212.0, "step": 15611 }, { "epoch": 1.986006869355044, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5559146404266357, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8657180070877075, "num_tokens": 592649910.0, "step": 15612 }, { "epoch": 1.9861340796336344, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.452730894088745, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8723106980323792, "num_tokens": 592689555.0, "step": 15613 }, { "epoch": 1.986261289912225, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5382778644561768, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8763540983200073, "num_tokens": 592731921.0, "step": 15614 }, { "epoch": 1.9863885001908153, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.413313865661621, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8757317066192627, "num_tokens": 592763631.0, "step": 15615 }, { "epoch": 1.9865157104694058, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5334153175354004, "learning_rate": 1e-05, "loss": 0.3908, "mean_token_accuracy": 0.890155553817749, "num_tokens": 592796961.0, "step": 15616 }, { "epoch": 1.9866429207479963, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3481531143188477, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.878963053226471, "num_tokens": 592831380.0, "step": 15617 }, { "epoch": 1.9867701310265868, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.576497793197632, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8645479679107666, "num_tokens": 592873606.0, "step": 15618 }, { "epoch": 1.9868973413051774, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2744197845458984, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8646562695503235, "num_tokens": 592919311.0, "step": 15619 }, { "epoch": 1.987024551583768, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6770663261413574, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8495783805847168, "num_tokens": 592961818.0, "step": 15620 }, { "epoch": 1.9871517618623584, "ewc_loss": 0.07275390625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.482269287109375e-05, "grad_norm": 3.2020602226257324, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8703965544700623, "num_tokens": 593003320.0, "step": 15621 }, { "epoch": 1.987278972140949, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8270950317382812, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8620250225067139, "num_tokens": 593050887.0, "step": 15622 }, { "epoch": 1.9874061824195395, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1598012447357178, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.861835241317749, "num_tokens": 593086296.0, "step": 15623 }, { "epoch": 1.98753339269813, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.054652690887451, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8543636798858643, "num_tokens": 593122135.0, "step": 15624 }, { "epoch": 1.9876606029767205, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.206894874572754, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8636647462844849, "num_tokens": 593158845.0, "step": 15625 }, { "epoch": 1.987787813255311, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.26332426071167, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8777550458908081, "num_tokens": 593199386.0, "step": 15626 }, { "epoch": 1.9879150235339016, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.427738904953003, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8801432251930237, "num_tokens": 593237061.0, "step": 15627 }, { "epoch": 1.9880422338124921, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.312050819396973, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8724477291107178, "num_tokens": 593275220.0, "step": 15628 }, { "epoch": 1.9881694440910826, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.691105842590332, "learning_rate": 1e-05, "loss": 0.5226, "mean_token_accuracy": 0.850892186164856, "num_tokens": 593314576.0, "step": 15629 }, { "epoch": 1.9882966543696732, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1127471923828125, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8793905973434448, "num_tokens": 593362905.0, "step": 15630 }, { "epoch": 1.9884238646482637, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.73797869682312, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.879150927066803, "num_tokens": 593402486.0, "step": 15631 }, { "epoch": 1.9885510749268542, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.8951103687286377, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8820770978927612, "num_tokens": 593440555.0, "step": 15632 }, { "epoch": 1.9886782852054448, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6814041137695312, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8711409568786621, "num_tokens": 593480488.0, "step": 15633 }, { "epoch": 1.988805495484035, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7131102085113525, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8778260946273804, "num_tokens": 593518615.0, "step": 15634 }, { "epoch": 1.9889327057626256, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.621126890182495, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8779433965682983, "num_tokens": 593556990.0, "step": 15635 }, { "epoch": 1.9890599160412161, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5896852016448975, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.862554132938385, "num_tokens": 593599438.0, "step": 15636 }, { "epoch": 1.9891871263198067, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.551237106323242, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8592044711112976, "num_tokens": 593635991.0, "step": 15637 }, { "epoch": 1.9893143365983972, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5061471462249756, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8725349307060242, "num_tokens": 593676600.0, "step": 15638 }, { "epoch": 1.9894415468769877, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4955430030822754, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8815621733665466, "num_tokens": 593709069.0, "step": 15639 }, { "epoch": 1.989568757155578, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.46823787689209, "learning_rate": 1e-05, "loss": 0.3845, "mean_token_accuracy": 0.8960912227630615, "num_tokens": 593740361.0, "step": 15640 }, { "epoch": 1.9896959674341685, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4429144859313965, "learning_rate": 1e-05, "loss": 0.3868, "mean_token_accuracy": 0.8867560625076294, "num_tokens": 593775424.0, "step": 15641 }, { "epoch": 1.989823177712759, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.55072283744812, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8646475076675415, "num_tokens": 593816562.0, "step": 15642 }, { "epoch": 1.9899503879913496, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.4153900146484375, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8732976913452148, "num_tokens": 593853272.0, "step": 15643 }, { "epoch": 1.9900775982699401, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.506526470184326, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8775758743286133, "num_tokens": 593885308.0, "step": 15644 }, { "epoch": 1.9902048085485307, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3242597579956055, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.8847893476486206, "num_tokens": 593918150.0, "step": 15645 }, { "epoch": 1.9903320188271212, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.55252742767334, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8673636317253113, "num_tokens": 593961947.0, "step": 15646 }, { "epoch": 1.9904592291057117, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.279876947402954, "learning_rate": 1e-05, "loss": 0.4902, "mean_token_accuracy": 0.8569319248199463, "num_tokens": 594000659.0, "step": 15647 }, { "epoch": 1.9905864393843022, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6302292346954346, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8726546764373779, "num_tokens": 594040033.0, "step": 15648 }, { "epoch": 1.9907136496628928, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.198154926300049, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8774998784065247, "num_tokens": 594080534.0, "step": 15649 }, { "epoch": 1.9908408599414833, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7795419692993164, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.8682988882064819, "num_tokens": 594118015.0, "step": 15650 }, { "epoch": 1.9909680702200738, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.127516031265259, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.87602698802948, "num_tokens": 594153839.0, "step": 15651 }, { "epoch": 1.9910952804986644, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.018368721008301, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8822540044784546, "num_tokens": 594195672.0, "step": 15652 }, { "epoch": 1.9912224907772549, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1347312927246094, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8665546178817749, "num_tokens": 594239178.0, "step": 15653 }, { "epoch": 1.9913497010558454, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.325988292694092, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.874605655670166, "num_tokens": 594274938.0, "step": 15654 }, { "epoch": 1.991476911334436, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.371849775314331, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8807238936424255, "num_tokens": 594310632.0, "step": 15655 }, { "epoch": 1.9916041216130265, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.461619853973389, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8664555549621582, "num_tokens": 594344308.0, "step": 15656 }, { "epoch": 1.991731331891617, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.7307794094085693, "learning_rate": 1e-05, "loss": 0.4726, "mean_token_accuracy": 0.8694674372673035, "num_tokens": 594385287.0, "step": 15657 }, { "epoch": 1.9918585421702073, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.283838748931885, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8825547695159912, "num_tokens": 594422521.0, "step": 15658 }, { "epoch": 1.9919857524487978, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.8379623889923096, "learning_rate": 1e-05, "loss": 0.5174, "mean_token_accuracy": 0.8571758270263672, "num_tokens": 594466837.0, "step": 15659 }, { "epoch": 1.9921129627273884, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.063500881195068, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8709456920623779, "num_tokens": 594501283.0, "step": 15660 }, { "epoch": 1.9922401730059789, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7620062828063965, "learning_rate": 1e-05, "loss": 0.5053, "mean_token_accuracy": 0.8572335839271545, "num_tokens": 594544627.0, "step": 15661 }, { "epoch": 1.9923673832845694, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.895557403564453, "learning_rate": 1e-05, "loss": 0.5255, "mean_token_accuracy": 0.8507704138755798, "num_tokens": 594586363.0, "step": 15662 }, { "epoch": 1.99249459356316, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.643707752227783, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8696606159210205, "num_tokens": 594624950.0, "step": 15663 }, { "epoch": 1.9926218038417502, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7818169593811035, "learning_rate": 1e-05, "loss": 0.5072, "mean_token_accuracy": 0.8569574952125549, "num_tokens": 594665103.0, "step": 15664 }, { "epoch": 1.9927490141203408, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5383517742156982, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.881700873374939, "num_tokens": 594699506.0, "step": 15665 }, { "epoch": 1.9928762243989313, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7314865589141846, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8664330244064331, "num_tokens": 594741771.0, "step": 15666 }, { "epoch": 1.9930034346775218, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.420591354370117, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8810603618621826, "num_tokens": 594784705.0, "step": 15667 }, { "epoch": 1.9931306449561124, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7494988441467285, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8773056268692017, "num_tokens": 594827800.0, "step": 15668 }, { "epoch": 1.993257855234703, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.357879638671875, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.868860125541687, "num_tokens": 594865150.0, "step": 15669 }, { "epoch": 1.9933850655132934, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8356246948242188, "learning_rate": 1e-05, "loss": 0.5071, "mean_token_accuracy": 0.8573035001754761, "num_tokens": 594903440.0, "step": 15670 }, { "epoch": 1.993512275791884, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2896881103515625, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8750014901161194, "num_tokens": 594936761.0, "step": 15671 }, { "epoch": 1.9936394860704745, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9711835384368896, "learning_rate": 1e-05, "loss": 0.4788, "mean_token_accuracy": 0.8592269420623779, "num_tokens": 594972298.0, "step": 15672 }, { "epoch": 1.993766696349065, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.301530122756958, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8715174794197083, "num_tokens": 595009530.0, "step": 15673 }, { "epoch": 1.9938939066276555, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.093019008636475, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8754845857620239, "num_tokens": 595048147.0, "step": 15674 }, { "epoch": 1.994021116906246, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4146311283111572, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8841389417648315, "num_tokens": 595083580.0, "step": 15675 }, { "epoch": 1.9941483271848366, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.114080905914307, "learning_rate": 1e-05, "loss": 0.5291, "mean_token_accuracy": 0.8514772653579712, "num_tokens": 595122989.0, "step": 15676 }, { "epoch": 1.9942755374634271, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.5700342655181885, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.86673504114151, "num_tokens": 595165378.0, "step": 15677 }, { "epoch": 1.9944027477420176, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.983588457107544, "learning_rate": 1e-05, "loss": 0.5375, "mean_token_accuracy": 0.8498983979225159, "num_tokens": 595203802.0, "step": 15678 }, { "epoch": 1.9945299580206082, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6426985263824463, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.867216169834137, "num_tokens": 595242056.0, "step": 15679 }, { "epoch": 1.9946571682991987, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.772796869277954, "learning_rate": 1e-05, "loss": 0.5026, "mean_token_accuracy": 0.8563922643661499, "num_tokens": 595276381.0, "step": 15680 }, { "epoch": 1.9947843785777892, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6346333026885986, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8717458248138428, "num_tokens": 595318551.0, "step": 15681 }, { "epoch": 1.9949115888563798, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.570371389389038, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8619692325592041, "num_tokens": 595361802.0, "step": 15682 }, { "epoch": 1.99503879913497, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.626558303833008, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8725993633270264, "num_tokens": 595402853.0, "step": 15683 }, { "epoch": 1.9951660094135606, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.4148061275482178, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8772676587104797, "num_tokens": 595443774.0, "step": 15684 }, { "epoch": 1.9952932196921511, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6336019039154053, "learning_rate": 1e-05, "loss": 0.4688, "mean_token_accuracy": 0.8650972247123718, "num_tokens": 595485395.0, "step": 15685 }, { "epoch": 1.9954204299707416, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.3145337104797363, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8655553460121155, "num_tokens": 595527724.0, "step": 15686 }, { "epoch": 1.9955476402493322, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.690901517868042, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8605179190635681, "num_tokens": 595569081.0, "step": 15687 }, { "epoch": 1.9956748505279227, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.23653244972229, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8608662486076355, "num_tokens": 595611400.0, "step": 15688 }, { "epoch": 1.995802060806513, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.811893939971924, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8689286112785339, "num_tokens": 595648063.0, "step": 15689 }, { "epoch": 1.9959292710851035, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1817150115966797, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8578256368637085, "num_tokens": 595688679.0, "step": 15690 }, { "epoch": 1.996056481363694, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9981436729431152, "learning_rate": 1e-05, "loss": 0.4894, "mean_token_accuracy": 0.8608875274658203, "num_tokens": 595726120.0, "step": 15691 }, { "epoch": 1.9961836916422846, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.209364891052246, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8717098236083984, "num_tokens": 595762020.0, "step": 15692 }, { "epoch": 1.9963109019208751, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.186121940612793, "learning_rate": 1e-05, "loss": 0.5259, "mean_token_accuracy": 0.8503643274307251, "num_tokens": 595797909.0, "step": 15693 }, { "epoch": 1.9964381121994657, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3737220764160156, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8672193288803101, "num_tokens": 595837542.0, "step": 15694 }, { "epoch": 1.9965653224780562, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.281960487365723, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8746480941772461, "num_tokens": 595874006.0, "step": 15695 }, { "epoch": 1.9966925327566467, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6285829544067383, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8644263744354248, "num_tokens": 595904098.0, "step": 15696 }, { "epoch": 1.9968197430352372, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.102930068969727, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8758665323257446, "num_tokens": 595945725.0, "step": 15697 }, { "epoch": 1.9969469533138278, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6980018615722656, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8700025081634521, "num_tokens": 595977826.0, "step": 15698 }, { "epoch": 1.9970741635924183, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9145073890686035, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8631251454353333, "num_tokens": 596011350.0, "step": 15699 }, { "epoch": 1.9972013738710088, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6484804153442383, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8848031759262085, "num_tokens": 596050688.0, "step": 15700 }, { "epoch": 1.9973285841495994, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7418665885925293, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8802866339683533, "num_tokens": 596090711.0, "step": 15701 }, { "epoch": 1.9974557944281899, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5770742893218994, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8692626953125, "num_tokens": 596130346.0, "step": 15702 }, { "epoch": 1.9975830047067804, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6353249549865723, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8611943125724792, "num_tokens": 596173035.0, "step": 15703 }, { "epoch": 1.997710214985371, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.520423650741577, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8539999723434448, "num_tokens": 596208105.0, "step": 15704 }, { "epoch": 1.9978374252639615, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5668017864227295, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8844587802886963, "num_tokens": 596247367.0, "step": 15705 }, { "epoch": 1.997964635542552, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4467904567718506, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8686770796775818, "num_tokens": 596288226.0, "step": 15706 }, { "epoch": 1.9980918458211423, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5560896396636963, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8599391579627991, "num_tokens": 596329422.0, "step": 15707 }, { "epoch": 1.9982190560997328, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.379171371459961, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8740927577018738, "num_tokens": 596367447.0, "step": 15708 }, { "epoch": 1.9983462663783234, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.594254970550537, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8672534227371216, "num_tokens": 596401564.0, "step": 15709 }, { "epoch": 1.9984734766569139, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3135130405426025, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8641061782836914, "num_tokens": 596441076.0, "step": 15710 }, { "epoch": 1.9986006869355044, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6494531631469727, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8788352012634277, "num_tokens": 596480360.0, "step": 15711 }, { "epoch": 1.998727897214095, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.262413740158081, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8603943586349487, "num_tokens": 596512836.0, "step": 15712 }, { "epoch": 1.9988551074926852, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.757525682449341, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8841280937194824, "num_tokens": 596549154.0, "step": 15713 }, { "epoch": 1.9989823177712758, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1809909343719482, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8780655860900879, "num_tokens": 596590528.0, "step": 15714 }, { "epoch": 1.9991095280498663, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.949502468109131, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8717520236968994, "num_tokens": 596623926.0, "step": 15715 }, { "epoch": 1.9992367383284568, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.189903497695923, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8724952936172485, "num_tokens": 596661102.0, "step": 15716 }, { "epoch": 1.9993639486070474, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.147222995758057, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8837335109710693, "num_tokens": 596701113.0, "step": 15717 }, { "epoch": 1.9994911588856379, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.338491678237915, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8810580372810364, "num_tokens": 596735130.0, "step": 15718 }, { "epoch": 1.9996183691642284, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.250430107116699, "learning_rate": 1e-05, "loss": 0.5117, "mean_token_accuracy": 0.8575249910354614, "num_tokens": 596774612.0, "step": 15719 }, { "epoch": 1.999745579442819, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.566162347793579, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8703536987304688, "num_tokens": 596811033.0, "step": 15720 }, { "epoch": 1.9998727897214095, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.137333869934082, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8596354722976685, "num_tokens": 596849169.0, "step": 15721 }, { "epoch": 2.0, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6723711490631104, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8773351907730103, "num_tokens": 596886661.0, "step": 15722 }, { "epoch": 2.0001272102785905, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9316704273223877, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8872492909431458, "num_tokens": 596927769.0, "step": 15723 }, { "epoch": 2.000254420557181, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.660902261734009, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8812507390975952, "num_tokens": 596966558.0, "step": 15724 }, { "epoch": 2.0003816308357716, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7442195415496826, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8860784769058228, "num_tokens": 597007424.0, "step": 15725 }, { "epoch": 2.000508841114362, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.608823776245117, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8811923265457153, "num_tokens": 597043544.0, "step": 15726 }, { "epoch": 2.0006360513929526, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6010611057281494, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8791198134422302, "num_tokens": 597081079.0, "step": 15727 }, { "epoch": 2.000763261671543, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5530316829681396, "learning_rate": 1e-05, "loss": 0.3848, "mean_token_accuracy": 0.8891108632087708, "num_tokens": 597122188.0, "step": 15728 }, { "epoch": 2.0008904719501337, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5083274841308594, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8788773417472839, "num_tokens": 597164835.0, "step": 15729 }, { "epoch": 2.0010176822287242, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5142884254455566, "learning_rate": 1e-05, "loss": 0.4004, "mean_token_accuracy": 0.8854237794876099, "num_tokens": 597201266.0, "step": 15730 }, { "epoch": 2.0011448925073148, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4506337642669678, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8842089176177979, "num_tokens": 597244095.0, "step": 15731 }, { "epoch": 2.0012721027859053, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4806768894195557, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8802032470703125, "num_tokens": 597285462.0, "step": 15732 }, { "epoch": 2.001399313064496, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.424090623855591, "learning_rate": 1e-05, "loss": 0.4023, "mean_token_accuracy": 0.8865917921066284, "num_tokens": 597323436.0, "step": 15733 }, { "epoch": 2.0015265233430863, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.468867778778076, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8763444423675537, "num_tokens": 597361775.0, "step": 15734 }, { "epoch": 2.0016537336216764, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.39347243309021, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.871545672416687, "num_tokens": 597402772.0, "step": 15735 }, { "epoch": 2.001780943900267, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4666647911071777, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.886717677116394, "num_tokens": 597444245.0, "step": 15736 }, { "epoch": 2.0019081541788575, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.371311664581299, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.882574737071991, "num_tokens": 597481742.0, "step": 15737 }, { "epoch": 2.002035364457448, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.483936071395874, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8715502023696899, "num_tokens": 597520339.0, "step": 15738 }, { "epoch": 2.0021625747360385, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.3407464027404785, "learning_rate": 1e-05, "loss": 0.3903, "mean_token_accuracy": 0.8876592516899109, "num_tokens": 597552756.0, "step": 15739 }, { "epoch": 2.002289785014629, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.5136337280273438, "learning_rate": 1e-05, "loss": 0.3893, "mean_token_accuracy": 0.8892583847045898, "num_tokens": 597582224.0, "step": 15740 }, { "epoch": 2.0024169952932196, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2841081619262695, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8817852139472961, "num_tokens": 597619121.0, "step": 15741 }, { "epoch": 2.00254420557181, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5886049270629883, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8712237477302551, "num_tokens": 597654074.0, "step": 15742 }, { "epoch": 2.0026714158504006, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1978020668029785, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.883470356464386, "num_tokens": 597695873.0, "step": 15743 }, { "epoch": 2.002798626128991, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.734464168548584, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8845382332801819, "num_tokens": 597733072.0, "step": 15744 }, { "epoch": 2.0029258364075817, "ewc_loss": 0.0732421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.5299530029296875e-05, "grad_norm": 3.105382204055786, "learning_rate": 1e-05, "loss": 0.3488, "mean_token_accuracy": 0.902131199836731, "num_tokens": 597772777.0, "step": 15745 }, { "epoch": 2.0030530466861722, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9954662322998047, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8715631365776062, "num_tokens": 597809276.0, "step": 15746 }, { "epoch": 2.0031802569647628, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.0860984325408936, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.875098466873169, "num_tokens": 597848956.0, "step": 15747 }, { "epoch": 2.0033074672433533, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.329807281494141, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8853569030761719, "num_tokens": 597890664.0, "step": 15748 }, { "epoch": 2.003434677521944, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3128528594970703, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8794947862625122, "num_tokens": 597931017.0, "step": 15749 }, { "epoch": 2.0035618878005343, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.565783977508545, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8931445479393005, "num_tokens": 597971552.0, "step": 15750 }, { "epoch": 2.003689098079125, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.770122528076172, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8715680837631226, "num_tokens": 598011554.0, "step": 15751 }, { "epoch": 2.0038163083577154, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.399359226226807, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8812046051025391, "num_tokens": 598050633.0, "step": 15752 }, { "epoch": 2.003943518636306, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.937314033508301, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.8886692523956299, "num_tokens": 598088022.0, "step": 15753 }, { "epoch": 2.0040707289148965, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.136112689971924, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8788415789604187, "num_tokens": 598126333.0, "step": 15754 }, { "epoch": 2.004197939193487, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.850799322128296, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8790778517723083, "num_tokens": 598161725.0, "step": 15755 }, { "epoch": 2.0043251494720775, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9347219467163086, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8846386671066284, "num_tokens": 598203158.0, "step": 15756 }, { "epoch": 2.004452359750668, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.7021167278289795, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.884144127368927, "num_tokens": 598241093.0, "step": 15757 }, { "epoch": 2.0045795700292586, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8257274627685547, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8831663727760315, "num_tokens": 598276389.0, "step": 15758 }, { "epoch": 2.0047067803078487, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5551273822784424, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8827986121177673, "num_tokens": 598316645.0, "step": 15759 }, { "epoch": 2.004833990586439, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.777609348297119, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8783392906188965, "num_tokens": 598361443.0, "step": 15760 }, { "epoch": 2.0049612008650297, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4352121353149414, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8831202983856201, "num_tokens": 598402934.0, "step": 15761 }, { "epoch": 2.0050884111436202, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.811347484588623, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8770132064819336, "num_tokens": 598436315.0, "step": 15762 }, { "epoch": 2.0052156214222108, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3395705223083496, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8884887099266052, "num_tokens": 598476562.0, "step": 15763 }, { "epoch": 2.0053428317008013, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.913120985031128, "learning_rate": 1e-05, "loss": 0.3865, "mean_token_accuracy": 0.8937563896179199, "num_tokens": 598515180.0, "step": 15764 }, { "epoch": 2.005470041979392, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2918128967285156, "learning_rate": 1e-05, "loss": 0.3783, "mean_token_accuracy": 0.8934374451637268, "num_tokens": 598554177.0, "step": 15765 }, { "epoch": 2.0055972522579824, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.079845428466797, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8789613246917725, "num_tokens": 598589091.0, "step": 15766 }, { "epoch": 2.005724462536573, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.368558406829834, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.8871057033538818, "num_tokens": 598624570.0, "step": 15767 }, { "epoch": 2.0058516728151634, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.170180320739746, "learning_rate": 1e-05, "loss": 0.3662, "mean_token_accuracy": 0.9013379812240601, "num_tokens": 598660350.0, "step": 15768 }, { "epoch": 2.005978883093754, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5397539138793945, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.868689775466919, "num_tokens": 598701823.0, "step": 15769 }, { "epoch": 2.0061060933723445, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.101160526275635, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.887296736240387, "num_tokens": 598738637.0, "step": 15770 }, { "epoch": 2.006233303650935, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.668447494506836, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8835747241973877, "num_tokens": 598773459.0, "step": 15771 }, { "epoch": 2.0063605139295255, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.895220994949341, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8870714902877808, "num_tokens": 598810098.0, "step": 15772 }, { "epoch": 2.006487724208116, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.685056209564209, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8730393648147583, "num_tokens": 598850701.0, "step": 15773 }, { "epoch": 2.0066149344867066, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.66926908493042, "learning_rate": 1e-05, "loss": 0.3745, "mean_token_accuracy": 0.8969875574111938, "num_tokens": 598887745.0, "step": 15774 }, { "epoch": 2.006742144765297, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6548359394073486, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8845805525779724, "num_tokens": 598928360.0, "step": 15775 }, { "epoch": 2.0068693550438876, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.49585223197937, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8847705721855164, "num_tokens": 598965628.0, "step": 15776 }, { "epoch": 2.006996565322478, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.642343044281006, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.880926251411438, "num_tokens": 599008445.0, "step": 15777 }, { "epoch": 2.0071237756010687, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3732337951660156, "learning_rate": 1e-05, "loss": 0.3708, "mean_token_accuracy": 0.8981918096542358, "num_tokens": 599041730.0, "step": 15778 }, { "epoch": 2.007250985879659, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.666210651397705, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8849091529846191, "num_tokens": 599082481.0, "step": 15779 }, { "epoch": 2.0073781961582498, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.275531768798828, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8804755210876465, "num_tokens": 599122503.0, "step": 15780 }, { "epoch": 2.0075054064368403, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7584517002105713, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8690512180328369, "num_tokens": 599158558.0, "step": 15781 }, { "epoch": 2.007632616715431, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2007975578308105, "learning_rate": 1e-05, "loss": 0.4001, "mean_token_accuracy": 0.8874286413192749, "num_tokens": 599196769.0, "step": 15782 }, { "epoch": 2.0077598269940213, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8989667892456055, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8847037553787231, "num_tokens": 599231050.0, "step": 15783 }, { "epoch": 2.0078870372726114, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1897644996643066, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.888302206993103, "num_tokens": 599271826.0, "step": 15784 }, { "epoch": 2.008014247551202, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.089512825012207, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8760570287704468, "num_tokens": 599311134.0, "step": 15785 }, { "epoch": 2.0081414578297925, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3012447357177734, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.885517954826355, "num_tokens": 599349240.0, "step": 15786 }, { "epoch": 2.008268668108383, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.21726655960083, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8840657472610474, "num_tokens": 599381624.0, "step": 15787 }, { "epoch": 2.0083958783869735, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4776875972747803, "learning_rate": 1e-05, "loss": 0.4084, "mean_token_accuracy": 0.8830416798591614, "num_tokens": 599421524.0, "step": 15788 }, { "epoch": 2.008523088665564, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.196229934692383, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8741937279701233, "num_tokens": 599459285.0, "step": 15789 }, { "epoch": 2.0086502989441546, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6328463554382324, "learning_rate": 1e-05, "loss": 0.3785, "mean_token_accuracy": 0.8962483406066895, "num_tokens": 599501682.0, "step": 15790 }, { "epoch": 2.008777509222745, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.032464981079102, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8928475379943848, "num_tokens": 599538518.0, "step": 15791 }, { "epoch": 2.0089047195013356, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.682011604309082, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8905687928199768, "num_tokens": 599572802.0, "step": 15792 }, { "epoch": 2.009031929779926, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.843606948852539, "learning_rate": 1e-05, "loss": 0.4045, "mean_token_accuracy": 0.8881881237030029, "num_tokens": 599614366.0, "step": 15793 }, { "epoch": 2.0091591400585167, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6321771144866943, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.881658136844635, "num_tokens": 599654814.0, "step": 15794 }, { "epoch": 2.0092863503371072, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6861627101898193, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8832721710205078, "num_tokens": 599689678.0, "step": 15795 }, { "epoch": 2.0094135606156978, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5817067623138428, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8778362274169922, "num_tokens": 599725932.0, "step": 15796 }, { "epoch": 2.0095407708942883, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5742650032043457, "learning_rate": 1e-05, "loss": 0.3863, "mean_token_accuracy": 0.89375901222229, "num_tokens": 599762041.0, "step": 15797 }, { "epoch": 2.009667981172879, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.507507801055908, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8870893716812134, "num_tokens": 599800706.0, "step": 15798 }, { "epoch": 2.0097951914514693, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5482370853424072, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8677265644073486, "num_tokens": 599833483.0, "step": 15799 }, { "epoch": 2.00992240173006, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4354207515716553, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8832062482833862, "num_tokens": 599872172.0, "step": 15800 }, { "epoch": 2.0100496120086504, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.532473087310791, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8768255710601807, "num_tokens": 599914149.0, "step": 15801 }, { "epoch": 2.010176822287241, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3762564659118652, "learning_rate": 1e-05, "loss": 0.4024, "mean_token_accuracy": 0.8842911124229431, "num_tokens": 599950603.0, "step": 15802 }, { "epoch": 2.0103040325658315, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.582953929901123, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8760393857955933, "num_tokens": 599991844.0, "step": 15803 }, { "epoch": 2.010431242844422, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2945656776428223, "learning_rate": 1e-05, "loss": 0.397, "mean_token_accuracy": 0.8872588872909546, "num_tokens": 600033303.0, "step": 15804 }, { "epoch": 2.0105584531230125, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6584343910217285, "learning_rate": 1e-05, "loss": 0.3725, "mean_token_accuracy": 0.8986252546310425, "num_tokens": 600065633.0, "step": 15805 }, { "epoch": 2.010685663401603, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.222879648208618, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8824172019958496, "num_tokens": 600104333.0, "step": 15806 }, { "epoch": 2.0108128736801936, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.805586338043213, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8769540786743164, "num_tokens": 600141329.0, "step": 15807 }, { "epoch": 2.0109400839587837, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.157028913497925, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8720157742500305, "num_tokens": 600184171.0, "step": 15808 }, { "epoch": 2.011067294237374, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.033839702606201, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8677083849906921, "num_tokens": 600220249.0, "step": 15809 }, { "epoch": 2.0111945045159647, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.211599349975586, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8786236047744751, "num_tokens": 600256015.0, "step": 15810 }, { "epoch": 2.0113217147945552, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2569580078125, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8824290633201599, "num_tokens": 600282893.0, "step": 15811 }, { "epoch": 2.0114489250731458, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.410083532333374, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8604274988174438, "num_tokens": 600320940.0, "step": 15812 }, { "epoch": 2.0115761353517363, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.314284324645996, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8649829626083374, "num_tokens": 600362054.0, "step": 15813 }, { "epoch": 2.011703345630327, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6784865856170654, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.867800772190094, "num_tokens": 600395080.0, "step": 15814 }, { "epoch": 2.0118305559089174, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.140021800994873, "learning_rate": 1e-05, "loss": 0.4044, "mean_token_accuracy": 0.890038013458252, "num_tokens": 600428060.0, "step": 15815 }, { "epoch": 2.011957766187508, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.721182346343994, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8880459070205688, "num_tokens": 600469833.0, "step": 15816 }, { "epoch": 2.0120849764660984, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.957871437072754, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.8855863809585571, "num_tokens": 600498006.0, "step": 15817 }, { "epoch": 2.012212186744689, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6856510639190674, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8840063810348511, "num_tokens": 600534744.0, "step": 15818 }, { "epoch": 2.0123393970232795, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7802493572235107, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8804110288619995, "num_tokens": 600570282.0, "step": 15819 }, { "epoch": 2.01246660730187, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6117005348205566, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8908678293228149, "num_tokens": 600603858.0, "step": 15820 }, { "epoch": 2.0125938175804605, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.717275857925415, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8688763976097107, "num_tokens": 600642078.0, "step": 15821 }, { "epoch": 2.012721027859051, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5396955013275146, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.877648115158081, "num_tokens": 600679424.0, "step": 15822 }, { "epoch": 2.0128482381376416, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.612244129180908, "learning_rate": 1e-05, "loss": 0.393, "mean_token_accuracy": 0.8899433612823486, "num_tokens": 600716310.0, "step": 15823 }, { "epoch": 2.012975448416232, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4390945434570312, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.870272696018219, "num_tokens": 600757031.0, "step": 15824 }, { "epoch": 2.0131026586948226, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.606963872909546, "learning_rate": 1e-05, "loss": 0.4125, "mean_token_accuracy": 0.8805878162384033, "num_tokens": 600795414.0, "step": 15825 }, { "epoch": 2.013229868973413, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.372349977493286, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.87038254737854, "num_tokens": 600829239.0, "step": 15826 }, { "epoch": 2.0133570792520037, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6284077167510986, "learning_rate": 1e-05, "loss": 0.3868, "mean_token_accuracy": 0.8905144929885864, "num_tokens": 600868155.0, "step": 15827 }, { "epoch": 2.013484289530594, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.295081853866577, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8754140734672546, "num_tokens": 600906797.0, "step": 15828 }, { "epoch": 2.0136114998091847, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7129480838775635, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8730559349060059, "num_tokens": 600948399.0, "step": 15829 }, { "epoch": 2.0137387100877753, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2266764640808105, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.877959132194519, "num_tokens": 600989682.0, "step": 15830 }, { "epoch": 2.013865920366366, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.867774486541748, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.877618134021759, "num_tokens": 601025387.0, "step": 15831 }, { "epoch": 2.0139931306449563, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.1966545581817627, "learning_rate": 1e-05, "loss": 0.3791, "mean_token_accuracy": 0.8933295607566833, "num_tokens": 601061124.0, "step": 15832 }, { "epoch": 2.0141203409235464, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.049220085144043, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8766969442367554, "num_tokens": 601105789.0, "step": 15833 }, { "epoch": 2.014247551202137, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2982962131500244, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.867324709892273, "num_tokens": 601138697.0, "step": 15834 }, { "epoch": 2.0143747614807275, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.172471523284912, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.8923647403717041, "num_tokens": 601172875.0, "step": 15835 }, { "epoch": 2.014501971759318, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4537508487701416, "learning_rate": 1e-05, "loss": 0.3842, "mean_token_accuracy": 0.8951526880264282, "num_tokens": 601206303.0, "step": 15836 }, { "epoch": 2.0146291820379085, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.149329662322998, "learning_rate": 1e-05, "loss": 0.3829, "mean_token_accuracy": 0.8973218202590942, "num_tokens": 601243879.0, "step": 15837 }, { "epoch": 2.014756392316499, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6101183891296387, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.883208155632019, "num_tokens": 601282639.0, "step": 15838 }, { "epoch": 2.0148836025950896, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.9884257316589355, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8789522647857666, "num_tokens": 601326204.0, "step": 15839 }, { "epoch": 2.01501081287368, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6592586040496826, "learning_rate": 1e-05, "loss": 0.4026, "mean_token_accuracy": 0.8865939378738403, "num_tokens": 601358936.0, "step": 15840 }, { "epoch": 2.0151380231522706, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.785565137863159, "learning_rate": 1e-05, "loss": 0.3852, "mean_token_accuracy": 0.8927961587905884, "num_tokens": 601398253.0, "step": 15841 }, { "epoch": 2.015265233430861, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6204421520233154, "learning_rate": 1e-05, "loss": 0.3424, "mean_token_accuracy": 0.9055854082107544, "num_tokens": 601434806.0, "step": 15842 }, { "epoch": 2.0153924437094517, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.621164560317993, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.8847519159317017, "num_tokens": 601471834.0, "step": 15843 }, { "epoch": 2.0155196539880422, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5800652503967285, "learning_rate": 1e-05, "loss": 0.3919, "mean_token_accuracy": 0.8915905952453613, "num_tokens": 601511256.0, "step": 15844 }, { "epoch": 2.0156468642666328, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.508920907974243, "learning_rate": 1e-05, "loss": 0.3794, "mean_token_accuracy": 0.8929089903831482, "num_tokens": 601547434.0, "step": 15845 }, { "epoch": 2.0157740745452233, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5453052520751953, "learning_rate": 1e-05, "loss": 0.3896, "mean_token_accuracy": 0.8894805312156677, "num_tokens": 601581375.0, "step": 15846 }, { "epoch": 2.015901284823814, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.45066499710083, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8774706721305847, "num_tokens": 601615067.0, "step": 15847 }, { "epoch": 2.0160284951024043, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5043962001800537, "learning_rate": 1e-05, "loss": 0.3505, "mean_token_accuracy": 0.8991764783859253, "num_tokens": 601654168.0, "step": 15848 }, { "epoch": 2.016155705380995, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4156124591827393, "learning_rate": 1e-05, "loss": 0.3784, "mean_token_accuracy": 0.8940605521202087, "num_tokens": 601687113.0, "step": 15849 }, { "epoch": 2.0162829156595854, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4928503036499023, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.874570906162262, "num_tokens": 601723993.0, "step": 15850 }, { "epoch": 2.016410125938176, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.398669958114624, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8658214807510376, "num_tokens": 601758522.0, "step": 15851 }, { "epoch": 2.0165373362167665, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.478184223175049, "learning_rate": 1e-05, "loss": 0.3504, "mean_token_accuracy": 0.9037771821022034, "num_tokens": 601794221.0, "step": 15852 }, { "epoch": 2.016664546495357, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3527073860168457, "learning_rate": 1e-05, "loss": 0.3863, "mean_token_accuracy": 0.8905977010726929, "num_tokens": 601830430.0, "step": 15853 }, { "epoch": 2.0167917567739475, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.486117124557495, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8834946155548096, "num_tokens": 601871834.0, "step": 15854 }, { "epoch": 2.016918967052538, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3283963203430176, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8776794672012329, "num_tokens": 601911432.0, "step": 15855 }, { "epoch": 2.0170461773311286, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.527189016342163, "learning_rate": 1e-05, "loss": 0.383, "mean_token_accuracy": 0.8920127749443054, "num_tokens": 601955592.0, "step": 15856 }, { "epoch": 2.0171733876097186, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2908666133880615, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8765766620635986, "num_tokens": 601996843.0, "step": 15857 }, { "epoch": 2.017300597888309, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5970449447631836, "learning_rate": 1e-05, "loss": 0.3911, "mean_token_accuracy": 0.8910675048828125, "num_tokens": 602034963.0, "step": 15858 }, { "epoch": 2.0174278081668997, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.230597496032715, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8812731504440308, "num_tokens": 602073460.0, "step": 15859 }, { "epoch": 2.0175550184454902, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7093772888183594, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8675346374511719, "num_tokens": 602117442.0, "step": 15860 }, { "epoch": 2.0176822287240808, "ewc_loss": 0.07373046875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1350224018096924, "learning_rate": 1e-05, "loss": 0.3862, "mean_token_accuracy": 0.8876146078109741, "num_tokens": 602161253.0, "step": 15861 }, { "epoch": 2.0178094390026713, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9086453914642334, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8741228580474854, "num_tokens": 602204324.0, "step": 15862 }, { "epoch": 2.017936649281262, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.104750871658325, "learning_rate": 1e-05, "loss": 0.3872, "mean_token_accuracy": 0.8868849277496338, "num_tokens": 602236335.0, "step": 15863 }, { "epoch": 2.0180638595598523, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.226129531860352, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8766312599182129, "num_tokens": 602271302.0, "step": 15864 }, { "epoch": 2.018191069838443, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.2351365089416504, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8883389830589294, "num_tokens": 602304103.0, "step": 15865 }, { "epoch": 2.0183182801170334, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.478147506713867, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8843768835067749, "num_tokens": 602343293.0, "step": 15866 }, { "epoch": 2.018445490395624, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.596668004989624, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8757498264312744, "num_tokens": 602383650.0, "step": 15867 }, { "epoch": 2.0185727006742145, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.459901809692383, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8920998573303223, "num_tokens": 602423436.0, "step": 15868 }, { "epoch": 2.018699910952805, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.887662649154663, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8830713033676147, "num_tokens": 602460909.0, "step": 15869 }, { "epoch": 2.0188271212313955, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.205645561218262, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8798892498016357, "num_tokens": 602499610.0, "step": 15870 }, { "epoch": 2.018954331509986, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.8619959354400635, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8777354955673218, "num_tokens": 602540783.0, "step": 15871 }, { "epoch": 2.0190815417885766, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.989015579223633, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8762627840042114, "num_tokens": 602579205.0, "step": 15872 }, { "epoch": 2.019208752067167, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7394039630889893, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8812350034713745, "num_tokens": 602613720.0, "step": 15873 }, { "epoch": 2.0193359623457576, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8453683853149414, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8881031274795532, "num_tokens": 602651954.0, "step": 15874 }, { "epoch": 2.019463172624348, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000278472900390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6169636249542236, "learning_rate": 1e-05, "loss": 0.3869, "mean_token_accuracy": 0.8941680788993835, "num_tokens": 602685553.0, "step": 15875 }, { "epoch": 2.0195903829029387, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7464399337768555, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8764764070510864, "num_tokens": 602727854.0, "step": 15876 }, { "epoch": 2.019717593181529, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.497476816177368, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8840329051017761, "num_tokens": 602765993.0, "step": 15877 }, { "epoch": 2.0198448034601197, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.733617067337036, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.8933507800102234, "num_tokens": 602807450.0, "step": 15878 }, { "epoch": 2.0199720137387103, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4025673866271973, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8656595945358276, "num_tokens": 602848689.0, "step": 15879 }, { "epoch": 2.020099224017301, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.769012451171875, "learning_rate": 1e-05, "loss": 0.3928, "mean_token_accuracy": 0.8951394557952881, "num_tokens": 602888010.0, "step": 15880 }, { "epoch": 2.0202264342958913, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3172435760498047, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8672329187393188, "num_tokens": 602923345.0, "step": 15881 }, { "epoch": 2.0203536445744814, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8687143325805664, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8842698931694031, "num_tokens": 602967320.0, "step": 15882 }, { "epoch": 2.020480854853072, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.260948657989502, "learning_rate": 1e-05, "loss": 0.3686, "mean_token_accuracy": 0.8928602337837219, "num_tokens": 603005177.0, "step": 15883 }, { "epoch": 2.0206080651316625, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.041367530822754, "learning_rate": 1e-05, "loss": 0.3876, "mean_token_accuracy": 0.8948395252227783, "num_tokens": 603039562.0, "step": 15884 }, { "epoch": 2.020735275410253, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3195133209228516, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8889321684837341, "num_tokens": 603079075.0, "step": 15885 }, { "epoch": 2.0208624856888435, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.17249059677124, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8817068338394165, "num_tokens": 603118077.0, "step": 15886 }, { "epoch": 2.020989695967434, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5116424560546875, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8748854994773865, "num_tokens": 603156921.0, "step": 15887 }, { "epoch": 2.0211169062460246, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.104724884033203, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8837898969650269, "num_tokens": 603202062.0, "step": 15888 }, { "epoch": 2.021244116524615, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6433041095733643, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8900634050369263, "num_tokens": 603236240.0, "step": 15889 }, { "epoch": 2.0213713268032056, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9210400581359863, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8784713745117188, "num_tokens": 603277376.0, "step": 15890 }, { "epoch": 2.021498537081796, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6671793460845947, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8829793930053711, "num_tokens": 603314740.0, "step": 15891 }, { "epoch": 2.0216257473603867, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7056996822357178, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8849418759346008, "num_tokens": 603354992.0, "step": 15892 }, { "epoch": 2.021752957638977, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6564629077911377, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8783628344535828, "num_tokens": 603387271.0, "step": 15893 }, { "epoch": 2.0218801679175677, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.52425217628479, "learning_rate": 1e-05, "loss": 0.3919, "mean_token_accuracy": 0.8910921812057495, "num_tokens": 603427092.0, "step": 15894 }, { "epoch": 2.0220073781961583, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6330392360687256, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8717583417892456, "num_tokens": 603465188.0, "step": 15895 }, { "epoch": 2.022134588474749, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.413299798965454, "learning_rate": 1e-05, "loss": 0.4125, "mean_token_accuracy": 0.8798158168792725, "num_tokens": 603499290.0, "step": 15896 }, { "epoch": 2.0222617987533393, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6520543098449707, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8857632875442505, "num_tokens": 603530594.0, "step": 15897 }, { "epoch": 2.02238900903193, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.29606556892395, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8843213319778442, "num_tokens": 603569337.0, "step": 15898 }, { "epoch": 2.0225162193105204, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.718964099884033, "learning_rate": 1e-05, "loss": 0.4065, "mean_token_accuracy": 0.8850136995315552, "num_tokens": 603609704.0, "step": 15899 }, { "epoch": 2.022643429589111, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.231785297393799, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8822757005691528, "num_tokens": 603647623.0, "step": 15900 }, { "epoch": 2.0227706398677014, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.843824625015259, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8862691521644592, "num_tokens": 603682237.0, "step": 15901 }, { "epoch": 2.022897850146292, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1855905055999756, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8772178292274475, "num_tokens": 603720348.0, "step": 15902 }, { "epoch": 2.0230250604248825, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.024155616760254, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8832987546920776, "num_tokens": 603759857.0, "step": 15903 }, { "epoch": 2.023152270703473, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2431881427764893, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8782405853271484, "num_tokens": 603796970.0, "step": 15904 }, { "epoch": 2.0232794809820636, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.181761264801025, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8797736763954163, "num_tokens": 603838385.0, "step": 15905 }, { "epoch": 2.0234066912606536, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.397442579269409, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8811354637145996, "num_tokens": 603884262.0, "step": 15906 }, { "epoch": 2.023533901539244, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2293829917907715, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8775357604026794, "num_tokens": 603925013.0, "step": 15907 }, { "epoch": 2.0236611118178347, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.600163698196411, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8722436428070068, "num_tokens": 603971173.0, "step": 15908 }, { "epoch": 2.0237883220964252, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.104974746704102, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8755188584327698, "num_tokens": 604009091.0, "step": 15909 }, { "epoch": 2.0239155323750158, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6836116313934326, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8821364045143127, "num_tokens": 604045163.0, "step": 15910 }, { "epoch": 2.0240427426536063, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.895115852355957, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.886206865310669, "num_tokens": 604091379.0, "step": 15911 }, { "epoch": 2.024169952932197, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.645021915435791, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8800984025001526, "num_tokens": 604132159.0, "step": 15912 }, { "epoch": 2.0242971632107873, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7478158473968506, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8728935122489929, "num_tokens": 604169689.0, "step": 15913 }, { "epoch": 2.024424373489378, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5887677669525146, "learning_rate": 1e-05, "loss": 0.36, "mean_token_accuracy": 0.9009727239608765, "num_tokens": 604201913.0, "step": 15914 }, { "epoch": 2.0245515837679684, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6378259658813477, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8747247457504272, "num_tokens": 604238275.0, "step": 15915 }, { "epoch": 2.024678794046559, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5013296604156494, "learning_rate": 1e-05, "loss": 0.3879, "mean_token_accuracy": 0.889934241771698, "num_tokens": 604279862.0, "step": 15916 }, { "epoch": 2.0248060043251495, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5734341144561768, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8805594444274902, "num_tokens": 604319091.0, "step": 15917 }, { "epoch": 2.02493321460374, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.439913034439087, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8745479583740234, "num_tokens": 604363614.0, "step": 15918 }, { "epoch": 2.0250604248823305, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5866379737854004, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8801321983337402, "num_tokens": 604394422.0, "step": 15919 }, { "epoch": 2.025187635160921, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.370954990386963, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8789730072021484, "num_tokens": 604432844.0, "step": 15920 }, { "epoch": 2.0253148454395116, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6065235137939453, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8823351263999939, "num_tokens": 604472074.0, "step": 15921 }, { "epoch": 2.025442055718102, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2916722297668457, "learning_rate": 1e-05, "loss": 0.3549, "mean_token_accuracy": 0.8982524871826172, "num_tokens": 604507827.0, "step": 15922 }, { "epoch": 2.0255692659966926, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6858420372009277, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8773394823074341, "num_tokens": 604548600.0, "step": 15923 }, { "epoch": 2.025696476275283, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2525248527526855, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8673387169837952, "num_tokens": 604582701.0, "step": 15924 }, { "epoch": 2.0258236865538737, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7992703914642334, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8725161552429199, "num_tokens": 604624089.0, "step": 15925 }, { "epoch": 2.025950896832464, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.207726240158081, "learning_rate": 1e-05, "loss": 0.3844, "mean_token_accuracy": 0.891148030757904, "num_tokens": 604655764.0, "step": 15926 }, { "epoch": 2.0260781071110547, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9700686931610107, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8911957144737244, "num_tokens": 604688000.0, "step": 15927 }, { "epoch": 2.0262053173896453, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2313590049743652, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8759620189666748, "num_tokens": 604724602.0, "step": 15928 }, { "epoch": 2.026332527668236, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.123233795166016, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8864396810531616, "num_tokens": 604761794.0, "step": 15929 }, { "epoch": 2.0264597379468263, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.365145206451416, "learning_rate": 1e-05, "loss": 0.3716, "mean_token_accuracy": 0.8963570594787598, "num_tokens": 604800239.0, "step": 15930 }, { "epoch": 2.0265869482254164, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1743011474609375, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8814002871513367, "num_tokens": 604837143.0, "step": 15931 }, { "epoch": 2.026714158504007, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5551657676696777, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8817803859710693, "num_tokens": 604874074.0, "step": 15932 }, { "epoch": 2.0268413687825975, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.056101322174072, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8760415315628052, "num_tokens": 604911639.0, "step": 15933 }, { "epoch": 2.026968579061188, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.648594856262207, "learning_rate": 1e-05, "loss": 0.3922, "mean_token_accuracy": 0.8937216401100159, "num_tokens": 604944963.0, "step": 15934 }, { "epoch": 2.0270957893397785, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8483734130859375, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8829967975616455, "num_tokens": 604986048.0, "step": 15935 }, { "epoch": 2.027222999618369, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.634658098220825, "learning_rate": 1e-05, "loss": 0.3982, "mean_token_accuracy": 0.8877763748168945, "num_tokens": 605023995.0, "step": 15936 }, { "epoch": 2.0273502098969596, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6563591957092285, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.8882648944854736, "num_tokens": 605064612.0, "step": 15937 }, { "epoch": 2.02747742017555, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6113553047180176, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8762152791023254, "num_tokens": 605099332.0, "step": 15938 }, { "epoch": 2.0276046304541406, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.51456880569458, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8833069205284119, "num_tokens": 605137616.0, "step": 15939 }, { "epoch": 2.027731840732731, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.578949213027954, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8800950050354004, "num_tokens": 605177974.0, "step": 15940 }, { "epoch": 2.0278590510113217, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4244017601013184, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8801885843276978, "num_tokens": 605219156.0, "step": 15941 }, { "epoch": 2.027986261289912, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5682642459869385, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8842309713363647, "num_tokens": 605261131.0, "step": 15942 }, { "epoch": 2.0281134715685027, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3572592735290527, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8777878880500793, "num_tokens": 605303640.0, "step": 15943 }, { "epoch": 2.0282406818470933, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5854337215423584, "learning_rate": 1e-05, "loss": 0.3807, "mean_token_accuracy": 0.8924393653869629, "num_tokens": 605341823.0, "step": 15944 }, { "epoch": 2.028367892125684, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3010222911834717, "learning_rate": 1e-05, "loss": 0.3824, "mean_token_accuracy": 0.892905056476593, "num_tokens": 605380282.0, "step": 15945 }, { "epoch": 2.0284951024042743, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.644747495651245, "learning_rate": 1e-05, "loss": 0.3921, "mean_token_accuracy": 0.888904869556427, "num_tokens": 605415077.0, "step": 15946 }, { "epoch": 2.028622312682865, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.238517999649048, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8824634552001953, "num_tokens": 605453805.0, "step": 15947 }, { "epoch": 2.0287495229614554, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.763864517211914, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8814182877540588, "num_tokens": 605488740.0, "step": 15948 }, { "epoch": 2.028876733240046, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1751694679260254, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8737756013870239, "num_tokens": 605529155.0, "step": 15949 }, { "epoch": 2.0290039435186364, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9296751022338867, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.882239818572998, "num_tokens": 605567610.0, "step": 15950 }, { "epoch": 2.029131153797227, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1653647422790527, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8788855075836182, "num_tokens": 605609485.0, "step": 15951 }, { "epoch": 2.0292583640758175, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.154569625854492, "learning_rate": 1e-05, "loss": 0.3976, "mean_token_accuracy": 0.8925430774688721, "num_tokens": 605644718.0, "step": 15952 }, { "epoch": 2.029385574354408, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.272876739501953, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8825512528419495, "num_tokens": 605686893.0, "step": 15953 }, { "epoch": 2.0295127846329986, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.338442802429199, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8830721974372864, "num_tokens": 605724007.0, "step": 15954 }, { "epoch": 2.0296399949115886, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.5664725303649902, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8741778135299683, "num_tokens": 605761213.0, "step": 15955 }, { "epoch": 2.029767205190179, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.259520530700684, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8841930627822876, "num_tokens": 605803274.0, "step": 15956 }, { "epoch": 2.0298944154687697, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7303643226623535, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8799763917922974, "num_tokens": 605838326.0, "step": 15957 }, { "epoch": 2.0300216257473602, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.06243371963501, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8851485848426819, "num_tokens": 605878882.0, "step": 15958 }, { "epoch": 2.0301488360259508, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.724153518676758, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.887102484703064, "num_tokens": 605913574.0, "step": 15959 }, { "epoch": 2.0302760463045413, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.883443593978882, "learning_rate": 1e-05, "loss": 0.3926, "mean_token_accuracy": 0.8902068138122559, "num_tokens": 605944845.0, "step": 15960 }, { "epoch": 2.030403256583132, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.656181573867798, "learning_rate": 1e-05, "loss": 0.4856, "mean_token_accuracy": 0.8607428669929504, "num_tokens": 605982604.0, "step": 15961 }, { "epoch": 2.0305304668617223, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.72184157371521, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.877547025680542, "num_tokens": 606023506.0, "step": 15962 }, { "epoch": 2.030657677140313, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.57283878326416, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8728654980659485, "num_tokens": 606063319.0, "step": 15963 }, { "epoch": 2.0307848874189034, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.626204490661621, "learning_rate": 1e-05, "loss": 0.3745, "mean_token_accuracy": 0.8968309164047241, "num_tokens": 606103393.0, "step": 15964 }, { "epoch": 2.030912097697494, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.488053798675537, "learning_rate": 1e-05, "loss": 0.3862, "mean_token_accuracy": 0.8924559354782104, "num_tokens": 606144191.0, "step": 15965 }, { "epoch": 2.0310393079760845, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5987274646759033, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8860504627227783, "num_tokens": 606183219.0, "step": 15966 }, { "epoch": 2.031166518254675, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.414445400238037, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8788139224052429, "num_tokens": 606221429.0, "step": 15967 }, { "epoch": 2.0312937285332655, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.607940673828125, "learning_rate": 1e-05, "loss": 0.3944, "mean_token_accuracy": 0.8873682618141174, "num_tokens": 606260810.0, "step": 15968 }, { "epoch": 2.031420938811856, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3306331634521484, "learning_rate": 1e-05, "loss": 0.39, "mean_token_accuracy": 0.8883922100067139, "num_tokens": 606299602.0, "step": 15969 }, { "epoch": 2.0315481490904466, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.659088373184204, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8864078521728516, "num_tokens": 606341130.0, "step": 15970 }, { "epoch": 2.031675359369037, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.266512870788574, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.886654257774353, "num_tokens": 606378341.0, "step": 15971 }, { "epoch": 2.0318025696476276, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.798046588897705, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8649874329566956, "num_tokens": 606413799.0, "step": 15972 }, { "epoch": 2.031929779926218, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.212010145187378, "learning_rate": 1e-05, "loss": 0.3802, "mean_token_accuracy": 0.8902314901351929, "num_tokens": 606447428.0, "step": 15973 }, { "epoch": 2.0320569902048087, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.966546058654785, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8819055557250977, "num_tokens": 606481652.0, "step": 15974 }, { "epoch": 2.032184200483399, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.207230567932129, "learning_rate": 1e-05, "loss": 0.36, "mean_token_accuracy": 0.8987140655517578, "num_tokens": 606519279.0, "step": 15975 }, { "epoch": 2.0323114107619897, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.163018226623535, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8827077150344849, "num_tokens": 606556399.0, "step": 15976 }, { "epoch": 2.0324386210405803, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3776233196258545, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8676812648773193, "num_tokens": 606596807.0, "step": 15977 }, { "epoch": 2.032565831319171, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.237023830413818, "learning_rate": 1e-05, "loss": 0.3871, "mean_token_accuracy": 0.8954153060913086, "num_tokens": 606629231.0, "step": 15978 }, { "epoch": 2.032693041597761, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.587280750274658, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8778426647186279, "num_tokens": 606669042.0, "step": 15979 }, { "epoch": 2.0328202518763514, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.117098331451416, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8727453947067261, "num_tokens": 606705550.0, "step": 15980 }, { "epoch": 2.032947462154942, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6903696060180664, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8760936260223389, "num_tokens": 606739866.0, "step": 15981 }, { "epoch": 2.0330746724335325, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.8957746028900146, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.881855845451355, "num_tokens": 606782923.0, "step": 15982 }, { "epoch": 2.033201882712123, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.668153762817383, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8864395618438721, "num_tokens": 606822831.0, "step": 15983 }, { "epoch": 2.0333290929907135, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.705244541168213, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8860878944396973, "num_tokens": 606861129.0, "step": 15984 }, { "epoch": 2.033456303269304, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.628019332885742, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8666064739227295, "num_tokens": 606902744.0, "step": 15985 }, { "epoch": 2.0335835135478946, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.57442045211792, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.874424934387207, "num_tokens": 606935137.0, "step": 15986 }, { "epoch": 2.033710723826485, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5655853748321533, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.886829137802124, "num_tokens": 606972215.0, "step": 15987 }, { "epoch": 2.0338379341050756, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.494676351547241, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.884208619594574, "num_tokens": 607004489.0, "step": 15988 }, { "epoch": 2.033965144383666, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.529831647872925, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8779640197753906, "num_tokens": 607039457.0, "step": 15989 }, { "epoch": 2.0340923546622567, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4256136417388916, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.876437783241272, "num_tokens": 607084863.0, "step": 15990 }, { "epoch": 2.034219564940847, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5114564895629883, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.872063159942627, "num_tokens": 607125013.0, "step": 15991 }, { "epoch": 2.0343467752194377, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.384892225265503, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8843075037002563, "num_tokens": 607160501.0, "step": 15992 }, { "epoch": 2.0344739854980283, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.501371145248413, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8745619654655457, "num_tokens": 607204621.0, "step": 15993 }, { "epoch": 2.034601195776619, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3448071479797363, "learning_rate": 1e-05, "loss": 0.3955, "mean_token_accuracy": 0.8898059129714966, "num_tokens": 607245904.0, "step": 15994 }, { "epoch": 2.0347284060552093, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5155723094940186, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8839328289031982, "num_tokens": 607286907.0, "step": 15995 }, { "epoch": 2.0348556163338, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3157317638397217, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8700253963470459, "num_tokens": 607325990.0, "step": 15996 }, { "epoch": 2.0349828266123904, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.57085919380188, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8748259544372559, "num_tokens": 607363000.0, "step": 15997 }, { "epoch": 2.035110036890981, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2449228763580322, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8811771869659424, "num_tokens": 607405913.0, "step": 15998 }, { "epoch": 2.0352372471695714, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.684462070465088, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8829782009124756, "num_tokens": 607440242.0, "step": 15999 }, { "epoch": 2.035364457448162, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1590774059295654, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.882676899433136, "num_tokens": 607485705.0, "step": 16000 }, { "epoch": 2.0354916677267525, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8723695278167725, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8742011189460754, "num_tokens": 607524026.0, "step": 16001 }, { "epoch": 2.035618878005343, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1086747646331787, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.876230001449585, "num_tokens": 607562925.0, "step": 16002 }, { "epoch": 2.0357460882839336, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.159903049468994, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8715739250183105, "num_tokens": 607598126.0, "step": 16003 }, { "epoch": 2.0358732985625236, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.1925647258758545, "learning_rate": 1e-05, "loss": 0.3898, "mean_token_accuracy": 0.8934895992279053, "num_tokens": 607631911.0, "step": 16004 }, { "epoch": 2.036000508841114, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.437672138214111, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8710391521453857, "num_tokens": 607669817.0, "step": 16005 }, { "epoch": 2.0361277191197047, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.5364301204681396, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.8851569890975952, "num_tokens": 607710713.0, "step": 16006 }, { "epoch": 2.036254929398295, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.446321487426758, "learning_rate": 1e-05, "loss": 0.3853, "mean_token_accuracy": 0.8976091146469116, "num_tokens": 607751540.0, "step": 16007 }, { "epoch": 2.0363821396768857, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.839322566986084, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8821587562561035, "num_tokens": 607792831.0, "step": 16008 }, { "epoch": 2.0365093499554763, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.229916095733643, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8810263872146606, "num_tokens": 607824175.0, "step": 16009 }, { "epoch": 2.036636560234067, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.864596128463745, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.871372640132904, "num_tokens": 607864302.0, "step": 16010 }, { "epoch": 2.0367637705126573, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.988874673843384, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8754978775978088, "num_tokens": 607903231.0, "step": 16011 }, { "epoch": 2.036890980791248, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002803802490234375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7452547550201416, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8870182037353516, "num_tokens": 607942795.0, "step": 16012 }, { "epoch": 2.0370181910698384, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8271307945251465, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8800491094589233, "num_tokens": 607978034.0, "step": 16013 }, { "epoch": 2.037145401348429, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.627372980117798, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8907991647720337, "num_tokens": 608014260.0, "step": 16014 }, { "epoch": 2.0372726116270194, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7335457801818848, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8720546960830688, "num_tokens": 608052637.0, "step": 16015 }, { "epoch": 2.03739982190561, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5197527408599854, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8768310546875, "num_tokens": 608085877.0, "step": 16016 }, { "epoch": 2.0375270321842005, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6995272636413574, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8802309036254883, "num_tokens": 608125059.0, "step": 16017 }, { "epoch": 2.037654242462791, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.408979892730713, "learning_rate": 1e-05, "loss": 0.3977, "mean_token_accuracy": 0.8881188035011292, "num_tokens": 608161357.0, "step": 16018 }, { "epoch": 2.0377814527413816, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.733215570449829, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8801928758621216, "num_tokens": 608196534.0, "step": 16019 }, { "epoch": 2.037908663019972, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3060052394866943, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8800752758979797, "num_tokens": 608237130.0, "step": 16020 }, { "epoch": 2.0380358732985626, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.835672616958618, "learning_rate": 1e-05, "loss": 0.3761, "mean_token_accuracy": 0.8984917402267456, "num_tokens": 608279064.0, "step": 16021 }, { "epoch": 2.038163083577153, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2468788623809814, "learning_rate": 1e-05, "loss": 0.3806, "mean_token_accuracy": 0.891409158706665, "num_tokens": 608316019.0, "step": 16022 }, { "epoch": 2.0382902938557437, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.003720283508301, "learning_rate": 1e-05, "loss": 0.3902, "mean_token_accuracy": 0.8934849500656128, "num_tokens": 608357199.0, "step": 16023 }, { "epoch": 2.038417504134334, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2867484092712402, "learning_rate": 1e-05, "loss": 0.4063, "mean_token_accuracy": 0.8807528018951416, "num_tokens": 608392149.0, "step": 16024 }, { "epoch": 2.0385447144129247, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.157249927520752, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.890679121017456, "num_tokens": 608431200.0, "step": 16025 }, { "epoch": 2.0386719246915153, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.436696767807007, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8872990608215332, "num_tokens": 608471832.0, "step": 16026 }, { "epoch": 2.038799134970106, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.194490432739258, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8726110458374023, "num_tokens": 608508362.0, "step": 16027 }, { "epoch": 2.0389263452486963, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6324095726013184, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8765188455581665, "num_tokens": 608547988.0, "step": 16028 }, { "epoch": 2.0390535555272864, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.014978885650635, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8909649848937988, "num_tokens": 608580448.0, "step": 16029 }, { "epoch": 2.039180765805877, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.7025885581970215, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8730257749557495, "num_tokens": 608618573.0, "step": 16030 }, { "epoch": 2.0393079760844675, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7729718685150146, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8797330856323242, "num_tokens": 608658161.0, "step": 16031 }, { "epoch": 2.039435186363058, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.680943012237549, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8750442266464233, "num_tokens": 608696599.0, "step": 16032 }, { "epoch": 2.0395623966416485, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.55910325050354, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8895442485809326, "num_tokens": 608738347.0, "step": 16033 }, { "epoch": 2.039689606920239, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.655696392059326, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8853941559791565, "num_tokens": 608779483.0, "step": 16034 }, { "epoch": 2.0398168171988296, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.411240577697754, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8796048164367676, "num_tokens": 608821369.0, "step": 16035 }, { "epoch": 2.03994402747742, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6782665252685547, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8808883428573608, "num_tokens": 608854749.0, "step": 16036 }, { "epoch": 2.0400712377560106, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.305259943008423, "learning_rate": 1e-05, "loss": 0.3831, "mean_token_accuracy": 0.8899228572845459, "num_tokens": 608893692.0, "step": 16037 }, { "epoch": 2.040198448034601, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.736738681793213, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8776349425315857, "num_tokens": 608930721.0, "step": 16038 }, { "epoch": 2.0403256583131917, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.240194797515869, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8741523027420044, "num_tokens": 608967022.0, "step": 16039 }, { "epoch": 2.040452868591782, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.853543281555176, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8642354011535645, "num_tokens": 609009995.0, "step": 16040 }, { "epoch": 2.0405800788703727, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.210728883743286, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8780367374420166, "num_tokens": 609044636.0, "step": 16041 }, { "epoch": 2.0407072891489633, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.008918285369873, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.8887990117073059, "num_tokens": 609076328.0, "step": 16042 }, { "epoch": 2.040834499427554, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.250131845474243, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8645890951156616, "num_tokens": 609118955.0, "step": 16043 }, { "epoch": 2.0409617097061443, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.156775951385498, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8702467679977417, "num_tokens": 609157958.0, "step": 16044 }, { "epoch": 2.041088919984735, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3962953090667725, "learning_rate": 1e-05, "loss": 0.4001, "mean_token_accuracy": 0.8860989809036255, "num_tokens": 609197636.0, "step": 16045 }, { "epoch": 2.0412161302633254, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.202542781829834, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8643280267715454, "num_tokens": 609239648.0, "step": 16046 }, { "epoch": 2.041343340541916, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5736215114593506, "learning_rate": 1e-05, "loss": 0.3872, "mean_token_accuracy": 0.8931553363800049, "num_tokens": 609281012.0, "step": 16047 }, { "epoch": 2.0414705508205064, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.086604595184326, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8811042308807373, "num_tokens": 609323775.0, "step": 16048 }, { "epoch": 2.041597761099097, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6565022468566895, "learning_rate": 1e-05, "loss": 0.3844, "mean_token_accuracy": 0.8950635194778442, "num_tokens": 609358960.0, "step": 16049 }, { "epoch": 2.0417249713776875, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.917389154434204, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8710073828697205, "num_tokens": 609395447.0, "step": 16050 }, { "epoch": 2.041852181656278, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.63108229637146, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.8947800993919373, "num_tokens": 609429811.0, "step": 16051 }, { "epoch": 2.0419793919348685, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7529308795928955, "learning_rate": 1e-05, "loss": 0.4063, "mean_token_accuracy": 0.8866859078407288, "num_tokens": 609466104.0, "step": 16052 }, { "epoch": 2.0421066022134586, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.576469898223877, "learning_rate": 1e-05, "loss": 0.385, "mean_token_accuracy": 0.8904587626457214, "num_tokens": 609503513.0, "step": 16053 }, { "epoch": 2.042233812492049, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.631622552871704, "learning_rate": 1e-05, "loss": 0.3842, "mean_token_accuracy": 0.8923308849334717, "num_tokens": 609546899.0, "step": 16054 }, { "epoch": 2.0423610227706397, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5159099102020264, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8757030367851257, "num_tokens": 609589297.0, "step": 16055 }, { "epoch": 2.04248823304923, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.58087158203125, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.8897701501846313, "num_tokens": 609624839.0, "step": 16056 }, { "epoch": 2.0426154433278207, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.456491470336914, "learning_rate": 1e-05, "loss": 0.3857, "mean_token_accuracy": 0.8894600868225098, "num_tokens": 609661708.0, "step": 16057 }, { "epoch": 2.0427426536064113, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.559905529022217, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8704513311386108, "num_tokens": 609704133.0, "step": 16058 }, { "epoch": 2.042869863885002, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3960137367248535, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8819146156311035, "num_tokens": 609740743.0, "step": 16059 }, { "epoch": 2.0429970741635923, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5771257877349854, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8726788759231567, "num_tokens": 609777520.0, "step": 16060 }, { "epoch": 2.043124284442183, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.331249475479126, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8829357028007507, "num_tokens": 609812974.0, "step": 16061 }, { "epoch": 2.0432514947207734, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.629910707473755, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8871936798095703, "num_tokens": 609853287.0, "step": 16062 }, { "epoch": 2.043378704999364, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.260368585586548, "learning_rate": 1e-05, "loss": 0.3721, "mean_token_accuracy": 0.8928025960922241, "num_tokens": 609888768.0, "step": 16063 }, { "epoch": 2.0435059152779544, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.737853765487671, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8884449005126953, "num_tokens": 609927976.0, "step": 16064 }, { "epoch": 2.043633125556545, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1931920051574707, "learning_rate": 1e-05, "loss": 0.3841, "mean_token_accuracy": 0.888384222984314, "num_tokens": 609964404.0, "step": 16065 }, { "epoch": 2.0437603358351355, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.915942668914795, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8816094398498535, "num_tokens": 610005188.0, "step": 16066 }, { "epoch": 2.043887546113726, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2047677040100098, "learning_rate": 1e-05, "loss": 0.3976, "mean_token_accuracy": 0.8886601328849792, "num_tokens": 610039118.0, "step": 16067 }, { "epoch": 2.0440147563923166, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.10167932510376, "learning_rate": 1e-05, "loss": 0.3997, "mean_token_accuracy": 0.8913432955741882, "num_tokens": 610074867.0, "step": 16068 }, { "epoch": 2.044141966670907, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3035948276519775, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8855760097503662, "num_tokens": 610111382.0, "step": 16069 }, { "epoch": 2.0442691769494976, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.223879337310791, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8803746104240417, "num_tokens": 610148026.0, "step": 16070 }, { "epoch": 2.044396387228088, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.525146961212158, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8880447149276733, "num_tokens": 610179841.0, "step": 16071 }, { "epoch": 2.0445235975066787, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00028228759765625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.161698818206787, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8782349824905396, "num_tokens": 610214224.0, "step": 16072 }, { "epoch": 2.044650807785269, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6364307403564453, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8839942812919617, "num_tokens": 610257640.0, "step": 16073 }, { "epoch": 2.0447780180638597, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9859139919281006, "learning_rate": 1e-05, "loss": 0.3918, "mean_token_accuracy": 0.8929577469825745, "num_tokens": 610295438.0, "step": 16074 }, { "epoch": 2.0449052283424503, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.649933099746704, "learning_rate": 1e-05, "loss": 0.3974, "mean_token_accuracy": 0.8899935483932495, "num_tokens": 610336390.0, "step": 16075 }, { "epoch": 2.045032438621041, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8073134422302246, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8873757719993591, "num_tokens": 610369860.0, "step": 16076 }, { "epoch": 2.045159648899631, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6152002811431885, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8851426243782043, "num_tokens": 610408116.0, "step": 16077 }, { "epoch": 2.0452868591782214, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6637110710144043, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8732916116714478, "num_tokens": 610444317.0, "step": 16078 }, { "epoch": 2.045414069456812, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.557548761367798, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8741669654846191, "num_tokens": 610485675.0, "step": 16079 }, { "epoch": 2.0455412797354025, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.55255389213562, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.887819766998291, "num_tokens": 610524697.0, "step": 16080 }, { "epoch": 2.045668490013993, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5002100467681885, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8863472938537598, "num_tokens": 610567470.0, "step": 16081 }, { "epoch": 2.0457957002925835, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.510256767272949, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.883621871471405, "num_tokens": 610603168.0, "step": 16082 }, { "epoch": 2.045922910571174, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.446521043777466, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8827033042907715, "num_tokens": 610643416.0, "step": 16083 }, { "epoch": 2.0460501208497646, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5034091472625732, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8854461908340454, "num_tokens": 610676513.0, "step": 16084 }, { "epoch": 2.046177331128355, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.40451717376709, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.8859803080558777, "num_tokens": 610710842.0, "step": 16085 }, { "epoch": 2.0463045414069456, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5115861892700195, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8688616752624512, "num_tokens": 610747770.0, "step": 16086 }, { "epoch": 2.046431751685536, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3465280532836914, "learning_rate": 1e-05, "loss": 0.3936, "mean_token_accuracy": 0.8883963823318481, "num_tokens": 610788780.0, "step": 16087 }, { "epoch": 2.0465589619641267, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5369248390197754, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.8858238458633423, "num_tokens": 610832930.0, "step": 16088 }, { "epoch": 2.046686172242717, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2925217151641846, "learning_rate": 1e-05, "loss": 0.3855, "mean_token_accuracy": 0.8877251148223877, "num_tokens": 610875828.0, "step": 16089 }, { "epoch": 2.0468133825213077, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.606926679611206, "learning_rate": 1e-05, "loss": 0.3929, "mean_token_accuracy": 0.8897974491119385, "num_tokens": 610913920.0, "step": 16090 }, { "epoch": 2.0469405927998983, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2208898067474365, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8826906681060791, "num_tokens": 610956887.0, "step": 16091 }, { "epoch": 2.047067803078489, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7406296730041504, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8737793564796448, "num_tokens": 611000889.0, "step": 16092 }, { "epoch": 2.0471950133570793, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.1544127464294434, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8821197748184204, "num_tokens": 611040573.0, "step": 16093 }, { "epoch": 2.04732222363567, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.947934150695801, "learning_rate": 1e-05, "loss": 0.3957, "mean_token_accuracy": 0.8903571963310242, "num_tokens": 611075925.0, "step": 16094 }, { "epoch": 2.0474494339142604, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.136157274246216, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8752032518386841, "num_tokens": 611116492.0, "step": 16095 }, { "epoch": 2.047576644192851, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.209649562835693, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8869757056236267, "num_tokens": 611155683.0, "step": 16096 }, { "epoch": 2.0477038544714414, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.2758965492248535, "learning_rate": 1e-05, "loss": 0.3382, "mean_token_accuracy": 0.9075971841812134, "num_tokens": 611194352.0, "step": 16097 }, { "epoch": 2.047831064750032, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.408680438995361, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.886631965637207, "num_tokens": 611229919.0, "step": 16098 }, { "epoch": 2.0479582750286225, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.618515968322754, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8807718753814697, "num_tokens": 611267394.0, "step": 16099 }, { "epoch": 2.048085485307213, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.320526123046875, "learning_rate": 1e-05, "loss": 0.5174, "mean_token_accuracy": 0.8544209003448486, "num_tokens": 611312009.0, "step": 16100 }, { "epoch": 2.0482126955858035, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.809439182281494, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8688703179359436, "num_tokens": 611346723.0, "step": 16101 }, { "epoch": 2.0483399058643936, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.096429824829102, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8865540623664856, "num_tokens": 611381803.0, "step": 16102 }, { "epoch": 2.048467116142984, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.763824701309204, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8813490867614746, "num_tokens": 611419579.0, "step": 16103 }, { "epoch": 2.0485943264215747, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9096810817718506, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8829038143157959, "num_tokens": 611456904.0, "step": 16104 }, { "epoch": 2.048721536700165, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6556215286254883, "learning_rate": 1e-05, "loss": 0.3872, "mean_token_accuracy": 0.8908732533454895, "num_tokens": 611495597.0, "step": 16105 }, { "epoch": 2.0488487469787557, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7880501747131348, "learning_rate": 1e-05, "loss": 0.3936, "mean_token_accuracy": 0.8878989219665527, "num_tokens": 611531307.0, "step": 16106 }, { "epoch": 2.0489759572573463, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.561276912689209, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.867084801197052, "num_tokens": 611567657.0, "step": 16107 }, { "epoch": 2.049103167535937, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.714400053024292, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8799819350242615, "num_tokens": 611608107.0, "step": 16108 }, { "epoch": 2.0492303778145273, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4779834747314453, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.879771888256073, "num_tokens": 611644695.0, "step": 16109 }, { "epoch": 2.049357588093118, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6962509155273438, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8862848281860352, "num_tokens": 611679330.0, "step": 16110 }, { "epoch": 2.0494847983717084, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3827102184295654, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8690165281295776, "num_tokens": 611718792.0, "step": 16111 }, { "epoch": 2.049612008650299, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.743187427520752, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8730934858322144, "num_tokens": 611752559.0, "step": 16112 }, { "epoch": 2.0497392189288894, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.294386863708496, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8867664933204651, "num_tokens": 611791616.0, "step": 16113 }, { "epoch": 2.04986642920748, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8616812229156494, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8720645308494568, "num_tokens": 611826577.0, "step": 16114 }, { "epoch": 2.0499936394860705, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.260195732116699, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8793247938156128, "num_tokens": 611861777.0, "step": 16115 }, { "epoch": 2.050120849764661, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.015598297119141, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8750221133232117, "num_tokens": 611897933.0, "step": 16116 }, { "epoch": 2.0502480600432516, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3107285499572754, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8817703723907471, "num_tokens": 611935285.0, "step": 16117 }, { "epoch": 2.050375270321842, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1235246658325195, "learning_rate": 1e-05, "loss": 0.3967, "mean_token_accuracy": 0.8916745781898499, "num_tokens": 611969879.0, "step": 16118 }, { "epoch": 2.0505024806004326, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4676530361175537, "learning_rate": 1e-05, "loss": 0.4024, "mean_token_accuracy": 0.8871305584907532, "num_tokens": 612003608.0, "step": 16119 }, { "epoch": 2.050629690879023, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.085919380187988, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8777698278427124, "num_tokens": 612044247.0, "step": 16120 }, { "epoch": 2.0507569011576137, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6038882732391357, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8751471042633057, "num_tokens": 612080776.0, "step": 16121 }, { "epoch": 2.050884111436204, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9229657649993896, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8711769580841064, "num_tokens": 612119162.0, "step": 16122 }, { "epoch": 2.0510113217147947, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.620542287826538, "learning_rate": 1e-05, "loss": 0.3868, "mean_token_accuracy": 0.8929656744003296, "num_tokens": 612157774.0, "step": 16123 }, { "epoch": 2.0511385319933853, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7326338291168213, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8835645318031311, "num_tokens": 612192169.0, "step": 16124 }, { "epoch": 2.051265742271976, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.624432325363159, "learning_rate": 1e-05, "loss": 0.3977, "mean_token_accuracy": 0.8874011635780334, "num_tokens": 612225936.0, "step": 16125 }, { "epoch": 2.0513929525505663, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.551957845687866, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8859433531761169, "num_tokens": 612262509.0, "step": 16126 }, { "epoch": 2.0515201628291564, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.609513282775879, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8781951665878296, "num_tokens": 612297952.0, "step": 16127 }, { "epoch": 2.051647373107747, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.425657033920288, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8836637735366821, "num_tokens": 612332512.0, "step": 16128 }, { "epoch": 2.0517745833863374, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6141350269317627, "learning_rate": 1e-05, "loss": 0.3978, "mean_token_accuracy": 0.8886551856994629, "num_tokens": 612370350.0, "step": 16129 }, { "epoch": 2.051901793664928, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3406693935394287, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8723227977752686, "num_tokens": 612404775.0, "step": 16130 }, { "epoch": 2.0520290039435185, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6481847763061523, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8818907737731934, "num_tokens": 612447226.0, "step": 16131 }, { "epoch": 2.052156214222109, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.273698568344116, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8712631464004517, "num_tokens": 612481358.0, "step": 16132 }, { "epoch": 2.0522834245006996, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7208428382873535, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8919765949249268, "num_tokens": 612519132.0, "step": 16133 }, { "epoch": 2.05241063477929, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2145333290100098, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8739105463027954, "num_tokens": 612559187.0, "step": 16134 }, { "epoch": 2.0525378450578806, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.862206220626831, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8870123028755188, "num_tokens": 612598208.0, "step": 16135 }, { "epoch": 2.052665055336471, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1870169639587402, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8843565583229065, "num_tokens": 612633951.0, "step": 16136 }, { "epoch": 2.0527922656150617, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.048076629638672, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8763549327850342, "num_tokens": 612673978.0, "step": 16137 }, { "epoch": 2.052919475893652, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2324655055999756, "learning_rate": 1e-05, "loss": 0.4039, "mean_token_accuracy": 0.8847676515579224, "num_tokens": 612720488.0, "step": 16138 }, { "epoch": 2.0530466861722427, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.235255718231201, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8785059452056885, "num_tokens": 612755009.0, "step": 16139 }, { "epoch": 2.0531738964508333, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4460532665252686, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8784227967262268, "num_tokens": 612790653.0, "step": 16140 }, { "epoch": 2.053301106729424, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.257033348083496, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8859769701957703, "num_tokens": 612829406.0, "step": 16141 }, { "epoch": 2.0534283170080143, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.63877272605896, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8716522455215454, "num_tokens": 612872304.0, "step": 16142 }, { "epoch": 2.053555527286605, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.107312202453613, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8794095516204834, "num_tokens": 612907909.0, "step": 16143 }, { "epoch": 2.0536827375651954, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.698437213897705, "learning_rate": 1e-05, "loss": 0.3938, "mean_token_accuracy": 0.8929314017295837, "num_tokens": 612943794.0, "step": 16144 }, { "epoch": 2.053809947843786, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.902573823928833, "learning_rate": 1e-05, "loss": 0.4054, "mean_token_accuracy": 0.8863312005996704, "num_tokens": 612987125.0, "step": 16145 }, { "epoch": 2.0539371581223764, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6566081047058105, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8735761642456055, "num_tokens": 613028232.0, "step": 16146 }, { "epoch": 2.054064368400967, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7394840717315674, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8843257427215576, "num_tokens": 613066651.0, "step": 16147 }, { "epoch": 2.0541915786795575, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5863089561462402, "learning_rate": 1e-05, "loss": 0.3978, "mean_token_accuracy": 0.8892826437950134, "num_tokens": 613105153.0, "step": 16148 }, { "epoch": 2.054318788958148, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.62345027923584, "learning_rate": 1e-05, "loss": 0.3892, "mean_token_accuracy": 0.8903791904449463, "num_tokens": 613146817.0, "step": 16149 }, { "epoch": 2.0544459992367385, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.516123056411743, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8748281598091125, "num_tokens": 613192644.0, "step": 16150 }, { "epoch": 2.0545732095153286, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.57224702835083, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8794321417808533, "num_tokens": 613232283.0, "step": 16151 }, { "epoch": 2.054700419793919, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4521241188049316, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.8875573873519897, "num_tokens": 613270916.0, "step": 16152 }, { "epoch": 2.0548276300725097, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.561741352081299, "learning_rate": 1e-05, "loss": 0.391, "mean_token_accuracy": 0.890457034111023, "num_tokens": 613303095.0, "step": 16153 }, { "epoch": 2.0549548403511, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3867952823638916, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8778239488601685, "num_tokens": 613341551.0, "step": 16154 }, { "epoch": 2.0550820506296907, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5821683406829834, "learning_rate": 1e-05, "loss": 0.3853, "mean_token_accuracy": 0.8908785581588745, "num_tokens": 613377511.0, "step": 16155 }, { "epoch": 2.0552092609082813, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.334550619125366, "learning_rate": 1e-05, "loss": 0.3897, "mean_token_accuracy": 0.888684868812561, "num_tokens": 613408534.0, "step": 16156 }, { "epoch": 2.055336471186872, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.628786563873291, "learning_rate": 1e-05, "loss": 0.3794, "mean_token_accuracy": 0.8944482207298279, "num_tokens": 613443461.0, "step": 16157 }, { "epoch": 2.0554636814654623, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2689599990844727, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8702587485313416, "num_tokens": 613482438.0, "step": 16158 }, { "epoch": 2.055590891744053, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.730480194091797, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.873445987701416, "num_tokens": 613521937.0, "step": 16159 }, { "epoch": 2.0557181020226434, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.198944568634033, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.870374858379364, "num_tokens": 613563471.0, "step": 16160 }, { "epoch": 2.055845312301234, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9019265174865723, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8810670971870422, "num_tokens": 613603753.0, "step": 16161 }, { "epoch": 2.0559725225798244, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1808929443359375, "learning_rate": 1e-05, "loss": 0.3905, "mean_token_accuracy": 0.8889454007148743, "num_tokens": 613637602.0, "step": 16162 }, { "epoch": 2.056099732858415, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.105260372161865, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8806010484695435, "num_tokens": 613673807.0, "step": 16163 }, { "epoch": 2.0562269431370055, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2888870239257812, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8799623250961304, "num_tokens": 613710702.0, "step": 16164 }, { "epoch": 2.056354153415596, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.242526054382324, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8786811232566833, "num_tokens": 613748890.0, "step": 16165 }, { "epoch": 2.0564813636941865, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5045862197875977, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.879155158996582, "num_tokens": 613791039.0, "step": 16166 }, { "epoch": 2.056608573972777, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.211947441101074, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8855289220809937, "num_tokens": 613821345.0, "step": 16167 }, { "epoch": 2.0567357842513676, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6849160194396973, "learning_rate": 1e-05, "loss": 0.5, "mean_token_accuracy": 0.8555924296379089, "num_tokens": 613857770.0, "step": 16168 }, { "epoch": 2.056862994529958, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.005280017852783, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.8870651721954346, "num_tokens": 613898844.0, "step": 16169 }, { "epoch": 2.0569902048085487, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.683549165725708, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8720055818557739, "num_tokens": 613940884.0, "step": 16170 }, { "epoch": 2.057117415087139, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8126285076141357, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8821905255317688, "num_tokens": 613981664.0, "step": 16171 }, { "epoch": 2.0572446253657297, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.65362548828125, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8758710026741028, "num_tokens": 614013965.0, "step": 16172 }, { "epoch": 2.0573718356443202, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6540327072143555, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8822230100631714, "num_tokens": 614048841.0, "step": 16173 }, { "epoch": 2.0574990459229108, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5792336463928223, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8837267160415649, "num_tokens": 614087677.0, "step": 16174 }, { "epoch": 2.057626256201501, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5420243740081787, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8811324238777161, "num_tokens": 614128546.0, "step": 16175 }, { "epoch": 2.0577534664800914, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5340592861175537, "learning_rate": 1e-05, "loss": 0.3972, "mean_token_accuracy": 0.8891491889953613, "num_tokens": 614163947.0, "step": 16176 }, { "epoch": 2.057880676758682, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4915523529052734, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8788134455680847, "num_tokens": 614202353.0, "step": 16177 }, { "epoch": 2.0580078870372724, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4852981567382812, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.877372145652771, "num_tokens": 614244781.0, "step": 16178 }, { "epoch": 2.058135097315863, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.463552951812744, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.884722113609314, "num_tokens": 614278335.0, "step": 16179 }, { "epoch": 2.0582623075944535, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.450436592102051, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8787181377410889, "num_tokens": 614317726.0, "step": 16180 }, { "epoch": 2.058389517873044, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4427340030670166, "learning_rate": 1e-05, "loss": 0.3933, "mean_token_accuracy": 0.8883509635925293, "num_tokens": 614355403.0, "step": 16181 }, { "epoch": 2.0585167281516346, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002841949462890625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.423527479171753, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.868538498878479, "num_tokens": 614404370.0, "step": 16182 }, { "epoch": 2.058643938430225, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4349260330200195, "learning_rate": 1e-05, "loss": 0.3828, "mean_token_accuracy": 0.8894290924072266, "num_tokens": 614439613.0, "step": 16183 }, { "epoch": 2.0587711487088156, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3935916423797607, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8752933144569397, "num_tokens": 614481634.0, "step": 16184 }, { "epoch": 2.058898358987406, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4497101306915283, "learning_rate": 1e-05, "loss": 0.3944, "mean_token_accuracy": 0.8867883682250977, "num_tokens": 614517178.0, "step": 16185 }, { "epoch": 2.0590255692659967, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3918092250823975, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8682856559753418, "num_tokens": 614551531.0, "step": 16186 }, { "epoch": 2.059152779544587, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4636847972869873, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8760473132133484, "num_tokens": 614589173.0, "step": 16187 }, { "epoch": 2.0592799898231777, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3321869373321533, "learning_rate": 1e-05, "loss": 0.4065, "mean_token_accuracy": 0.8799588680267334, "num_tokens": 614633657.0, "step": 16188 }, { "epoch": 2.0594072001017683, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5146520137786865, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8800068497657776, "num_tokens": 614670013.0, "step": 16189 }, { "epoch": 2.059534410380359, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.288417100906372, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8751989603042603, "num_tokens": 614706414.0, "step": 16190 }, { "epoch": 2.0596616206589493, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.585510730743408, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8760618567466736, "num_tokens": 614746397.0, "step": 16191 }, { "epoch": 2.05978883093754, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2179243564605713, "learning_rate": 1e-05, "loss": 0.3861, "mean_token_accuracy": 0.8892252445220947, "num_tokens": 614785587.0, "step": 16192 }, { "epoch": 2.0599160412161304, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7176854610443115, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8852547407150269, "num_tokens": 614823304.0, "step": 16193 }, { "epoch": 2.060043251494721, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.148118019104004, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8834235668182373, "num_tokens": 614857544.0, "step": 16194 }, { "epoch": 2.0601704617733114, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9382383823394775, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8772345781326294, "num_tokens": 614893095.0, "step": 16195 }, { "epoch": 2.060297672051902, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.088380813598633, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8765776753425598, "num_tokens": 614932634.0, "step": 16196 }, { "epoch": 2.0604248823304925, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.252198696136475, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8792269229888916, "num_tokens": 614975039.0, "step": 16197 }, { "epoch": 2.060552092609083, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.231294631958008, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8817815780639648, "num_tokens": 615019092.0, "step": 16198 }, { "epoch": 2.0606793028876735, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.543671607971191, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.891150951385498, "num_tokens": 615060183.0, "step": 16199 }, { "epoch": 2.0608065131662636, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.671088457107544, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8799309134483337, "num_tokens": 615098376.0, "step": 16200 }, { "epoch": 2.060933723444854, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.483916282653809, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8877803087234497, "num_tokens": 615134401.0, "step": 16201 }, { "epoch": 2.0610609337234447, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 3.938486099243164, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.88461834192276, "num_tokens": 615176145.0, "step": 16202 }, { "epoch": 2.061188144002035, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.21261739730835, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8754551410675049, "num_tokens": 615210281.0, "step": 16203 }, { "epoch": 2.0613153542806257, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.8792386054992676, "learning_rate": 1e-05, "loss": 0.379, "mean_token_accuracy": 0.8990966081619263, "num_tokens": 615246952.0, "step": 16204 }, { "epoch": 2.0614425645592163, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.9976019859313965, "learning_rate": 1e-05, "loss": 0.3955, "mean_token_accuracy": 0.8939268589019775, "num_tokens": 615285807.0, "step": 16205 }, { "epoch": 2.061569774837807, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7483291625976562, "learning_rate": 1e-05, "loss": 0.3833, "mean_token_accuracy": 0.893520712852478, "num_tokens": 615319641.0, "step": 16206 }, { "epoch": 2.0616969851163973, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8714451789855957, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8683268427848816, "num_tokens": 615358072.0, "step": 16207 }, { "epoch": 2.061824195394988, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.612987518310547, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8828727006912231, "num_tokens": 615393924.0, "step": 16208 }, { "epoch": 2.0619514056735784, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7900784015655518, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8818906545639038, "num_tokens": 615436771.0, "step": 16209 }, { "epoch": 2.062078615952169, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.483613967895508, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8775771260261536, "num_tokens": 615473577.0, "step": 16210 }, { "epoch": 2.0622058262307594, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8063344955444336, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8650597333908081, "num_tokens": 615507585.0, "step": 16211 }, { "epoch": 2.06233303650935, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3788459300994873, "learning_rate": 1e-05, "loss": 0.3921, "mean_token_accuracy": 0.8873956203460693, "num_tokens": 615542638.0, "step": 16212 }, { "epoch": 2.0624602467879405, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8845925331115723, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8766499757766724, "num_tokens": 615579567.0, "step": 16213 }, { "epoch": 2.062587457066531, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3173582553863525, "learning_rate": 1e-05, "loss": 0.3855, "mean_token_accuracy": 0.8904484510421753, "num_tokens": 615616563.0, "step": 16214 }, { "epoch": 2.0627146673451215, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.021429538726807, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8710916042327881, "num_tokens": 615653243.0, "step": 16215 }, { "epoch": 2.062841877623712, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3313002586364746, "learning_rate": 1e-05, "loss": 0.381, "mean_token_accuracy": 0.889380931854248, "num_tokens": 615689552.0, "step": 16216 }, { "epoch": 2.0629690879023026, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.171360492706299, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8722392320632935, "num_tokens": 615723803.0, "step": 16217 }, { "epoch": 2.063096298180893, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.504058361053467, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8816003799438477, "num_tokens": 615761181.0, "step": 16218 }, { "epoch": 2.0632235084594837, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.118412017822266, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8893865346908569, "num_tokens": 615802200.0, "step": 16219 }, { "epoch": 2.063350718738074, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6297595500946045, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8866264224052429, "num_tokens": 615843299.0, "step": 16220 }, { "epoch": 2.0634779290166647, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9619619846343994, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8798756003379822, "num_tokens": 615886148.0, "step": 16221 }, { "epoch": 2.0636051392952552, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.673610210418701, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8747043013572693, "num_tokens": 615930595.0, "step": 16222 }, { "epoch": 2.0637323495738458, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.756185293197632, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8809354901313782, "num_tokens": 615966873.0, "step": 16223 }, { "epoch": 2.0638595598524363, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6583149433135986, "learning_rate": 1e-05, "loss": 0.3723, "mean_token_accuracy": 0.894868791103363, "num_tokens": 616002021.0, "step": 16224 }, { "epoch": 2.0639867701310264, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5676121711730957, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8823205232620239, "num_tokens": 616043692.0, "step": 16225 }, { "epoch": 2.064113980409617, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6308836936950684, "learning_rate": 1e-05, "loss": 0.39, "mean_token_accuracy": 0.8922612071037292, "num_tokens": 616083811.0, "step": 16226 }, { "epoch": 2.0642411906882074, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4283206462860107, "learning_rate": 1e-05, "loss": 0.3728, "mean_token_accuracy": 0.8938277959823608, "num_tokens": 616121714.0, "step": 16227 }, { "epoch": 2.064368400966798, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.640350580215454, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8793454170227051, "num_tokens": 616156571.0, "step": 16228 }, { "epoch": 2.0644956112453885, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3346428871154785, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.876285195350647, "num_tokens": 616197198.0, "step": 16229 }, { "epoch": 2.064622821523979, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6856021881103516, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8703950643539429, "num_tokens": 616233495.0, "step": 16230 }, { "epoch": 2.0647500318025696, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2524642944335938, "learning_rate": 1e-05, "loss": 0.3749, "mean_token_accuracy": 0.8942505121231079, "num_tokens": 616268894.0, "step": 16231 }, { "epoch": 2.06487724208116, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7980456352233887, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8819386959075928, "num_tokens": 616302899.0, "step": 16232 }, { "epoch": 2.0650044523597506, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1993350982666016, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8628289699554443, "num_tokens": 616342937.0, "step": 16233 }, { "epoch": 2.065131662638341, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.938951253890991, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8813444375991821, "num_tokens": 616391581.0, "step": 16234 }, { "epoch": 2.0652588729169317, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1950125694274902, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.8855881690979004, "num_tokens": 616430303.0, "step": 16235 }, { "epoch": 2.065386083195522, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.147756099700928, "learning_rate": 1e-05, "loss": 0.3966, "mean_token_accuracy": 0.8917326927185059, "num_tokens": 616468294.0, "step": 16236 }, { "epoch": 2.0655132934741127, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.334563970565796, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8845853805541992, "num_tokens": 616503551.0, "step": 16237 }, { "epoch": 2.0656405037527032, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.256378173828125, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8906936645507812, "num_tokens": 616541374.0, "step": 16238 }, { "epoch": 2.065767714031294, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5546610355377197, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8861455917358398, "num_tokens": 616575480.0, "step": 16239 }, { "epoch": 2.0658949243098843, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.188956260681152, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8802151083946228, "num_tokens": 616620323.0, "step": 16240 }, { "epoch": 2.066022134588475, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.680647850036621, "learning_rate": 1e-05, "loss": 0.3895, "mean_token_accuracy": 0.8961035013198853, "num_tokens": 616659843.0, "step": 16241 }, { "epoch": 2.0661493448670654, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.007906913757324, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8861925601959229, "num_tokens": 616696065.0, "step": 16242 }, { "epoch": 2.066276555145656, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.69822359085083, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8863290548324585, "num_tokens": 616729444.0, "step": 16243 }, { "epoch": 2.0664037654242464, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8009755611419678, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8820629715919495, "num_tokens": 616773351.0, "step": 16244 }, { "epoch": 2.066530975702837, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6555683612823486, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8800604343414307, "num_tokens": 616812161.0, "step": 16245 }, { "epoch": 2.0666581859814275, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6639978885650635, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8659953474998474, "num_tokens": 616847623.0, "step": 16246 }, { "epoch": 2.066785396260018, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5731306076049805, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8799801468849182, "num_tokens": 616884078.0, "step": 16247 }, { "epoch": 2.0669126065386085, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5540547370910645, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8804223537445068, "num_tokens": 616924291.0, "step": 16248 }, { "epoch": 2.0670398168171986, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.507047176361084, "learning_rate": 1e-05, "loss": 0.3991, "mean_token_accuracy": 0.8881148099899292, "num_tokens": 616964580.0, "step": 16249 }, { "epoch": 2.067167027095789, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.503797769546509, "learning_rate": 1e-05, "loss": 0.3849, "mean_token_accuracy": 0.8922011852264404, "num_tokens": 617000574.0, "step": 16250 }, { "epoch": 2.0672942373743797, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4558658599853516, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8818480372428894, "num_tokens": 617043323.0, "step": 16251 }, { "epoch": 2.06742144765297, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4846529960632324, "learning_rate": 1e-05, "loss": 0.4011, "mean_token_accuracy": 0.8858828544616699, "num_tokens": 617082585.0, "step": 16252 }, { "epoch": 2.0675486579315607, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4207546710968018, "learning_rate": 1e-05, "loss": 0.3994, "mean_token_accuracy": 0.8849295377731323, "num_tokens": 617118384.0, "step": 16253 }, { "epoch": 2.0676758682101513, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4813642501831055, "learning_rate": 1e-05, "loss": 0.3664, "mean_token_accuracy": 0.8986146450042725, "num_tokens": 617153354.0, "step": 16254 }, { "epoch": 2.067803078488742, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3754324913024902, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8840345144271851, "num_tokens": 617192829.0, "step": 16255 }, { "epoch": 2.0679302887673323, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5052123069763184, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8835194110870361, "num_tokens": 617230824.0, "step": 16256 }, { "epoch": 2.068057499045923, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.332669496536255, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8732771873474121, "num_tokens": 617268791.0, "step": 16257 }, { "epoch": 2.0681847093245134, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.562215805053711, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8807864189147949, "num_tokens": 617305366.0, "step": 16258 }, { "epoch": 2.068311919603104, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2767012119293213, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8810495138168335, "num_tokens": 617342294.0, "step": 16259 }, { "epoch": 2.0684391298816944, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.660353422164917, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8816616535186768, "num_tokens": 617380456.0, "step": 16260 }, { "epoch": 2.068566340160285, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1805477142333984, "learning_rate": 1e-05, "loss": 0.3857, "mean_token_accuracy": 0.8880578279495239, "num_tokens": 617420705.0, "step": 16261 }, { "epoch": 2.0686935504388755, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8208906650543213, "learning_rate": 1e-05, "loss": 0.4046, "mean_token_accuracy": 0.8867334127426147, "num_tokens": 617462740.0, "step": 16262 }, { "epoch": 2.068820760717466, "ewc_loss": 0.07421875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1243653297424316, "learning_rate": 1e-05, "loss": 0.3799, "mean_token_accuracy": 0.8895351886749268, "num_tokens": 617497496.0, "step": 16263 }, { "epoch": 2.0689479709960565, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.068634510040283, "learning_rate": 1e-05, "loss": 0.3907, "mean_token_accuracy": 0.8921797275543213, "num_tokens": 617535297.0, "step": 16264 }, { "epoch": 2.069075181274647, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.181581497192383, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.881893515586853, "num_tokens": 617568332.0, "step": 16265 }, { "epoch": 2.0692023915532376, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.321840763092041, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.878237247467041, "num_tokens": 617608934.0, "step": 16266 }, { "epoch": 2.069329601831828, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.431403636932373, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8725712299346924, "num_tokens": 617643794.0, "step": 16267 }, { "epoch": 2.0694568121104187, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.40898323059082, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8859248757362366, "num_tokens": 617676448.0, "step": 16268 }, { "epoch": 2.069584022389009, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.7428946495056152, "learning_rate": 1e-05, "loss": 0.5021, "mean_token_accuracy": 0.8574222326278687, "num_tokens": 617713460.0, "step": 16269 }, { "epoch": 2.0697112326675997, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.225745677947998, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8745158910751343, "num_tokens": 617749767.0, "step": 16270 }, { "epoch": 2.0698384429461902, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.8176913261413574, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8660601377487183, "num_tokens": 617784991.0, "step": 16271 }, { "epoch": 2.0699656532247808, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.999065637588501, "learning_rate": 1e-05, "loss": 0.3915, "mean_token_accuracy": 0.8908654451370239, "num_tokens": 617820241.0, "step": 16272 }, { "epoch": 2.070092863503371, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7236521244049072, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8705834150314331, "num_tokens": 617861838.0, "step": 16273 }, { "epoch": 2.0702200737819614, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8500828742980957, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8813709616661072, "num_tokens": 617893001.0, "step": 16274 }, { "epoch": 2.070347284060552, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6055796146392822, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8773455023765564, "num_tokens": 617936361.0, "step": 16275 }, { "epoch": 2.0704744943391424, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.756180763244629, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8778821229934692, "num_tokens": 617968267.0, "step": 16276 }, { "epoch": 2.070601704617733, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.499622106552124, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8799872398376465, "num_tokens": 618007710.0, "step": 16277 }, { "epoch": 2.0707289148963235, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7334201335906982, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8813634514808655, "num_tokens": 618043185.0, "step": 16278 }, { "epoch": 2.070856125174914, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3966238498687744, "learning_rate": 1e-05, "loss": 0.383, "mean_token_accuracy": 0.8915228843688965, "num_tokens": 618085460.0, "step": 16279 }, { "epoch": 2.0709833354535045, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7927465438842773, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8905165195465088, "num_tokens": 618117155.0, "step": 16280 }, { "epoch": 2.071110545732095, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3175363540649414, "learning_rate": 1e-05, "loss": 0.3983, "mean_token_accuracy": 0.8900777697563171, "num_tokens": 618153380.0, "step": 16281 }, { "epoch": 2.0712377560106856, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8898138999938965, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8826664090156555, "num_tokens": 618190125.0, "step": 16282 }, { "epoch": 2.071364966289276, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.289696455001831, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8822636604309082, "num_tokens": 618225281.0, "step": 16283 }, { "epoch": 2.0714921765678667, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.019155979156494, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8873958587646484, "num_tokens": 618262213.0, "step": 16284 }, { "epoch": 2.071619386846457, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.367064952850342, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.8623299598693848, "num_tokens": 618292780.0, "step": 16285 }, { "epoch": 2.0717465971250477, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.086047649383545, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8662986755371094, "num_tokens": 618332968.0, "step": 16286 }, { "epoch": 2.0718738074036382, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.465873956680298, "learning_rate": 1e-05, "loss": 0.3827, "mean_token_accuracy": 0.8912602663040161, "num_tokens": 618369585.0, "step": 16287 }, { "epoch": 2.0720010176822288, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.042712211608887, "learning_rate": 1e-05, "loss": 0.3797, "mean_token_accuracy": 0.8971346616744995, "num_tokens": 618410388.0, "step": 16288 }, { "epoch": 2.0721282279608193, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5652353763580322, "learning_rate": 1e-05, "loss": 0.3763, "mean_token_accuracy": 0.8961943984031677, "num_tokens": 618453591.0, "step": 16289 }, { "epoch": 2.07225543823941, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9045252799987793, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8850088119506836, "num_tokens": 618489757.0, "step": 16290 }, { "epoch": 2.0723826485180004, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6195788383483887, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8864402770996094, "num_tokens": 618524733.0, "step": 16291 }, { "epoch": 2.072509858796591, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7084708213806152, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8778203725814819, "num_tokens": 618564734.0, "step": 16292 }, { "epoch": 2.0726370690751814, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6051721572875977, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8841184377670288, "num_tokens": 618602503.0, "step": 16293 }, { "epoch": 2.072764279353772, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.535851001739502, "learning_rate": 1e-05, "loss": 0.3799, "mean_token_accuracy": 0.8935060501098633, "num_tokens": 618639907.0, "step": 16294 }, { "epoch": 2.0728914896323625, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.595167875289917, "learning_rate": 1e-05, "loss": 0.4065, "mean_token_accuracy": 0.8866472244262695, "num_tokens": 618681366.0, "step": 16295 }, { "epoch": 2.073018699910953, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.403661012649536, "learning_rate": 1e-05, "loss": 0.3593, "mean_token_accuracy": 0.8970904350280762, "num_tokens": 618718610.0, "step": 16296 }, { "epoch": 2.0731459101895435, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6167640686035156, "learning_rate": 1e-05, "loss": 0.3906, "mean_token_accuracy": 0.8880118131637573, "num_tokens": 618757741.0, "step": 16297 }, { "epoch": 2.0732731204681336, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.314169406890869, "learning_rate": 1e-05, "loss": 0.3853, "mean_token_accuracy": 0.8895674347877502, "num_tokens": 618798189.0, "step": 16298 }, { "epoch": 2.073400330746724, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6715362071990967, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8814849257469177, "num_tokens": 618837308.0, "step": 16299 }, { "epoch": 2.0735275410253147, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.268641710281372, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8675191402435303, "num_tokens": 618876058.0, "step": 16300 }, { "epoch": 2.073654751303905, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.746744155883789, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8732360005378723, "num_tokens": 618916246.0, "step": 16301 }, { "epoch": 2.0737819615824957, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.193046808242798, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.874604344367981, "num_tokens": 618957639.0, "step": 16302 }, { "epoch": 2.0739091718610863, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.887037515640259, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8905846476554871, "num_tokens": 618999505.0, "step": 16303 }, { "epoch": 2.074036382139677, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1781256198883057, "learning_rate": 1e-05, "loss": 0.3871, "mean_token_accuracy": 0.8907912969589233, "num_tokens": 619034807.0, "step": 16304 }, { "epoch": 2.0741635924182673, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.085536479949951, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8911110162734985, "num_tokens": 619073262.0, "step": 16305 }, { "epoch": 2.074290802696858, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2664730548858643, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8743774890899658, "num_tokens": 619112706.0, "step": 16306 }, { "epoch": 2.0744180129754484, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000286102294921875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2575507164001465, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8743760585784912, "num_tokens": 619144444.0, "step": 16307 }, { "epoch": 2.074545223254039, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.4853157997131348, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8902947902679443, "num_tokens": 619183291.0, "step": 16308 }, { "epoch": 2.0746724335326294, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.222172737121582, "learning_rate": 1e-05, "loss": 0.4044, "mean_token_accuracy": 0.8906867504119873, "num_tokens": 619223865.0, "step": 16309 }, { "epoch": 2.07479964381122, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.667799949645996, "learning_rate": 1e-05, "loss": 0.393, "mean_token_accuracy": 0.8909206390380859, "num_tokens": 619257418.0, "step": 16310 }, { "epoch": 2.0749268540898105, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.047966480255127, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8756054043769836, "num_tokens": 619298724.0, "step": 16311 }, { "epoch": 2.075054064368401, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7008090019226074, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8837475776672363, "num_tokens": 619332497.0, "step": 16312 }, { "epoch": 2.0751812746469915, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8599464893341064, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8824050426483154, "num_tokens": 619363371.0, "step": 16313 }, { "epoch": 2.075308484925582, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6557106971740723, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8863937854766846, "num_tokens": 619396068.0, "step": 16314 }, { "epoch": 2.0754356952041726, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6882386207580566, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8814423084259033, "num_tokens": 619433728.0, "step": 16315 }, { "epoch": 2.075562905482763, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5773985385894775, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8731059432029724, "num_tokens": 619476185.0, "step": 16316 }, { "epoch": 2.0756901157613536, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5736923217773438, "learning_rate": 1e-05, "loss": 0.3564, "mean_token_accuracy": 0.9008692502975464, "num_tokens": 619513184.0, "step": 16317 }, { "epoch": 2.075817326039944, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.517050266265869, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8756985068321228, "num_tokens": 619550168.0, "step": 16318 }, { "epoch": 2.0759445363185347, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5271925926208496, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8777408599853516, "num_tokens": 619588960.0, "step": 16319 }, { "epoch": 2.0760717465971252, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.475010871887207, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.883010745048523, "num_tokens": 619624907.0, "step": 16320 }, { "epoch": 2.0761989568757158, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.469754457473755, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.8883091807365417, "num_tokens": 619670576.0, "step": 16321 }, { "epoch": 2.0763261671543063, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4330227375030518, "learning_rate": 1e-05, "loss": 0.3879, "mean_token_accuracy": 0.8901426792144775, "num_tokens": 619710452.0, "step": 16322 }, { "epoch": 2.0764533774328964, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.475842237472534, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8835926055908203, "num_tokens": 619750104.0, "step": 16323 }, { "epoch": 2.076580587711487, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3910632133483887, "learning_rate": 1e-05, "loss": 0.3972, "mean_token_accuracy": 0.8876815438270569, "num_tokens": 619791364.0, "step": 16324 }, { "epoch": 2.0767077979900774, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5078022480010986, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8785274624824524, "num_tokens": 619831138.0, "step": 16325 }, { "epoch": 2.076835008268668, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3570752143859863, "learning_rate": 1e-05, "loss": 0.4029, "mean_token_accuracy": 0.8853848576545715, "num_tokens": 619862897.0, "step": 16326 }, { "epoch": 2.0769622185472585, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5298521518707275, "learning_rate": 1e-05, "loss": 0.3885, "mean_token_accuracy": 0.8918653726577759, "num_tokens": 619897096.0, "step": 16327 }, { "epoch": 2.077089428825849, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2857470512390137, "learning_rate": 1e-05, "loss": 0.3735, "mean_token_accuracy": 0.8951915502548218, "num_tokens": 619937458.0, "step": 16328 }, { "epoch": 2.0772166391044395, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.625748872756958, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8705348372459412, "num_tokens": 619974735.0, "step": 16329 }, { "epoch": 2.07734384938303, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.219919443130493, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.877913236618042, "num_tokens": 620012497.0, "step": 16330 }, { "epoch": 2.0774710596616206, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.756732940673828, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8781983256340027, "num_tokens": 620051120.0, "step": 16331 }, { "epoch": 2.077598269940211, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.142038106918335, "learning_rate": 1e-05, "loss": 0.4045, "mean_token_accuracy": 0.8823059797286987, "num_tokens": 620091479.0, "step": 16332 }, { "epoch": 2.0777254802188017, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9829468727111816, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.8835086822509766, "num_tokens": 620126509.0, "step": 16333 }, { "epoch": 2.077852690497392, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.139676094055176, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8849262595176697, "num_tokens": 620164278.0, "step": 16334 }, { "epoch": 2.0779799007759827, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.269802570343018, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8742507696151733, "num_tokens": 620202388.0, "step": 16335 }, { "epoch": 2.0781071110545732, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.326467275619507, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8867030143737793, "num_tokens": 620236981.0, "step": 16336 }, { "epoch": 2.0782343213331638, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.445254802703857, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.886072039604187, "num_tokens": 620275946.0, "step": 16337 }, { "epoch": 2.0783615316117543, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.6718497276306152, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8684269785881042, "num_tokens": 620317385.0, "step": 16338 }, { "epoch": 2.078488741890345, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.322394847869873, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8882704973220825, "num_tokens": 620353625.0, "step": 16339 }, { "epoch": 2.0786159521689354, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.8546059131622314, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8646044731140137, "num_tokens": 620389421.0, "step": 16340 }, { "epoch": 2.078743162447526, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.117207050323486, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8870446085929871, "num_tokens": 620428233.0, "step": 16341 }, { "epoch": 2.0788703727261164, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.813243865966797, "learning_rate": 1e-05, "loss": 0.503, "mean_token_accuracy": 0.8631116151809692, "num_tokens": 620465156.0, "step": 16342 }, { "epoch": 2.078997583004707, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.880845785140991, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.881192684173584, "num_tokens": 620501731.0, "step": 16343 }, { "epoch": 2.0791247932832975, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.679119825363159, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8775388598442078, "num_tokens": 620538507.0, "step": 16344 }, { "epoch": 2.079252003561888, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7443161010742188, "learning_rate": 1e-05, "loss": 0.3708, "mean_token_accuracy": 0.8970240354537964, "num_tokens": 620577355.0, "step": 16345 }, { "epoch": 2.0793792138404785, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5770134925842285, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8769177794456482, "num_tokens": 620613320.0, "step": 16346 }, { "epoch": 2.0795064241190686, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6643528938293457, "learning_rate": 1e-05, "loss": 0.3766, "mean_token_accuracy": 0.8937122821807861, "num_tokens": 620655167.0, "step": 16347 }, { "epoch": 2.079633634397659, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.467900276184082, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8663553595542908, "num_tokens": 620697898.0, "step": 16348 }, { "epoch": 2.0797608446762497, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6471869945526123, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8877396583557129, "num_tokens": 620736548.0, "step": 16349 }, { "epoch": 2.07988805495484, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3807787895202637, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8720201253890991, "num_tokens": 620781368.0, "step": 16350 }, { "epoch": 2.0800152652334307, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.691112995147705, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8824657797813416, "num_tokens": 620821749.0, "step": 16351 }, { "epoch": 2.0801424755120212, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2927935123443604, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8785953521728516, "num_tokens": 620857906.0, "step": 16352 }, { "epoch": 2.0802696857906118, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8111228942871094, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.882549524307251, "num_tokens": 620897124.0, "step": 16353 }, { "epoch": 2.0803968960692023, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2376203536987305, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.871411144733429, "num_tokens": 620932123.0, "step": 16354 }, { "epoch": 2.080524106347793, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.979329824447632, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8825342655181885, "num_tokens": 620970906.0, "step": 16355 }, { "epoch": 2.0806513166263834, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2529096603393555, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8737948536872864, "num_tokens": 621007387.0, "step": 16356 }, { "epoch": 2.080778526904974, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.165093898773193, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8833787441253662, "num_tokens": 621043723.0, "step": 16357 }, { "epoch": 2.0809057371835644, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.389920949935913, "learning_rate": 1e-05, "loss": 0.3876, "mean_token_accuracy": 0.890085756778717, "num_tokens": 621083686.0, "step": 16358 }, { "epoch": 2.081032947462155, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.216842174530029, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8822638988494873, "num_tokens": 621125758.0, "step": 16359 }, { "epoch": 2.0811601577407455, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6066055297851562, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.882597804069519, "num_tokens": 621163407.0, "step": 16360 }, { "epoch": 2.081287368019336, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.088515758514404, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8751025199890137, "num_tokens": 621203321.0, "step": 16361 }, { "epoch": 2.0814145782979265, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.700479030609131, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8798151016235352, "num_tokens": 621239777.0, "step": 16362 }, { "epoch": 2.081541788576517, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8771848678588867, "learning_rate": 1e-05, "loss": 0.4045, "mean_token_accuracy": 0.8869209885597229, "num_tokens": 621271841.0, "step": 16363 }, { "epoch": 2.0816689988551076, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.678901195526123, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8837958574295044, "num_tokens": 621309393.0, "step": 16364 }, { "epoch": 2.081796209133698, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.687556743621826, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8700934648513794, "num_tokens": 621344362.0, "step": 16365 }, { "epoch": 2.0819234194122886, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6381478309631348, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8646059036254883, "num_tokens": 621383999.0, "step": 16366 }, { "epoch": 2.082050629690879, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5224130153656006, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.887692391872406, "num_tokens": 621424475.0, "step": 16367 }, { "epoch": 2.0821778399694697, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.603846788406372, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8699510097503662, "num_tokens": 621458241.0, "step": 16368 }, { "epoch": 2.0823050502480602, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.433746576309204, "learning_rate": 1e-05, "loss": 0.3892, "mean_token_accuracy": 0.8896974325180054, "num_tokens": 621495424.0, "step": 16369 }, { "epoch": 2.0824322605266508, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.591578722000122, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8821800351142883, "num_tokens": 621531256.0, "step": 16370 }, { "epoch": 2.082559470805241, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3575565814971924, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8837751746177673, "num_tokens": 621571665.0, "step": 16371 }, { "epoch": 2.0826866810838314, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6237785816192627, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8866704702377319, "num_tokens": 621604552.0, "step": 16372 }, { "epoch": 2.082813891362422, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.296623468399048, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8626054525375366, "num_tokens": 621651677.0, "step": 16373 }, { "epoch": 2.0829411016410124, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6705329418182373, "learning_rate": 1e-05, "loss": 0.3948, "mean_token_accuracy": 0.8891352415084839, "num_tokens": 621695108.0, "step": 16374 }, { "epoch": 2.083068311919603, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2452962398529053, "learning_rate": 1e-05, "loss": 0.3823, "mean_token_accuracy": 0.8909069895744324, "num_tokens": 621725768.0, "step": 16375 }, { "epoch": 2.0831955221981935, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7789998054504395, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8726183176040649, "num_tokens": 621768559.0, "step": 16376 }, { "epoch": 2.083322732476784, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1820313930511475, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.880052924156189, "num_tokens": 621805464.0, "step": 16377 }, { "epoch": 2.0834499427553745, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.111660957336426, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8724044561386108, "num_tokens": 621843129.0, "step": 16378 }, { "epoch": 2.083577153033965, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.208439588546753, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8822518587112427, "num_tokens": 621882450.0, "step": 16379 }, { "epoch": 2.0837043633125556, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.255491256713867, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.880359947681427, "num_tokens": 621915797.0, "step": 16380 }, { "epoch": 2.083831573591146, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4027481079101562, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8730885982513428, "num_tokens": 621953563.0, "step": 16381 }, { "epoch": 2.0839587838697367, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.342190265655518, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8950096368789673, "num_tokens": 621989579.0, "step": 16382 }, { "epoch": 2.084085994148327, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6517553329467773, "learning_rate": 1e-05, "loss": 0.3843, "mean_token_accuracy": 0.8954277038574219, "num_tokens": 622027446.0, "step": 16383 }, { "epoch": 2.0842132044269177, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.21898078918457, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8716009855270386, "num_tokens": 622064972.0, "step": 16384 }, { "epoch": 2.0843404147055082, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7601473331451416, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8933890461921692, "num_tokens": 622103260.0, "step": 16385 }, { "epoch": 2.0844676249840988, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.009481430053711, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8856791257858276, "num_tokens": 622140894.0, "step": 16386 }, { "epoch": 2.0845948352626893, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.727952718734741, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8894423842430115, "num_tokens": 622177537.0, "step": 16387 }, { "epoch": 2.08472204554128, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.846836566925049, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8847336173057556, "num_tokens": 622211172.0, "step": 16388 }, { "epoch": 2.0848492558198704, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.645075798034668, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8786773085594177, "num_tokens": 622244805.0, "step": 16389 }, { "epoch": 2.084976466098461, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.710387945175171, "learning_rate": 1e-05, "loss": 0.3899, "mean_token_accuracy": 0.8929930925369263, "num_tokens": 622279496.0, "step": 16390 }, { "epoch": 2.0851036763770514, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.539350986480713, "learning_rate": 1e-05, "loss": 0.3941, "mean_token_accuracy": 0.8908001184463501, "num_tokens": 622318818.0, "step": 16391 }, { "epoch": 2.085230886655642, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.644993782043457, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8862752318382263, "num_tokens": 622359313.0, "step": 16392 }, { "epoch": 2.0853580969342325, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4641544818878174, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.877350926399231, "num_tokens": 622392840.0, "step": 16393 }, { "epoch": 2.085485307212823, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6302261352539062, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8665136694908142, "num_tokens": 622433522.0, "step": 16394 }, { "epoch": 2.0856125174914135, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3927125930786133, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8825639486312866, "num_tokens": 622468805.0, "step": 16395 }, { "epoch": 2.0857397277700036, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.646308660507202, "learning_rate": 1e-05, "loss": 0.389, "mean_token_accuracy": 0.8918389081954956, "num_tokens": 622511462.0, "step": 16396 }, { "epoch": 2.085866938048594, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.310333728790283, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8851813077926636, "num_tokens": 622546246.0, "step": 16397 }, { "epoch": 2.0859941483271847, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7301430702209473, "learning_rate": 1e-05, "loss": 0.3922, "mean_token_accuracy": 0.8911058902740479, "num_tokens": 622583113.0, "step": 16398 }, { "epoch": 2.086121358605775, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2422032356262207, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8788831233978271, "num_tokens": 622621414.0, "step": 16399 }, { "epoch": 2.0862485688843657, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.880746841430664, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.865399181842804, "num_tokens": 622660079.0, "step": 16400 }, { "epoch": 2.0863757791629562, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.220282554626465, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8804894685745239, "num_tokens": 622700413.0, "step": 16401 }, { "epoch": 2.0865029894415468, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.054538249969482, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8873587846755981, "num_tokens": 622736448.0, "step": 16402 }, { "epoch": 2.0866301997201373, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2925686836242676, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8821603059768677, "num_tokens": 622777014.0, "step": 16403 }, { "epoch": 2.086757409998728, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.177608013153076, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8848413228988647, "num_tokens": 622817608.0, "step": 16404 }, { "epoch": 2.0868846202773184, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4884767532348633, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8814336061477661, "num_tokens": 622853090.0, "step": 16405 }, { "epoch": 2.087011830555909, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.130153656005859, "learning_rate": 1e-05, "loss": 0.3858, "mean_token_accuracy": 0.8957083225250244, "num_tokens": 622891843.0, "step": 16406 }, { "epoch": 2.0871390408344994, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6387314796447754, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.86088627576828, "num_tokens": 622930364.0, "step": 16407 }, { "epoch": 2.08726625111309, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.963052988052368, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8633036017417908, "num_tokens": 622970199.0, "step": 16408 }, { "epoch": 2.0873934613916805, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.672764301300049, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8745508193969727, "num_tokens": 623003669.0, "step": 16409 }, { "epoch": 2.087520671670271, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7388923168182373, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8670175075531006, "num_tokens": 623043110.0, "step": 16410 }, { "epoch": 2.0876478819488615, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6437828540802, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8669257164001465, "num_tokens": 623082474.0, "step": 16411 }, { "epoch": 2.087775092227452, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5623297691345215, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8663086891174316, "num_tokens": 623119650.0, "step": 16412 }, { "epoch": 2.0879023025060426, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.634645938873291, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.870832085609436, "num_tokens": 623150359.0, "step": 16413 }, { "epoch": 2.088029512784633, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4289090633392334, "learning_rate": 1e-05, "loss": 0.4123, "mean_token_accuracy": 0.8819196224212646, "num_tokens": 623189251.0, "step": 16414 }, { "epoch": 2.0881567230632236, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.612816333770752, "learning_rate": 1e-05, "loss": 0.3805, "mean_token_accuracy": 0.8941192626953125, "num_tokens": 623231763.0, "step": 16415 }, { "epoch": 2.088283933341814, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.340001106262207, "learning_rate": 1e-05, "loss": 0.3897, "mean_token_accuracy": 0.8893665671348572, "num_tokens": 623269322.0, "step": 16416 }, { "epoch": 2.0884111436204047, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.667325735092163, "learning_rate": 1e-05, "loss": 0.3978, "mean_token_accuracy": 0.8885538578033447, "num_tokens": 623303765.0, "step": 16417 }, { "epoch": 2.0885383538989952, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.255873918533325, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8750606775283813, "num_tokens": 623355484.0, "step": 16418 }, { "epoch": 2.0886655641775858, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.766425132751465, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.877448320388794, "num_tokens": 623399441.0, "step": 16419 }, { "epoch": 2.0887927744561763, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2183940410614014, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8588939905166626, "num_tokens": 623438690.0, "step": 16420 }, { "epoch": 2.0889199847347664, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9109721183776855, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8745152354240417, "num_tokens": 623476291.0, "step": 16421 }, { "epoch": 2.089047195013357, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.184678316116333, "learning_rate": 1e-05, "loss": 0.3924, "mean_token_accuracy": 0.8920132517814636, "num_tokens": 623513830.0, "step": 16422 }, { "epoch": 2.0891744052919474, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.100407600402832, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8753399848937988, "num_tokens": 623554789.0, "step": 16423 }, { "epoch": 2.089301615570538, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2866342067718506, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.883398175239563, "num_tokens": 623593646.0, "step": 16424 }, { "epoch": 2.0894288258491285, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.240337371826172, "learning_rate": 1e-05, "loss": 0.4084, "mean_token_accuracy": 0.8899214267730713, "num_tokens": 623630646.0, "step": 16425 }, { "epoch": 2.089556036127719, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.4964640140533447, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.86090487241745, "num_tokens": 623675488.0, "step": 16426 }, { "epoch": 2.0896832464063095, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.212153434753418, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8712197542190552, "num_tokens": 623719717.0, "step": 16427 }, { "epoch": 2.0898104566849, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6596548557281494, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8879841566085815, "num_tokens": 623753248.0, "step": 16428 }, { "epoch": 2.0899376669634906, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0592803955078125, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8670468926429749, "num_tokens": 623791562.0, "step": 16429 }, { "epoch": 2.090064877242081, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6953229904174805, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8674582242965698, "num_tokens": 623834187.0, "step": 16430 }, { "epoch": 2.0901920875206716, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.862661123275757, "learning_rate": 1e-05, "loss": 0.4811, "mean_token_accuracy": 0.8682861328125, "num_tokens": 623869664.0, "step": 16431 }, { "epoch": 2.090319297799262, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.647024154663086, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8791709542274475, "num_tokens": 623907461.0, "step": 16432 }, { "epoch": 2.0904465080778527, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.69891095161438, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8778396248817444, "num_tokens": 623947679.0, "step": 16433 }, { "epoch": 2.0905737183564432, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5620322227478027, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8860516548156738, "num_tokens": 623992659.0, "step": 16434 }, { "epoch": 2.0907009286350338, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.617445230484009, "learning_rate": 1e-05, "loss": 0.3849, "mean_token_accuracy": 0.892784833908081, "num_tokens": 624028305.0, "step": 16435 }, { "epoch": 2.0908281389136243, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5036377906799316, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.881974458694458, "num_tokens": 624062585.0, "step": 16436 }, { "epoch": 2.090955349192215, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.577954053878784, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8837989568710327, "num_tokens": 624100525.0, "step": 16437 }, { "epoch": 2.0910825594708053, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4231796264648438, "learning_rate": 1e-05, "loss": 0.3886, "mean_token_accuracy": 0.8896676301956177, "num_tokens": 624135982.0, "step": 16438 }, { "epoch": 2.091209769749396, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.570129871368408, "learning_rate": 1e-05, "loss": 0.3883, "mean_token_accuracy": 0.8897230625152588, "num_tokens": 624172867.0, "step": 16439 }, { "epoch": 2.0913369800279864, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3538312911987305, "learning_rate": 1e-05, "loss": 0.3666, "mean_token_accuracy": 0.8957270383834839, "num_tokens": 624208588.0, "step": 16440 }, { "epoch": 2.091464190306577, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.640713691711426, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.876848042011261, "num_tokens": 624241060.0, "step": 16441 }, { "epoch": 2.0915914005851675, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2777135372161865, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8749116659164429, "num_tokens": 624284508.0, "step": 16442 }, { "epoch": 2.091718610863758, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7303202152252197, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8834050893783569, "num_tokens": 624326787.0, "step": 16443 }, { "epoch": 2.0918458211423485, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2086398601531982, "learning_rate": 1e-05, "loss": 0.3855, "mean_token_accuracy": 0.8902976512908936, "num_tokens": 624363375.0, "step": 16444 }, { "epoch": 2.0919730314209386, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9010777473449707, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8747138977050781, "num_tokens": 624405508.0, "step": 16445 }, { "epoch": 2.092100241699529, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1983461380004883, "learning_rate": 1e-05, "loss": 0.3802, "mean_token_accuracy": 0.8902862071990967, "num_tokens": 624438505.0, "step": 16446 }, { "epoch": 2.0922274519781197, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.084301948547363, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8826314210891724, "num_tokens": 624475463.0, "step": 16447 }, { "epoch": 2.09235466225671, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.299325942993164, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8872079849243164, "num_tokens": 624511577.0, "step": 16448 }, { "epoch": 2.0924818725353007, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.208238124847412, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.882798433303833, "num_tokens": 624552213.0, "step": 16449 }, { "epoch": 2.0926090828138912, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.493380308151245, "learning_rate": 1e-05, "loss": 0.4026, "mean_token_accuracy": 0.8876854181289673, "num_tokens": 624589026.0, "step": 16450 }, { "epoch": 2.0927362930924818, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.166726589202881, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8790041208267212, "num_tokens": 624623245.0, "step": 16451 }, { "epoch": 2.0928635033710723, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6518640518188477, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8765813112258911, "num_tokens": 624659713.0, "step": 16452 }, { "epoch": 2.092990713649663, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.986922264099121, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8843761682510376, "num_tokens": 624695539.0, "step": 16453 }, { "epoch": 2.0931179239282534, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6670310497283936, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8740417957305908, "num_tokens": 624735311.0, "step": 16454 }, { "epoch": 2.093245134206844, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.785588026046753, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8850189447402954, "num_tokens": 624771941.0, "step": 16455 }, { "epoch": 2.0933723444854344, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.622054100036621, "learning_rate": 1e-05, "loss": 0.3873, "mean_token_accuracy": 0.8922522068023682, "num_tokens": 624809963.0, "step": 16456 }, { "epoch": 2.093499554764025, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6271514892578125, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8791934251785278, "num_tokens": 624849839.0, "step": 16457 }, { "epoch": 2.0936267650426155, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5872161388397217, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8803834915161133, "num_tokens": 624883920.0, "step": 16458 }, { "epoch": 2.093753975321206, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5246260166168213, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8882849812507629, "num_tokens": 624916024.0, "step": 16459 }, { "epoch": 2.0938811855997965, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5402109622955322, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.884602427482605, "num_tokens": 624950978.0, "step": 16460 }, { "epoch": 2.094008395878387, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4501967430114746, "learning_rate": 1e-05, "loss": 0.3812, "mean_token_accuracy": 0.8902063965797424, "num_tokens": 624992177.0, "step": 16461 }, { "epoch": 2.0941356061569776, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.520772695541382, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8822914361953735, "num_tokens": 625028211.0, "step": 16462 }, { "epoch": 2.094262816435568, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4048845767974854, "learning_rate": 1e-05, "loss": 0.3685, "mean_token_accuracy": 0.897498369216919, "num_tokens": 625065895.0, "step": 16463 }, { "epoch": 2.0943900267141586, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5147006511688232, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.8887930512428284, "num_tokens": 625100521.0, "step": 16464 }, { "epoch": 2.094517236992749, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.362436532974243, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8820520043373108, "num_tokens": 625144810.0, "step": 16465 }, { "epoch": 2.0946444472713397, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.552581787109375, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.878401517868042, "num_tokens": 625178001.0, "step": 16466 }, { "epoch": 2.09477165754993, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.318669080734253, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8820750713348389, "num_tokens": 625214857.0, "step": 16467 }, { "epoch": 2.0948988678285207, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.594647169113159, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8831655383110046, "num_tokens": 625253396.0, "step": 16468 }, { "epoch": 2.095026078107111, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.245208501815796, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8675049543380737, "num_tokens": 625291922.0, "step": 16469 }, { "epoch": 2.0951532883857014, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7148497104644775, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8795660734176636, "num_tokens": 625326548.0, "step": 16470 }, { "epoch": 2.095280498664292, "ewc_loss": 0.07470703125, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.174619436264038, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8870217800140381, "num_tokens": 625361354.0, "step": 16471 }, { "epoch": 2.0954077089428824, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002880096435546875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8854804039001465, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8714593648910522, "num_tokens": 625400885.0, "step": 16472 }, { "epoch": 2.095534919221473, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1179044246673584, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8834117650985718, "num_tokens": 625438975.0, "step": 16473 }, { "epoch": 2.0956621295000635, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.161906719207764, "learning_rate": 1e-05, "loss": 0.4903, "mean_token_accuracy": 0.862725019454956, "num_tokens": 625475384.0, "step": 16474 }, { "epoch": 2.095789339778654, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.2253546714782715, "learning_rate": 1e-05, "loss": 0.3772, "mean_token_accuracy": 0.8960190415382385, "num_tokens": 625508422.0, "step": 16475 }, { "epoch": 2.0959165500572445, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.401175498962402, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8892633318901062, "num_tokens": 625546434.0, "step": 16476 }, { "epoch": 2.096043760335835, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.533681631088257, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.885142982006073, "num_tokens": 625584005.0, "step": 16477 }, { "epoch": 2.0961709706144256, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.418659210205078, "learning_rate": 1e-05, "loss": 0.3892, "mean_token_accuracy": 0.8978227376937866, "num_tokens": 625616713.0, "step": 16478 }, { "epoch": 2.096298180893016, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.8100664615631104, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8827632665634155, "num_tokens": 625657738.0, "step": 16479 }, { "epoch": 2.0964253911716066, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.188971996307373, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8784555196762085, "num_tokens": 625693485.0, "step": 16480 }, { "epoch": 2.096552601450197, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.828490734100342, "learning_rate": 1e-05, "loss": 0.4004, "mean_token_accuracy": 0.8883341550827026, "num_tokens": 625727219.0, "step": 16481 }, { "epoch": 2.0966798117287877, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9745349884033203, "learning_rate": 1e-05, "loss": 0.3867, "mean_token_accuracy": 0.8939485549926758, "num_tokens": 625762858.0, "step": 16482 }, { "epoch": 2.0968070220073782, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.7292726039886475, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.879173755645752, "num_tokens": 625799752.0, "step": 16483 }, { "epoch": 2.0969342322859688, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8083131313323975, "learning_rate": 1e-05, "loss": 0.3824, "mean_token_accuracy": 0.8931017518043518, "num_tokens": 625835453.0, "step": 16484 }, { "epoch": 2.0970614425645593, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6239359378814697, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8669419884681702, "num_tokens": 625875181.0, "step": 16485 }, { "epoch": 2.09718865284315, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.707892894744873, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8930214047431946, "num_tokens": 625918851.0, "step": 16486 }, { "epoch": 2.0973158631217403, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5103883743286133, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8831245303153992, "num_tokens": 625954859.0, "step": 16487 }, { "epoch": 2.097443073400331, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.692155122756958, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8864161968231201, "num_tokens": 625986042.0, "step": 16488 }, { "epoch": 2.0975702836789214, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4188058376312256, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8607296943664551, "num_tokens": 626023604.0, "step": 16489 }, { "epoch": 2.097697493957512, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.714465618133545, "learning_rate": 1e-05, "loss": 0.3705, "mean_token_accuracy": 0.8985741138458252, "num_tokens": 626059193.0, "step": 16490 }, { "epoch": 2.0978247042361025, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.313483238220215, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8745743036270142, "num_tokens": 626099994.0, "step": 16491 }, { "epoch": 2.097951914514693, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.823431968688965, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8809726238250732, "num_tokens": 626133105.0, "step": 16492 }, { "epoch": 2.0980791247932835, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.255028009414673, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8895102143287659, "num_tokens": 626171837.0, "step": 16493 }, { "epoch": 2.0982063350718736, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.97825026512146, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8769473433494568, "num_tokens": 626214057.0, "step": 16494 }, { "epoch": 2.098333545350464, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.266697406768799, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8862102627754211, "num_tokens": 626249046.0, "step": 16495 }, { "epoch": 2.0984607556290547, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.163125514984131, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8695232272148132, "num_tokens": 626283991.0, "step": 16496 }, { "epoch": 2.098587965907645, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.415860176086426, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8786981105804443, "num_tokens": 626321919.0, "step": 16497 }, { "epoch": 2.0987151761862357, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.193819999694824, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8936424851417542, "num_tokens": 626360281.0, "step": 16498 }, { "epoch": 2.0988423864648262, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6034936904907227, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8825235366821289, "num_tokens": 626404314.0, "step": 16499 }, { "epoch": 2.0989695967434168, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.046019077301025, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8665494918823242, "num_tokens": 626452743.0, "step": 16500 }, { "epoch": 2.0990968070220073, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6814587116241455, "learning_rate": 1e-05, "loss": 0.3843, "mean_token_accuracy": 0.8971247673034668, "num_tokens": 626494310.0, "step": 16501 }, { "epoch": 2.099224017300598, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8291192054748535, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8828147053718567, "num_tokens": 626534995.0, "step": 16502 }, { "epoch": 2.0993512275791884, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.667200803756714, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.8900383114814758, "num_tokens": 626575084.0, "step": 16503 }, { "epoch": 2.099478437857779, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6358187198638916, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8876360058784485, "num_tokens": 626608922.0, "step": 16504 }, { "epoch": 2.0996056481363694, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.640805959701538, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8896838426589966, "num_tokens": 626645480.0, "step": 16505 }, { "epoch": 2.09973285841496, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.47916579246521, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8746356964111328, "num_tokens": 626687641.0, "step": 16506 }, { "epoch": 2.0998600686935505, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6306698322296143, "learning_rate": 1e-05, "loss": 0.3861, "mean_token_accuracy": 0.8919947147369385, "num_tokens": 626720624.0, "step": 16507 }, { "epoch": 2.099987278972141, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3815383911132812, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.85959792137146, "num_tokens": 626760174.0, "step": 16508 }, { "epoch": 2.1001144892507315, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6455001831054688, "learning_rate": 1e-05, "loss": 0.3634, "mean_token_accuracy": 0.8966070413589478, "num_tokens": 626792970.0, "step": 16509 }, { "epoch": 2.100241699529322, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.31592059135437, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8795970678329468, "num_tokens": 626826793.0, "step": 16510 }, { "epoch": 2.1003689098079126, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7015528678894043, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8838255405426025, "num_tokens": 626862938.0, "step": 16511 }, { "epoch": 2.100496120086503, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2438673973083496, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8827073574066162, "num_tokens": 626900738.0, "step": 16512 }, { "epoch": 2.1006233303650936, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.898714303970337, "learning_rate": 1e-05, "loss": 0.3948, "mean_token_accuracy": 0.8904730081558228, "num_tokens": 626932670.0, "step": 16513 }, { "epoch": 2.100750540643684, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1976654529571533, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8781984448432922, "num_tokens": 626971683.0, "step": 16514 }, { "epoch": 2.1008777509222747, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0463409423828125, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8739063739776611, "num_tokens": 627009955.0, "step": 16515 }, { "epoch": 2.101004961200865, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2450780868530273, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.868822455406189, "num_tokens": 627052469.0, "step": 16516 }, { "epoch": 2.1011321714794557, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.244784832000732, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8790767788887024, "num_tokens": 627088314.0, "step": 16517 }, { "epoch": 2.1012593817580463, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.426800012588501, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8822349905967712, "num_tokens": 627132030.0, "step": 16518 }, { "epoch": 2.1013865920366364, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.277359962463379, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8772664666175842, "num_tokens": 627171135.0, "step": 16519 }, { "epoch": 2.101513802315227, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6521718502044678, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8836931586265564, "num_tokens": 627206303.0, "step": 16520 }, { "epoch": 2.1016410125938174, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.131605625152588, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8784507513046265, "num_tokens": 627242859.0, "step": 16521 }, { "epoch": 2.101768222872408, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.724071979522705, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8772925138473511, "num_tokens": 627284659.0, "step": 16522 }, { "epoch": 2.1018954331509985, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.917470693588257, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.8876740336418152, "num_tokens": 627320012.0, "step": 16523 }, { "epoch": 2.102022643429589, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.681063175201416, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8846946954727173, "num_tokens": 627358117.0, "step": 16524 }, { "epoch": 2.1021498537081795, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.764165163040161, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8685424327850342, "num_tokens": 627393331.0, "step": 16525 }, { "epoch": 2.10227706398677, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6074540615081787, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8706805109977722, "num_tokens": 627433905.0, "step": 16526 }, { "epoch": 2.1024042742653606, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.633553981781006, "learning_rate": 1e-05, "loss": 0.3922, "mean_token_accuracy": 0.8942743539810181, "num_tokens": 627469673.0, "step": 16527 }, { "epoch": 2.102531484543951, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5460269451141357, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8774925470352173, "num_tokens": 627502917.0, "step": 16528 }, { "epoch": 2.1026586948225416, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.563985586166382, "learning_rate": 1e-05, "loss": 0.3847, "mean_token_accuracy": 0.8945420980453491, "num_tokens": 627536632.0, "step": 16529 }, { "epoch": 2.102785905101132, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.465897560119629, "learning_rate": 1e-05, "loss": 0.3944, "mean_token_accuracy": 0.8883165121078491, "num_tokens": 627580820.0, "step": 16530 }, { "epoch": 2.1029131153797227, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.54268479347229, "learning_rate": 1e-05, "loss": 0.366, "mean_token_accuracy": 0.8998614549636841, "num_tokens": 627616895.0, "step": 16531 }, { "epoch": 2.1030403256583132, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4061129093170166, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8748202323913574, "num_tokens": 627655648.0, "step": 16532 }, { "epoch": 2.1031675359369038, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.562006711959839, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.881196141242981, "num_tokens": 627692259.0, "step": 16533 }, { "epoch": 2.1032947462154943, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3435018062591553, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8849249482154846, "num_tokens": 627731601.0, "step": 16534 }, { "epoch": 2.103421956494085, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6029858589172363, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8809322118759155, "num_tokens": 627768381.0, "step": 16535 }, { "epoch": 2.1035491667726753, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2708382606506348, "learning_rate": 1e-05, "loss": 0.3955, "mean_token_accuracy": 0.8876348733901978, "num_tokens": 627808327.0, "step": 16536 }, { "epoch": 2.103676377051266, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7172305583953857, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8832718133926392, "num_tokens": 627838779.0, "step": 16537 }, { "epoch": 2.1038035873298564, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.200044631958008, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8798978328704834, "num_tokens": 627875634.0, "step": 16538 }, { "epoch": 2.103930797608447, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.873548984527588, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8811924457550049, "num_tokens": 627918350.0, "step": 16539 }, { "epoch": 2.1040580078870375, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1738810539245605, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8782422542572021, "num_tokens": 627955649.0, "step": 16540 }, { "epoch": 2.104185218165628, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.100442409515381, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8815562725067139, "num_tokens": 627989374.0, "step": 16541 }, { "epoch": 2.1043124284442185, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3948564529418945, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8705167174339294, "num_tokens": 628026338.0, "step": 16542 }, { "epoch": 2.1044396387228086, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.141683578491211, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8873693943023682, "num_tokens": 628062550.0, "step": 16543 }, { "epoch": 2.104566849001399, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3944766521453857, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8761263489723206, "num_tokens": 628104868.0, "step": 16544 }, { "epoch": 2.1046940592799896, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.16522741317749, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8747706413269043, "num_tokens": 628142009.0, "step": 16545 }, { "epoch": 2.10482126955858, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.549039840698242, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8788108229637146, "num_tokens": 628180992.0, "step": 16546 }, { "epoch": 2.1049484798371707, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.078861713409424, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8820161819458008, "num_tokens": 628215636.0, "step": 16547 }, { "epoch": 2.1050756901157612, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6188154220581055, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.889288604259491, "num_tokens": 628256993.0, "step": 16548 }, { "epoch": 2.1052029003943518, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.921318769454956, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8783004879951477, "num_tokens": 628293408.0, "step": 16549 }, { "epoch": 2.1053301106729423, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.625861406326294, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8833521008491516, "num_tokens": 628329353.0, "step": 16550 }, { "epoch": 2.105457320951533, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.766779899597168, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.869530439376831, "num_tokens": 628364312.0, "step": 16551 }, { "epoch": 2.1055845312301233, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5741164684295654, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8846207857131958, "num_tokens": 628401354.0, "step": 16552 }, { "epoch": 2.105711741508714, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6521499156951904, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8886007070541382, "num_tokens": 628432840.0, "step": 16553 }, { "epoch": 2.1058389517873044, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5164291858673096, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8774126768112183, "num_tokens": 628476300.0, "step": 16554 }, { "epoch": 2.105966162065895, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5718419551849365, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8805094957351685, "num_tokens": 628514248.0, "step": 16555 }, { "epoch": 2.1060933723444855, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.455474615097046, "learning_rate": 1e-05, "loss": 0.4044, "mean_token_accuracy": 0.8845914602279663, "num_tokens": 628557161.0, "step": 16556 }, { "epoch": 2.106220582623076, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5503592491149902, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8873447179794312, "num_tokens": 628591073.0, "step": 16557 }, { "epoch": 2.1063477929016665, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4082515239715576, "learning_rate": 1e-05, "loss": 0.3994, "mean_token_accuracy": 0.8861531019210815, "num_tokens": 628627215.0, "step": 16558 }, { "epoch": 2.106475003180257, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.553121566772461, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8798867464065552, "num_tokens": 628669858.0, "step": 16559 }, { "epoch": 2.1066022134588476, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3544909954071045, "learning_rate": 1e-05, "loss": 0.3671, "mean_token_accuracy": 0.8977096080780029, "num_tokens": 628704504.0, "step": 16560 }, { "epoch": 2.106729423737438, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.603273630142212, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8674157857894897, "num_tokens": 628743444.0, "step": 16561 }, { "epoch": 2.1068566340160286, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2975683212280273, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8695584535598755, "num_tokens": 628782256.0, "step": 16562 }, { "epoch": 2.106983844294619, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6726274490356445, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8857695460319519, "num_tokens": 628822150.0, "step": 16563 }, { "epoch": 2.1071110545732097, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.222019672393799, "learning_rate": 1e-05, "loss": 0.3833, "mean_token_accuracy": 0.8925886750221252, "num_tokens": 628862445.0, "step": 16564 }, { "epoch": 2.1072382648518, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8236823081970215, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.882651686668396, "num_tokens": 628897166.0, "step": 16565 }, { "epoch": 2.1073654751303907, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.181001901626587, "learning_rate": 1e-05, "loss": 0.3934, "mean_token_accuracy": 0.8872632384300232, "num_tokens": 628933443.0, "step": 16566 }, { "epoch": 2.107492685408981, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0123114585876465, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8891984820365906, "num_tokens": 628969614.0, "step": 16567 }, { "epoch": 2.1076198956875714, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2032968997955322, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8850849866867065, "num_tokens": 629010386.0, "step": 16568 }, { "epoch": 2.107747105966162, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2478556632995605, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8740723729133606, "num_tokens": 629041388.0, "step": 16569 }, { "epoch": 2.1078743162447524, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.4208264350891113, "learning_rate": 1e-05, "loss": 0.3903, "mean_token_accuracy": 0.8921995162963867, "num_tokens": 629077306.0, "step": 16570 }, { "epoch": 2.108001526523343, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.278481483459473, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8895200490951538, "num_tokens": 629110636.0, "step": 16571 }, { "epoch": 2.1081287368019335, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.634730339050293, "learning_rate": 1e-05, "loss": 0.3765, "mean_token_accuracy": 0.8994891047477722, "num_tokens": 629145834.0, "step": 16572 }, { "epoch": 2.108255947080524, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.137914180755615, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8789833784103394, "num_tokens": 629187148.0, "step": 16573 }, { "epoch": 2.1083831573591145, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.711852550506592, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8919149041175842, "num_tokens": 629224391.0, "step": 16574 }, { "epoch": 2.108510367637705, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9466552734375, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.884566068649292, "num_tokens": 629261074.0, "step": 16575 }, { "epoch": 2.1086375779162956, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6849117279052734, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8816779851913452, "num_tokens": 629296568.0, "step": 16576 }, { "epoch": 2.108764788194886, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7584540843963623, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8849101066589355, "num_tokens": 629342784.0, "step": 16577 }, { "epoch": 2.1088919984734766, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.609572172164917, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8882511258125305, "num_tokens": 629377705.0, "step": 16578 }, { "epoch": 2.109019208752067, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6393439769744873, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8866446018218994, "num_tokens": 629414823.0, "step": 16579 }, { "epoch": 2.1091464190306577, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.542017936706543, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.871478796005249, "num_tokens": 629452404.0, "step": 16580 }, { "epoch": 2.109273629309248, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5699124336242676, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.8861016035079956, "num_tokens": 629486147.0, "step": 16581 }, { "epoch": 2.1094008395878387, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.470705986022949, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8869343996047974, "num_tokens": 629523581.0, "step": 16582 }, { "epoch": 2.1095280498664293, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5370705127716064, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.8828735947608948, "num_tokens": 629562125.0, "step": 16583 }, { "epoch": 2.10965526014502, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4021928310394287, "learning_rate": 1e-05, "loss": 0.3898, "mean_token_accuracy": 0.8919327855110168, "num_tokens": 629601786.0, "step": 16584 }, { "epoch": 2.1097824704236103, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5420076847076416, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8891977667808533, "num_tokens": 629645022.0, "step": 16585 }, { "epoch": 2.109909680702201, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3557467460632324, "learning_rate": 1e-05, "loss": 0.3754, "mean_token_accuracy": 0.8943653106689453, "num_tokens": 629675604.0, "step": 16586 }, { "epoch": 2.1100368909807914, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5850274562835693, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8751845955848694, "num_tokens": 629718508.0, "step": 16587 }, { "epoch": 2.110164101259382, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2956771850585938, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8653246760368347, "num_tokens": 629758476.0, "step": 16588 }, { "epoch": 2.1102913115379724, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6760096549987793, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8755044341087341, "num_tokens": 629793692.0, "step": 16589 }, { "epoch": 2.110418521816563, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.2297158241271973, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8750723600387573, "num_tokens": 629829750.0, "step": 16590 }, { "epoch": 2.1105457320951535, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.822486400604248, "learning_rate": 1e-05, "loss": 0.4751, "mean_token_accuracy": 0.8640559315681458, "num_tokens": 629865832.0, "step": 16591 }, { "epoch": 2.1106729423737436, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1760051250457764, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8723175525665283, "num_tokens": 629906450.0, "step": 16592 }, { "epoch": 2.110800152652334, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.027074337005615, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8674622774124146, "num_tokens": 629939650.0, "step": 16593 }, { "epoch": 2.1109273629309246, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.202293634414673, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8832060098648071, "num_tokens": 629977692.0, "step": 16594 }, { "epoch": 2.111054573209515, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.233309745788574, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8835296034812927, "num_tokens": 630019111.0, "step": 16595 }, { "epoch": 2.1111817834881057, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.3849146366119385, "learning_rate": 1e-05, "loss": 0.3869, "mean_token_accuracy": 0.8891536593437195, "num_tokens": 630054924.0, "step": 16596 }, { "epoch": 2.1113089937666962, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.325327396392822, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.87458336353302, "num_tokens": 630100344.0, "step": 16597 }, { "epoch": 2.1114362040452868, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.642616033554077, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8817756772041321, "num_tokens": 630137296.0, "step": 16598 }, { "epoch": 2.1115634143238773, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.200359344482422, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.884223997592926, "num_tokens": 630177477.0, "step": 16599 }, { "epoch": 2.111690624602468, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7481987476348877, "learning_rate": 1e-05, "loss": 0.3731, "mean_token_accuracy": 0.899124026298523, "num_tokens": 630205830.0, "step": 16600 }, { "epoch": 2.1118178348810583, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9940764904022217, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.881180465221405, "num_tokens": 630247114.0, "step": 16601 }, { "epoch": 2.111945045159649, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.718803882598877, "learning_rate": 1e-05, "loss": 0.3942, "mean_token_accuracy": 0.8940521478652954, "num_tokens": 630279848.0, "step": 16602 }, { "epoch": 2.1120722554382394, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.800314426422119, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8920282125473022, "num_tokens": 630320993.0, "step": 16603 }, { "epoch": 2.11219946571683, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.646505117416382, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8603351712226868, "num_tokens": 630354703.0, "step": 16604 }, { "epoch": 2.1123266759954205, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.672653913497925, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8765140175819397, "num_tokens": 630388902.0, "step": 16605 }, { "epoch": 2.112453886274011, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5388071537017822, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8839545249938965, "num_tokens": 630431446.0, "step": 16606 }, { "epoch": 2.1125810965526015, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.610896348953247, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8774700164794922, "num_tokens": 630468334.0, "step": 16607 }, { "epoch": 2.112708306831192, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4736487865448, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8869456052780151, "num_tokens": 630498607.0, "step": 16608 }, { "epoch": 2.1128355171097826, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.593153715133667, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8840569257736206, "num_tokens": 630534900.0, "step": 16609 }, { "epoch": 2.112962727388373, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.391542673110962, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8718180060386658, "num_tokens": 630577045.0, "step": 16610 }, { "epoch": 2.1130899376669636, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.617265462875366, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8850119113922119, "num_tokens": 630613477.0, "step": 16611 }, { "epoch": 2.113217147945554, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3149983882904053, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.881296694278717, "num_tokens": 630651868.0, "step": 16612 }, { "epoch": 2.1133443582241447, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.69146466255188, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8842687010765076, "num_tokens": 630691483.0, "step": 16613 }, { "epoch": 2.113471568502735, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.239438533782959, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.885079562664032, "num_tokens": 630730392.0, "step": 16614 }, { "epoch": 2.1135987787813257, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.833141565322876, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.8879398107528687, "num_tokens": 630764336.0, "step": 16615 }, { "epoch": 2.1137259890599163, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.190093994140625, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.873374879360199, "num_tokens": 630805912.0, "step": 16616 }, { "epoch": 2.1138531993385064, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.045462131500244, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8751482367515564, "num_tokens": 630844667.0, "step": 16617 }, { "epoch": 2.113980409617097, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2340526580810547, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.882299542427063, "num_tokens": 630888217.0, "step": 16618 }, { "epoch": 2.1141076198956874, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.23682165145874, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8920680284500122, "num_tokens": 630922415.0, "step": 16619 }, { "epoch": 2.114234830174278, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4617063999176025, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8829766511917114, "num_tokens": 630955469.0, "step": 16620 }, { "epoch": 2.1143620404528685, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.240487575531006, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8855181932449341, "num_tokens": 630994641.0, "step": 16621 }, { "epoch": 2.114489250731459, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.656386137008667, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8719402551651001, "num_tokens": 631034671.0, "step": 16622 }, { "epoch": 2.1146164610100495, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.078647136688232, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8771228194236755, "num_tokens": 631073518.0, "step": 16623 }, { "epoch": 2.11474367128864, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.71720027923584, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8646212816238403, "num_tokens": 631111631.0, "step": 16624 }, { "epoch": 2.1148708815672306, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.869540214538574, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8797199130058289, "num_tokens": 631148081.0, "step": 16625 }, { "epoch": 2.114998091845821, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6764416694641113, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8764655590057373, "num_tokens": 631183176.0, "step": 16626 }, { "epoch": 2.1151253021244116, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.694889545440674, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8829922676086426, "num_tokens": 631213841.0, "step": 16627 }, { "epoch": 2.115252512403002, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6131317615509033, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8679261803627014, "num_tokens": 631249410.0, "step": 16628 }, { "epoch": 2.1153797226815927, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.567931652069092, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.878761887550354, "num_tokens": 631292153.0, "step": 16629 }, { "epoch": 2.115506932960183, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.554264545440674, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8796562552452087, "num_tokens": 631330461.0, "step": 16630 }, { "epoch": 2.1156341432387737, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4989097118377686, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8759278059005737, "num_tokens": 631367384.0, "step": 16631 }, { "epoch": 2.1157613535173643, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.51552677154541, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8750612139701843, "num_tokens": 631402558.0, "step": 16632 }, { "epoch": 2.115888563795955, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4543347358703613, "learning_rate": 1e-05, "loss": 0.4681, "mean_token_accuracy": 0.860844075679779, "num_tokens": 631444040.0, "step": 16633 }, { "epoch": 2.1160157740745453, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002899169921875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4773354530334473, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.8887917995452881, "num_tokens": 631486350.0, "step": 16634 }, { "epoch": 2.116142984353136, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4397354125976562, "learning_rate": 1e-05, "loss": 0.3864, "mean_token_accuracy": 0.8886678218841553, "num_tokens": 631519145.0, "step": 16635 }, { "epoch": 2.1162701946317264, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.460343837738037, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8764608502388, "num_tokens": 631554637.0, "step": 16636 }, { "epoch": 2.116397404910317, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4121193885803223, "learning_rate": 1e-05, "loss": 0.3896, "mean_token_accuracy": 0.88949054479599, "num_tokens": 631593580.0, "step": 16637 }, { "epoch": 2.1165246151889074, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4507079124450684, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8787989616394043, "num_tokens": 631630410.0, "step": 16638 }, { "epoch": 2.116651825467498, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4199044704437256, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8829094767570496, "num_tokens": 631661894.0, "step": 16639 }, { "epoch": 2.116779035746088, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4453601837158203, "learning_rate": 1e-05, "loss": 0.4024, "mean_token_accuracy": 0.8804324865341187, "num_tokens": 631694514.0, "step": 16640 }, { "epoch": 2.1169062460246786, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3905887603759766, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8719399571418762, "num_tokens": 631737715.0, "step": 16641 }, { "epoch": 2.117033456303269, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4526894092559814, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8709651231765747, "num_tokens": 631773972.0, "step": 16642 }, { "epoch": 2.1171606665818596, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3702383041381836, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8780895471572876, "num_tokens": 631812292.0, "step": 16643 }, { "epoch": 2.11728787686045, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4608993530273438, "learning_rate": 1e-05, "loss": 0.386, "mean_token_accuracy": 0.8910174369812012, "num_tokens": 631849518.0, "step": 16644 }, { "epoch": 2.1174150871390407, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.341512680053711, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.8816027641296387, "num_tokens": 631889821.0, "step": 16645 }, { "epoch": 2.1175422974176312, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.499434471130371, "learning_rate": 1e-05, "loss": 0.3782, "mean_token_accuracy": 0.8930233716964722, "num_tokens": 631925192.0, "step": 16646 }, { "epoch": 2.1176695076962218, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2888388633728027, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8809579610824585, "num_tokens": 631967876.0, "step": 16647 }, { "epoch": 2.1177967179748123, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.574793815612793, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8853647112846375, "num_tokens": 632006026.0, "step": 16648 }, { "epoch": 2.117923928253403, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.212857723236084, "learning_rate": 1e-05, "loss": 0.3729, "mean_token_accuracy": 0.8949415683746338, "num_tokens": 632045886.0, "step": 16649 }, { "epoch": 2.1180511385319933, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.70768404006958, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8867553472518921, "num_tokens": 632083583.0, "step": 16650 }, { "epoch": 2.118178348810584, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1300671100616455, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8857840299606323, "num_tokens": 632124427.0, "step": 16651 }, { "epoch": 2.1183055590891744, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9595565795898438, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8772475719451904, "num_tokens": 632159787.0, "step": 16652 }, { "epoch": 2.118432769367765, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.067124128341675, "learning_rate": 1e-05, "loss": 0.3843, "mean_token_accuracy": 0.8884874582290649, "num_tokens": 632194737.0, "step": 16653 }, { "epoch": 2.1185599796463555, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.3259596824646, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8856470584869385, "num_tokens": 632230737.0, "step": 16654 }, { "epoch": 2.118687189924946, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.2483537197113037, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8786946535110474, "num_tokens": 632271374.0, "step": 16655 }, { "epoch": 2.1188144002035365, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 4.6587724685668945, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8887360095977783, "num_tokens": 632305326.0, "step": 16656 }, { "epoch": 2.118941610482127, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.803358316421509, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8789371252059937, "num_tokens": 632338807.0, "step": 16657 }, { "epoch": 2.1190688207607176, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.549505233764648, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8689062595367432, "num_tokens": 632373933.0, "step": 16658 }, { "epoch": 2.119196031039308, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.063998699188232, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.879906415939331, "num_tokens": 632414149.0, "step": 16659 }, { "epoch": 2.1193232413178986, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1598029136657715, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8863030672073364, "num_tokens": 632453622.0, "step": 16660 }, { "epoch": 2.119450451596489, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.9338221549987793, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8692286014556885, "num_tokens": 632488398.0, "step": 16661 }, { "epoch": 2.1195776618750797, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9763169288635254, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8767859935760498, "num_tokens": 632524013.0, "step": 16662 }, { "epoch": 2.11970487215367, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7566161155700684, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8766883611679077, "num_tokens": 632562247.0, "step": 16663 }, { "epoch": 2.1198320824322607, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.853827953338623, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8721562623977661, "num_tokens": 632600066.0, "step": 16664 }, { "epoch": 2.119959292710851, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6042845249176025, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8758720755577087, "num_tokens": 632638157.0, "step": 16665 }, { "epoch": 2.1200865029894413, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.791372537612915, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8588175177574158, "num_tokens": 632682534.0, "step": 16666 }, { "epoch": 2.120213713268032, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4664533138275146, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8883125185966492, "num_tokens": 632717574.0, "step": 16667 }, { "epoch": 2.1203409235466224, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.815873622894287, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.8845255374908447, "num_tokens": 632754113.0, "step": 16668 }, { "epoch": 2.120468133825213, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3521969318389893, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8683733940124512, "num_tokens": 632797776.0, "step": 16669 }, { "epoch": 2.1205953441038035, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.933600664138794, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8799979090690613, "num_tokens": 632834841.0, "step": 16670 }, { "epoch": 2.120722554382394, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.287296772003174, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8888401985168457, "num_tokens": 632872698.0, "step": 16671 }, { "epoch": 2.1208497646609845, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1222124099731445, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8792296648025513, "num_tokens": 632911067.0, "step": 16672 }, { "epoch": 2.120976974939575, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3602654933929443, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8862094283103943, "num_tokens": 632952037.0, "step": 16673 }, { "epoch": 2.1211041852181656, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.276106834411621, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8763777613639832, "num_tokens": 632984667.0, "step": 16674 }, { "epoch": 2.121231395496756, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5931143760681152, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8707266449928284, "num_tokens": 633027165.0, "step": 16675 }, { "epoch": 2.1213586057753466, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.18272066116333, "learning_rate": 1e-05, "loss": 0.4039, "mean_token_accuracy": 0.8911468982696533, "num_tokens": 633068636.0, "step": 16676 }, { "epoch": 2.121485816053937, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7459652423858643, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8892155885696411, "num_tokens": 633103546.0, "step": 16677 }, { "epoch": 2.1216130263325277, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9594175815582275, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8721214532852173, "num_tokens": 633141529.0, "step": 16678 }, { "epoch": 2.121740236611118, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.744126081466675, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.8907639980316162, "num_tokens": 633177324.0, "step": 16679 }, { "epoch": 2.1218674468897087, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7135062217712402, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8820369839668274, "num_tokens": 633216726.0, "step": 16680 }, { "epoch": 2.1219946571682993, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6967732906341553, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8764955401420593, "num_tokens": 633255583.0, "step": 16681 }, { "epoch": 2.12212186744689, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5266246795654297, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8849679231643677, "num_tokens": 633295167.0, "step": 16682 }, { "epoch": 2.1222490777254803, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6701393127441406, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8672246336936951, "num_tokens": 633336662.0, "step": 16683 }, { "epoch": 2.122376288004071, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.392648935317993, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8858774900436401, "num_tokens": 633379826.0, "step": 16684 }, { "epoch": 2.1225034982826614, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6924221515655518, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8702701926231384, "num_tokens": 633421447.0, "step": 16685 }, { "epoch": 2.122630708561252, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.287499189376831, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.8888530731201172, "num_tokens": 633462704.0, "step": 16686 }, { "epoch": 2.1227579188398424, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.776247978210449, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8806272745132446, "num_tokens": 633496452.0, "step": 16687 }, { "epoch": 2.122885129118433, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2143588066101074, "learning_rate": 1e-05, "loss": 0.3638, "mean_token_accuracy": 0.8992900252342224, "num_tokens": 633534560.0, "step": 16688 }, { "epoch": 2.1230123393970235, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.925143003463745, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8860859870910645, "num_tokens": 633572429.0, "step": 16689 }, { "epoch": 2.1231395496756136, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.222109079360962, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8730084896087646, "num_tokens": 633603588.0, "step": 16690 }, { "epoch": 2.123266759954204, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.107492923736572, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.884649932384491, "num_tokens": 633640975.0, "step": 16691 }, { "epoch": 2.1233939702327946, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.300893545150757, "learning_rate": 1e-05, "loss": 0.4102, "mean_token_accuracy": 0.8856422901153564, "num_tokens": 633684340.0, "step": 16692 }, { "epoch": 2.123521180511385, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.263364315032959, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.883824348449707, "num_tokens": 633720969.0, "step": 16693 }, { "epoch": 2.1236483907899757, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.540196657180786, "learning_rate": 1e-05, "loss": 0.3795, "mean_token_accuracy": 0.8934025764465332, "num_tokens": 633751726.0, "step": 16694 }, { "epoch": 2.123775601068566, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.19631290435791, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8864591717720032, "num_tokens": 633784645.0, "step": 16695 }, { "epoch": 2.1239028113471567, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6945641040802, "learning_rate": 1e-05, "loss": 0.3802, "mean_token_accuracy": 0.8944960832595825, "num_tokens": 633816028.0, "step": 16696 }, { "epoch": 2.1240300216257473, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.000401020050049, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.872934103012085, "num_tokens": 633859018.0, "step": 16697 }, { "epoch": 2.124157231904338, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6950817108154297, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8748304843902588, "num_tokens": 633897549.0, "step": 16698 }, { "epoch": 2.1242844421829283, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8143153190612793, "learning_rate": 1e-05, "loss": 0.3907, "mean_token_accuracy": 0.8914898633956909, "num_tokens": 633933111.0, "step": 16699 }, { "epoch": 2.124411652461519, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6354615688323975, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8696821331977844, "num_tokens": 633972409.0, "step": 16700 }, { "epoch": 2.1245388627401094, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6791088581085205, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8634465336799622, "num_tokens": 634007112.0, "step": 16701 }, { "epoch": 2.1246660730187, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.563533067703247, "learning_rate": 1e-05, "loss": 0.3779, "mean_token_accuracy": 0.8957210779190063, "num_tokens": 634043509.0, "step": 16702 }, { "epoch": 2.1247932832972904, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5840587615966797, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.881956934928894, "num_tokens": 634081883.0, "step": 16703 }, { "epoch": 2.124920493575881, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.570326805114746, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8913788795471191, "num_tokens": 634116203.0, "step": 16704 }, { "epoch": 2.1250477038544715, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.502803325653076, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8835904598236084, "num_tokens": 634152538.0, "step": 16705 }, { "epoch": 2.125174914133062, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4838359355926514, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8853436708450317, "num_tokens": 634188563.0, "step": 16706 }, { "epoch": 2.1253021244116526, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4631199836730957, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8677141070365906, "num_tokens": 634235164.0, "step": 16707 }, { "epoch": 2.125429334690243, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.44655179977417, "learning_rate": 1e-05, "loss": 0.4011, "mean_token_accuracy": 0.888032078742981, "num_tokens": 634275136.0, "step": 16708 }, { "epoch": 2.1255565449688336, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.459169626235962, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8869026899337769, "num_tokens": 634312056.0, "step": 16709 }, { "epoch": 2.125683755247424, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4375953674316406, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8647186160087585, "num_tokens": 634345737.0, "step": 16710 }, { "epoch": 2.1258109655260147, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4546618461608887, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.880313515663147, "num_tokens": 634381122.0, "step": 16711 }, { "epoch": 2.125938175804605, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3958542346954346, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8801023960113525, "num_tokens": 634420529.0, "step": 16712 }, { "epoch": 2.1260653860831957, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4539034366607666, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8853188753128052, "num_tokens": 634460540.0, "step": 16713 }, { "epoch": 2.1261925963617863, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.379000186920166, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8702828884124756, "num_tokens": 634497818.0, "step": 16714 }, { "epoch": 2.1263198066403763, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4932401180267334, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8733410239219666, "num_tokens": 634532262.0, "step": 16715 }, { "epoch": 2.126447016918967, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3353493213653564, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8795268535614014, "num_tokens": 634567316.0, "step": 16716 }, { "epoch": 2.1265742271975574, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.581433057785034, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8777909278869629, "num_tokens": 634608603.0, "step": 16717 }, { "epoch": 2.126701437476148, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.254286766052246, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8795197606086731, "num_tokens": 634646539.0, "step": 16718 }, { "epoch": 2.1268286477547385, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.681718111038208, "learning_rate": 1e-05, "loss": 0.3966, "mean_token_accuracy": 0.8876733183860779, "num_tokens": 634673998.0, "step": 16719 }, { "epoch": 2.126955858033329, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.0002918243408203125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.167651653289795, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8818210363388062, "num_tokens": 634707609.0, "step": 16720 }, { "epoch": 2.1270830683119195, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8708691596984863, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8845524787902832, "num_tokens": 634742893.0, "step": 16721 }, { "epoch": 2.12721027859051, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.0956833362579346, "learning_rate": 1e-05, "loss": 0.4097, "mean_token_accuracy": 0.8857864141464233, "num_tokens": 634780887.0, "step": 16722 }, { "epoch": 2.1273374888691006, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.170981407165527, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8838133811950684, "num_tokens": 634819512.0, "step": 16723 }, { "epoch": 2.127464699147691, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.1854519844055176, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8664385080337524, "num_tokens": 634857213.0, "step": 16724 }, { "epoch": 2.1275919094262816, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.477869033813477, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8734023571014404, "num_tokens": 634899221.0, "step": 16725 }, { "epoch": 2.127719119704872, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.542189598083496, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.8956010341644287, "num_tokens": 634943486.0, "step": 16726 }, { "epoch": 2.1278463299834627, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.517820835113525, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8738422393798828, "num_tokens": 634981303.0, "step": 16727 }, { "epoch": 2.127973540262053, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 3.9069466590881348, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.877636194229126, "num_tokens": 635021411.0, "step": 16728 }, { "epoch": 2.1281007505406437, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.244760990142822, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8898690938949585, "num_tokens": 635063449.0, "step": 16729 }, { "epoch": 2.1282279608192343, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.9002976417541504, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8676197528839111, "num_tokens": 635107052.0, "step": 16730 }, { "epoch": 2.128355171097825, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.0306291580200195, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.8905766010284424, "num_tokens": 635141036.0, "step": 16731 }, { "epoch": 2.1284823813764153, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7665746212005615, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.871617317199707, "num_tokens": 635181060.0, "step": 16732 }, { "epoch": 2.128609591655006, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8767969608306885, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8797072768211365, "num_tokens": 635222679.0, "step": 16733 }, { "epoch": 2.1287368019335964, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.637988328933716, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8841348886489868, "num_tokens": 635258351.0, "step": 16734 }, { "epoch": 2.128864012212187, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.784708261489868, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.862438440322876, "num_tokens": 635300647.0, "step": 16735 }, { "epoch": 2.1289912224907774, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5208141803741455, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8810425996780396, "num_tokens": 635335239.0, "step": 16736 }, { "epoch": 2.129118432769368, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7707159519195557, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8853311538696289, "num_tokens": 635370786.0, "step": 16737 }, { "epoch": 2.129245643047958, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3976447582244873, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.881279468536377, "num_tokens": 635410033.0, "step": 16738 }, { "epoch": 2.129372853326549, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8243777751922607, "learning_rate": 1e-05, "loss": 0.3938, "mean_token_accuracy": 0.8922160863876343, "num_tokens": 635447157.0, "step": 16739 }, { "epoch": 2.129500063605139, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3134031295776367, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8876853585243225, "num_tokens": 635485580.0, "step": 16740 }, { "epoch": 2.1296272738837296, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9639768600463867, "learning_rate": 1e-05, "loss": 0.3778, "mean_token_accuracy": 0.8944734930992126, "num_tokens": 635520289.0, "step": 16741 }, { "epoch": 2.12975448416232, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.300252914428711, "learning_rate": 1e-05, "loss": 0.3946, "mean_token_accuracy": 0.8861578702926636, "num_tokens": 635552390.0, "step": 16742 }, { "epoch": 2.1298816944409107, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.143528461456299, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8787925243377686, "num_tokens": 635588281.0, "step": 16743 }, { "epoch": 2.130008904719501, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.416144371032715, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8682764172554016, "num_tokens": 635632229.0, "step": 16744 }, { "epoch": 2.1301361149980917, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.199153900146484, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8853594660758972, "num_tokens": 635671315.0, "step": 16745 }, { "epoch": 2.1302633252766823, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6201727390289307, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8763078451156616, "num_tokens": 635711765.0, "step": 16746 }, { "epoch": 2.130390535555273, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.071422576904297, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8639011383056641, "num_tokens": 635747710.0, "step": 16747 }, { "epoch": 2.1305177458338633, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7164225578308105, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8689476847648621, "num_tokens": 635781355.0, "step": 16748 }, { "epoch": 2.130644956112454, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.819434881210327, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8814283013343811, "num_tokens": 635821340.0, "step": 16749 }, { "epoch": 2.1307721663910444, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7029647827148438, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8711339235305786, "num_tokens": 635857943.0, "step": 16750 }, { "epoch": 2.130899376669635, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.601595163345337, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8749663829803467, "num_tokens": 635896808.0, "step": 16751 }, { "epoch": 2.1310265869482254, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6796631813049316, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8861500024795532, "num_tokens": 635934952.0, "step": 16752 }, { "epoch": 2.131153797226816, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4316813945770264, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8813443779945374, "num_tokens": 635972968.0, "step": 16753 }, { "epoch": 2.1312810075054065, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.688732862472534, "learning_rate": 1e-05, "loss": 0.3743, "mean_token_accuracy": 0.8928811550140381, "num_tokens": 636004761.0, "step": 16754 }, { "epoch": 2.131408217783997, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3264262676239014, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.874049186706543, "num_tokens": 636043003.0, "step": 16755 }, { "epoch": 2.1315354280625876, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7367289066314697, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8803266882896423, "num_tokens": 636083580.0, "step": 16756 }, { "epoch": 2.131662638341178, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.241297721862793, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8842754364013672, "num_tokens": 636124248.0, "step": 16757 }, { "epoch": 2.1317898486197686, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.862062692642212, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.872557520866394, "num_tokens": 636164665.0, "step": 16758 }, { "epoch": 2.131917058898359, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2133970260620117, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8778395056724548, "num_tokens": 636201526.0, "step": 16759 }, { "epoch": 2.1320442691769497, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.022706508636475, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8794326186180115, "num_tokens": 636247793.0, "step": 16760 }, { "epoch": 2.13217147945554, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.247835636138916, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8881244659423828, "num_tokens": 636289989.0, "step": 16761 }, { "epoch": 2.1322986897341307, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.197264194488525, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.867512583732605, "num_tokens": 636328729.0, "step": 16762 }, { "epoch": 2.132425900012721, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.4426984786987305, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.879650354385376, "num_tokens": 636362593.0, "step": 16763 }, { "epoch": 2.1325531102913113, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.210997581481934, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8888545036315918, "num_tokens": 636402790.0, "step": 16764 }, { "epoch": 2.132680320569902, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.606996774673462, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8777443170547485, "num_tokens": 636443052.0, "step": 16765 }, { "epoch": 2.1328075308484924, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.082655906677246, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8809368014335632, "num_tokens": 636481345.0, "step": 16766 }, { "epoch": 2.132934741127083, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.674984931945801, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8805643320083618, "num_tokens": 636516638.0, "step": 16767 }, { "epoch": 2.1330619514056735, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.8957574367523193, "learning_rate": 1e-05, "loss": 0.3818, "mean_token_accuracy": 0.8940829634666443, "num_tokens": 636551606.0, "step": 16768 }, { "epoch": 2.133189161684264, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.643913745880127, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8743572235107422, "num_tokens": 636589298.0, "step": 16769 }, { "epoch": 2.1333163719628545, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7396011352539062, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8804844617843628, "num_tokens": 636622010.0, "step": 16770 }, { "epoch": 2.133443582241445, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5787105560302734, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8835264444351196, "num_tokens": 636665355.0, "step": 16771 }, { "epoch": 2.1335707925200356, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6217195987701416, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8869133591651917, "num_tokens": 636706135.0, "step": 16772 }, { "epoch": 2.133698002798626, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5062758922576904, "learning_rate": 1e-05, "loss": 0.3781, "mean_token_accuracy": 0.8937591314315796, "num_tokens": 636744587.0, "step": 16773 }, { "epoch": 2.1338252130772166, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5726611614227295, "learning_rate": 1e-05, "loss": 0.3744, "mean_token_accuracy": 0.8920260667800903, "num_tokens": 636784175.0, "step": 16774 }, { "epoch": 2.133952423355807, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.46075439453125, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8702863454818726, "num_tokens": 636821752.0, "step": 16775 }, { "epoch": 2.1340796336343977, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.575960874557495, "learning_rate": 1e-05, "loss": 0.4857, "mean_token_accuracy": 0.8660721182823181, "num_tokens": 636855507.0, "step": 16776 }, { "epoch": 2.134206843912988, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.391968011856079, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8751722574234009, "num_tokens": 636895401.0, "step": 16777 }, { "epoch": 2.1343340541915787, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.574723720550537, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8621562123298645, "num_tokens": 636940225.0, "step": 16778 }, { "epoch": 2.1344612644701693, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3331546783447266, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8752380013465881, "num_tokens": 636976685.0, "step": 16779 }, { "epoch": 2.13458847474876, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6247780323028564, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8804603815078735, "num_tokens": 637018170.0, "step": 16780 }, { "epoch": 2.1347156850273503, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.262863874435425, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8818433284759521, "num_tokens": 637055115.0, "step": 16781 }, { "epoch": 2.134842895305941, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7490646839141846, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8686240911483765, "num_tokens": 637093964.0, "step": 16782 }, { "epoch": 2.1349701055845314, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1838183403015137, "learning_rate": 1e-05, "loss": 0.3596, "mean_token_accuracy": 0.9001526236534119, "num_tokens": 637135242.0, "step": 16783 }, { "epoch": 2.135097315863122, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9322125911712646, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8794611692428589, "num_tokens": 637173407.0, "step": 16784 }, { "epoch": 2.1352245261417124, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1895675659179688, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8791199922561646, "num_tokens": 637207613.0, "step": 16785 }, { "epoch": 2.135351736420303, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.153749465942383, "learning_rate": 1e-05, "loss": 0.3876, "mean_token_accuracy": 0.8939749002456665, "num_tokens": 637242010.0, "step": 16786 }, { "epoch": 2.1354789466988935, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.300179958343506, "learning_rate": 1e-05, "loss": 0.3997, "mean_token_accuracy": 0.8854024410247803, "num_tokens": 637284034.0, "step": 16787 }, { "epoch": 2.1356061569774836, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.300015926361084, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8879230618476868, "num_tokens": 637325657.0, "step": 16788 }, { "epoch": 2.135733367256074, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.572901964187622, "learning_rate": 1e-05, "loss": 0.3975, "mean_token_accuracy": 0.8904566764831543, "num_tokens": 637360946.0, "step": 16789 }, { "epoch": 2.1358605775346646, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.231656551361084, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8769596815109253, "num_tokens": 637397540.0, "step": 16790 }, { "epoch": 2.135987787813255, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7237372398376465, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8859113454818726, "num_tokens": 637443148.0, "step": 16791 }, { "epoch": 2.1361149980918457, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0143818855285645, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8889142274856567, "num_tokens": 637480796.0, "step": 16792 }, { "epoch": 2.136242208370436, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.720224142074585, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8755593299865723, "num_tokens": 637525465.0, "step": 16793 }, { "epoch": 2.1363694186490267, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8154749870300293, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8827481269836426, "num_tokens": 637563235.0, "step": 16794 }, { "epoch": 2.1364966289276173, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.649866819381714, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.8882249593734741, "num_tokens": 637600061.0, "step": 16795 }, { "epoch": 2.136623839206208, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6581759452819824, "learning_rate": 1e-05, "loss": 0.3809, "mean_token_accuracy": 0.89522784948349, "num_tokens": 637640877.0, "step": 16796 }, { "epoch": 2.1367510494847983, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5792338848114014, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8765299320220947, "num_tokens": 637679248.0, "step": 16797 }, { "epoch": 2.136878259763389, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.571589231491089, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.876437246799469, "num_tokens": 637715973.0, "step": 16798 }, { "epoch": 2.1370054700419794, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5161874294281006, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8864116072654724, "num_tokens": 637753358.0, "step": 16799 }, { "epoch": 2.13713268032057, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5077829360961914, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8808032870292664, "num_tokens": 637798608.0, "step": 16800 }, { "epoch": 2.1372598905991604, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4569413661956787, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8824948668479919, "num_tokens": 637837173.0, "step": 16801 }, { "epoch": 2.137387100877751, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4965364933013916, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8879966735839844, "num_tokens": 637874072.0, "step": 16802 }, { "epoch": 2.1375143111563415, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4293148517608643, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.871911883354187, "num_tokens": 637909369.0, "step": 16803 }, { "epoch": 2.137641521434932, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4847660064697266, "learning_rate": 1e-05, "loss": 0.39, "mean_token_accuracy": 0.8915039896965027, "num_tokens": 637946526.0, "step": 16804 }, { "epoch": 2.1377687317135226, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3831613063812256, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8755879402160645, "num_tokens": 637984476.0, "step": 16805 }, { "epoch": 2.137895941992113, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5119779109954834, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8758504986763, "num_tokens": 638019098.0, "step": 16806 }, { "epoch": 2.1380231522707036, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3300399780273438, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8810209035873413, "num_tokens": 638059049.0, "step": 16807 }, { "epoch": 2.138150362549294, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.558079719543457, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8812909126281738, "num_tokens": 638095887.0, "step": 16808 }, { "epoch": 2.1382775728278847, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.276930332183838, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8623863458633423, "num_tokens": 638141855.0, "step": 16809 }, { "epoch": 2.138404783106475, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.637951135635376, "learning_rate": 1e-05, "loss": 0.3886, "mean_token_accuracy": 0.8927052617073059, "num_tokens": 638184491.0, "step": 16810 }, { "epoch": 2.1385319933850653, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2110188007354736, "learning_rate": 1e-05, "loss": 0.3752, "mean_token_accuracy": 0.8909468650817871, "num_tokens": 638218341.0, "step": 16811 }, { "epoch": 2.1386592036636562, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7899017333984375, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8749794960021973, "num_tokens": 638257422.0, "step": 16812 }, { "epoch": 2.1387864139422463, "ewc_loss": 0.0751953125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1311392784118652, "learning_rate": 1e-05, "loss": 0.3997, "mean_token_accuracy": 0.8848739862442017, "num_tokens": 638298053.0, "step": 16813 }, { "epoch": 2.138913624220837, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.028475284576416, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8628876209259033, "num_tokens": 638337870.0, "step": 16814 }, { "epoch": 2.1390408344994274, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.1526284217834473, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8807329535484314, "num_tokens": 638371587.0, "step": 16815 }, { "epoch": 2.139168044778018, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.300282955169678, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8797037601470947, "num_tokens": 638416879.0, "step": 16816 }, { "epoch": 2.1392952550566084, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3627383708953857, "learning_rate": 1e-05, "loss": 0.3907, "mean_token_accuracy": 0.8905712366104126, "num_tokens": 638452809.0, "step": 16817 }, { "epoch": 2.139422465335199, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.45792818069458, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8743678331375122, "num_tokens": 638487027.0, "step": 16818 }, { "epoch": 2.1395496756137895, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.7289462089538574, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8851323127746582, "num_tokens": 638517922.0, "step": 16819 }, { "epoch": 2.13967688589238, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.279271602630615, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8704843521118164, "num_tokens": 638562360.0, "step": 16820 }, { "epoch": 2.1398040961709706, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.8413381576538086, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8756102919578552, "num_tokens": 638595649.0, "step": 16821 }, { "epoch": 2.139931306449561, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.061694622039795, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8824920654296875, "num_tokens": 638632588.0, "step": 16822 }, { "epoch": 2.1400585167281516, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7813775539398193, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8833968639373779, "num_tokens": 638663000.0, "step": 16823 }, { "epoch": 2.140185727006742, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.862496852874756, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8778724670410156, "num_tokens": 638703737.0, "step": 16824 }, { "epoch": 2.1403129372853327, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6586732864379883, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.868108868598938, "num_tokens": 638745526.0, "step": 16825 }, { "epoch": 2.140440147563923, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7494773864746094, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8855212926864624, "num_tokens": 638780865.0, "step": 16826 }, { "epoch": 2.1405673578425137, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5547373294830322, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8782601952552795, "num_tokens": 638822274.0, "step": 16827 }, { "epoch": 2.1406945681211043, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6981003284454346, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8693432807922363, "num_tokens": 638859077.0, "step": 16828 }, { "epoch": 2.140821778399695, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4466471672058105, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8803465962409973, "num_tokens": 638898055.0, "step": 16829 }, { "epoch": 2.1409489886782853, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6992294788360596, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8671424388885498, "num_tokens": 638939597.0, "step": 16830 }, { "epoch": 2.141076198956876, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3728411197662354, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8757047057151794, "num_tokens": 638978268.0, "step": 16831 }, { "epoch": 2.1412034092354664, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7596688270568848, "learning_rate": 1e-05, "loss": 0.3847, "mean_token_accuracy": 0.893548846244812, "num_tokens": 639010554.0, "step": 16832 }, { "epoch": 2.141330619514057, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2756316661834717, "learning_rate": 1e-05, "loss": 0.351, "mean_token_accuracy": 0.9028993844985962, "num_tokens": 639046721.0, "step": 16833 }, { "epoch": 2.1414578297926474, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8978090286254883, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8747597336769104, "num_tokens": 639080078.0, "step": 16834 }, { "epoch": 2.141585040071238, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.247540235519409, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8837940692901611, "num_tokens": 639119469.0, "step": 16835 }, { "epoch": 2.141712250349828, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.067549705505371, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8773801326751709, "num_tokens": 639159954.0, "step": 16836 }, { "epoch": 2.141839460628419, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3246166706085205, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8842214941978455, "num_tokens": 639202455.0, "step": 16837 }, { "epoch": 2.141966670907009, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.193241119384766, "learning_rate": 1e-05, "loss": 0.4001, "mean_token_accuracy": 0.8904470801353455, "num_tokens": 639240889.0, "step": 16838 }, { "epoch": 2.1420938811855996, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000293731689453125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5239484310150146, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8812638521194458, "num_tokens": 639275079.0, "step": 16839 }, { "epoch": 2.14222109146419, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.136982440948486, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8723641633987427, "num_tokens": 639308166.0, "step": 16840 }, { "epoch": 2.1423483017427807, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6656908988952637, "learning_rate": 1e-05, "loss": 0.4157, "mean_token_accuracy": 0.8908318877220154, "num_tokens": 639341862.0, "step": 16841 }, { "epoch": 2.142475512021371, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.918816328048706, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8905420303344727, "num_tokens": 639381065.0, "step": 16842 }, { "epoch": 2.1426027222999617, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6706933975219727, "learning_rate": 1e-05, "loss": 0.4009, "mean_token_accuracy": 0.8887343406677246, "num_tokens": 639421232.0, "step": 16843 }, { "epoch": 2.1427299325785523, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7156009674072266, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8751844167709351, "num_tokens": 639459462.0, "step": 16844 }, { "epoch": 2.142857142857143, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.649807929992676, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.880926251411438, "num_tokens": 639499759.0, "step": 16845 }, { "epoch": 2.1429843531357333, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5519161224365234, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8827021718025208, "num_tokens": 639534285.0, "step": 16846 }, { "epoch": 2.143111563414324, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6237728595733643, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8766831159591675, "num_tokens": 639576058.0, "step": 16847 }, { "epoch": 2.1432387736929144, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4381206035614014, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8842151761054993, "num_tokens": 639610464.0, "step": 16848 }, { "epoch": 2.143365983971505, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.613454818725586, "learning_rate": 1e-05, "loss": 0.3805, "mean_token_accuracy": 0.8936551213264465, "num_tokens": 639645327.0, "step": 16849 }, { "epoch": 2.1434931942500954, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.360947847366333, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8707824945449829, "num_tokens": 639684501.0, "step": 16850 }, { "epoch": 2.143620404528686, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.640597343444824, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8746541738510132, "num_tokens": 639721687.0, "step": 16851 }, { "epoch": 2.1437476148072765, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.294621229171753, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8630236387252808, "num_tokens": 639761503.0, "step": 16852 }, { "epoch": 2.143874825085867, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6977291107177734, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8839738368988037, "num_tokens": 639800938.0, "step": 16853 }, { "epoch": 2.1440020353644575, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.231498956680298, "learning_rate": 1e-05, "loss": 0.3669, "mean_token_accuracy": 0.8947628736495972, "num_tokens": 639835318.0, "step": 16854 }, { "epoch": 2.144129245643048, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.806899070739746, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8860108852386475, "num_tokens": 639878155.0, "step": 16855 }, { "epoch": 2.1442564559216386, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1774048805236816, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8856136798858643, "num_tokens": 639920369.0, "step": 16856 }, { "epoch": 2.144383666200229, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9886438846588135, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8835695385932922, "num_tokens": 639964690.0, "step": 16857 }, { "epoch": 2.1445108764788197, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2096118927001953, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8689801692962646, "num_tokens": 640002969.0, "step": 16858 }, { "epoch": 2.14463808675741, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.18467378616333, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8920458555221558, "num_tokens": 640044061.0, "step": 16859 }, { "epoch": 2.1447652970360007, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.363135576248169, "learning_rate": 1e-05, "loss": 0.4071, "mean_token_accuracy": 0.8857020139694214, "num_tokens": 640082803.0, "step": 16860 }, { "epoch": 2.144892507314591, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.272852897644043, "learning_rate": 1e-05, "loss": 0.3903, "mean_token_accuracy": 0.8987360000610352, "num_tokens": 640123223.0, "step": 16861 }, { "epoch": 2.1450197175931813, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.586322069168091, "learning_rate": 1e-05, "loss": 0.3806, "mean_token_accuracy": 0.8972583413124084, "num_tokens": 640160674.0, "step": 16862 }, { "epoch": 2.145146927871772, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.163334846496582, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.888580858707428, "num_tokens": 640197347.0, "step": 16863 }, { "epoch": 2.1452741381503624, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7143328189849854, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8811765909194946, "num_tokens": 640235599.0, "step": 16864 }, { "epoch": 2.145401348428953, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9548590183258057, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.876653790473938, "num_tokens": 640277753.0, "step": 16865 }, { "epoch": 2.1455285587075434, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6960461139678955, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8857415914535522, "num_tokens": 640317486.0, "step": 16866 }, { "epoch": 2.145655768986134, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.774055242538452, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8647456765174866, "num_tokens": 640357240.0, "step": 16867 }, { "epoch": 2.1457829792647245, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6218814849853516, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8849475383758545, "num_tokens": 640392429.0, "step": 16868 }, { "epoch": 2.145910189543315, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.647963523864746, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8700231313705444, "num_tokens": 640432529.0, "step": 16869 }, { "epoch": 2.1460373998219056, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.553396224975586, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8826805353164673, "num_tokens": 640468443.0, "step": 16870 }, { "epoch": 2.146164610100496, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5788238048553467, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8709865212440491, "num_tokens": 640505647.0, "step": 16871 }, { "epoch": 2.1462918203790866, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4775357246398926, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8684781193733215, "num_tokens": 640551180.0, "step": 16872 }, { "epoch": 2.146419030657677, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5412418842315674, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8867513537406921, "num_tokens": 640590939.0, "step": 16873 }, { "epoch": 2.1465462409362677, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4291160106658936, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8663029670715332, "num_tokens": 640625688.0, "step": 16874 }, { "epoch": 2.146673451214858, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5544300079345703, "learning_rate": 1e-05, "loss": 0.4962, "mean_token_accuracy": 0.8600748777389526, "num_tokens": 640669142.0, "step": 16875 }, { "epoch": 2.1468006614934487, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.351970911026001, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8717725276947021, "num_tokens": 640713345.0, "step": 16876 }, { "epoch": 2.1469278717720393, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6032748222351074, "learning_rate": 1e-05, "loss": 0.4828, "mean_token_accuracy": 0.8591840267181396, "num_tokens": 640750930.0, "step": 16877 }, { "epoch": 2.14705508205063, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.277848720550537, "learning_rate": 1e-05, "loss": 0.3966, "mean_token_accuracy": 0.8869507312774658, "num_tokens": 640792029.0, "step": 16878 }, { "epoch": 2.1471822923292203, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6901581287384033, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8859211206436157, "num_tokens": 640833952.0, "step": 16879 }, { "epoch": 2.147309502607811, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2140755653381348, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8796364068984985, "num_tokens": 640867980.0, "step": 16880 }, { "epoch": 2.1474367128864014, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8518407344818115, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8671355247497559, "num_tokens": 640913241.0, "step": 16881 }, { "epoch": 2.147563923164992, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1658847332000732, "learning_rate": 1e-05, "loss": 0.3713, "mean_token_accuracy": 0.896776020526886, "num_tokens": 640946705.0, "step": 16882 }, { "epoch": 2.1476911334435824, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.065914154052734, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8888751268386841, "num_tokens": 640985658.0, "step": 16883 }, { "epoch": 2.147818343722173, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.224534511566162, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8738008737564087, "num_tokens": 641025963.0, "step": 16884 }, { "epoch": 2.1479455540007635, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.270501136779785, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8875048160552979, "num_tokens": 641070853.0, "step": 16885 }, { "epoch": 2.1480727642793536, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.456653118133545, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.8846656084060669, "num_tokens": 641106336.0, "step": 16886 }, { "epoch": 2.148199974557944, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.293890476226807, "learning_rate": 1e-05, "loss": 0.3886, "mean_token_accuracy": 0.8972241282463074, "num_tokens": 641147842.0, "step": 16887 }, { "epoch": 2.1483271848365346, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.679476261138916, "learning_rate": 1e-05, "loss": 0.3957, "mean_token_accuracy": 0.8907111883163452, "num_tokens": 641186329.0, "step": 16888 }, { "epoch": 2.148454395115125, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.131939888000488, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8735581636428833, "num_tokens": 641227892.0, "step": 16889 }, { "epoch": 2.1485816053937157, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7446229457855225, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8864539861679077, "num_tokens": 641265611.0, "step": 16890 }, { "epoch": 2.148708815672306, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9318041801452637, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8620242476463318, "num_tokens": 641299564.0, "step": 16891 }, { "epoch": 2.1488360259508967, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6841561794281006, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8823633193969727, "num_tokens": 641335892.0, "step": 16892 }, { "epoch": 2.1489632362294873, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7600715160369873, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8799992799758911, "num_tokens": 641375470.0, "step": 16893 }, { "epoch": 2.149090446508078, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6171650886535645, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.8808599710464478, "num_tokens": 641408793.0, "step": 16894 }, { "epoch": 2.1492176567866683, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6478066444396973, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8782474398612976, "num_tokens": 641444949.0, "step": 16895 }, { "epoch": 2.149344867065259, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.537611484527588, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8793402910232544, "num_tokens": 641482775.0, "step": 16896 }, { "epoch": 2.1494720773438494, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.577939987182617, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.8857095241546631, "num_tokens": 641520385.0, "step": 16897 }, { "epoch": 2.14959928762244, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.456827163696289, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8689250349998474, "num_tokens": 641563885.0, "step": 16898 }, { "epoch": 2.1497264979010304, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.566002130508423, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8808770775794983, "num_tokens": 641600868.0, "step": 16899 }, { "epoch": 2.149853708179621, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.381943464279175, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8792498111724854, "num_tokens": 641643936.0, "step": 16900 }, { "epoch": 2.1499809184582115, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6103439331054688, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8735479116439819, "num_tokens": 641680248.0, "step": 16901 }, { "epoch": 2.150108128736802, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.31095814704895, "learning_rate": 1e-05, "loss": 0.3964, "mean_token_accuracy": 0.8879864811897278, "num_tokens": 641721280.0, "step": 16902 }, { "epoch": 2.1502353390153925, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.68392276763916, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8778015375137329, "num_tokens": 641760168.0, "step": 16903 }, { "epoch": 2.150362549293983, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.241034984588623, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8666324615478516, "num_tokens": 641798517.0, "step": 16904 }, { "epoch": 2.1504897595725736, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8068835735321045, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8748724460601807, "num_tokens": 641841486.0, "step": 16905 }, { "epoch": 2.150616969851164, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.203059196472168, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8720498085021973, "num_tokens": 641880042.0, "step": 16906 }, { "epoch": 2.1507441801297547, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9948606491088867, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8831067085266113, "num_tokens": 641916069.0, "step": 16907 }, { "epoch": 2.150871390408345, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2283003330230713, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8778172731399536, "num_tokens": 641951298.0, "step": 16908 }, { "epoch": 2.1509986006869353, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.179060459136963, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.883642315864563, "num_tokens": 641987577.0, "step": 16909 }, { "epoch": 2.1511258109655262, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3768489360809326, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8649722337722778, "num_tokens": 642031914.0, "step": 16910 }, { "epoch": 2.1512530212441163, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.252432346343994, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8759433627128601, "num_tokens": 642069729.0, "step": 16911 }, { "epoch": 2.151380231522707, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6065354347229004, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8782503008842468, "num_tokens": 642106831.0, "step": 16912 }, { "epoch": 2.1515074418012974, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.125891208648682, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8884974718093872, "num_tokens": 642142408.0, "step": 16913 }, { "epoch": 2.151634652079888, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6894710063934326, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8754381537437439, "num_tokens": 642181860.0, "step": 16914 }, { "epoch": 2.1517618623584784, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.92484188079834, "learning_rate": 1e-05, "loss": 0.3931, "mean_token_accuracy": 0.8950331211090088, "num_tokens": 642219073.0, "step": 16915 }, { "epoch": 2.151889072637069, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6747498512268066, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8775036931037903, "num_tokens": 642254872.0, "step": 16916 }, { "epoch": 2.1520162829156595, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7792744636535645, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8910514116287231, "num_tokens": 642295379.0, "step": 16917 }, { "epoch": 2.15214349319425, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.612163543701172, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8640585541725159, "num_tokens": 642333316.0, "step": 16918 }, { "epoch": 2.1522707034728406, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.623502016067505, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.871868908405304, "num_tokens": 642373677.0, "step": 16919 }, { "epoch": 2.152397913751431, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5383102893829346, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8701373934745789, "num_tokens": 642416066.0, "step": 16920 }, { "epoch": 2.1525251240300216, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.558342218399048, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8676308393478394, "num_tokens": 642456818.0, "step": 16921 }, { "epoch": 2.152652334308612, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.469373941421509, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8775776624679565, "num_tokens": 642498021.0, "step": 16922 }, { "epoch": 2.1527795445872027, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.539761781692505, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8715717792510986, "num_tokens": 642533999.0, "step": 16923 }, { "epoch": 2.152906754865793, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4159634113311768, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8801175355911255, "num_tokens": 642571943.0, "step": 16924 }, { "epoch": 2.1530339651443837, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5344388484954834, "learning_rate": 1e-05, "loss": 0.3964, "mean_token_accuracy": 0.8851375579833984, "num_tokens": 642605145.0, "step": 16925 }, { "epoch": 2.1531611754229742, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.364260673522949, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.877199113368988, "num_tokens": 642644686.0, "step": 16926 }, { "epoch": 2.1532883857015648, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5778048038482666, "learning_rate": 1e-05, "loss": 0.3965, "mean_token_accuracy": 0.8868489861488342, "num_tokens": 642677815.0, "step": 16927 }, { "epoch": 2.1534155959801553, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2920725345611572, "learning_rate": 1e-05, "loss": 0.3819, "mean_token_accuracy": 0.8950603008270264, "num_tokens": 642716245.0, "step": 16928 }, { "epoch": 2.153542806258746, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6723625659942627, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8779129981994629, "num_tokens": 642750428.0, "step": 16929 }, { "epoch": 2.1536700165373364, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2290687561035156, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8735759258270264, "num_tokens": 642790349.0, "step": 16930 }, { "epoch": 2.153797226815927, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7883269786834717, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8849532604217529, "num_tokens": 642835063.0, "step": 16931 }, { "epoch": 2.1539244370945174, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1729934215545654, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8772878646850586, "num_tokens": 642875013.0, "step": 16932 }, { "epoch": 2.154051647373108, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.998513698577881, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.870866060256958, "num_tokens": 642914934.0, "step": 16933 }, { "epoch": 2.154178857651698, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1944096088409424, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8805098533630371, "num_tokens": 642949368.0, "step": 16934 }, { "epoch": 2.1543060679302886, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.229065895080566, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8709642887115479, "num_tokens": 642988083.0, "step": 16935 }, { "epoch": 2.154433278208879, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3710384368896484, "learning_rate": 1e-05, "loss": 0.4841, "mean_token_accuracy": 0.8615046739578247, "num_tokens": 643028868.0, "step": 16936 }, { "epoch": 2.1545604884874696, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.323394775390625, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8770726919174194, "num_tokens": 643072251.0, "step": 16937 }, { "epoch": 2.15468769876606, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6243107318878174, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8823681473731995, "num_tokens": 643109864.0, "step": 16938 }, { "epoch": 2.1548149090446507, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.200788974761963, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.892872154712677, "num_tokens": 643147868.0, "step": 16939 }, { "epoch": 2.154942119323241, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.73927640914917, "learning_rate": 1e-05, "loss": 0.3579, "mean_token_accuracy": 0.9033196568489075, "num_tokens": 643182173.0, "step": 16940 }, { "epoch": 2.1550693296018317, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.008541107177734, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8742143511772156, "num_tokens": 643217006.0, "step": 16941 }, { "epoch": 2.1551965398804223, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.721698760986328, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8628171682357788, "num_tokens": 643253724.0, "step": 16942 }, { "epoch": 2.155323750159013, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.812938928604126, "learning_rate": 1e-05, "loss": 0.4749, "mean_token_accuracy": 0.8659923076629639, "num_tokens": 643294633.0, "step": 16943 }, { "epoch": 2.1554509604376033, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.638113260269165, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.872340202331543, "num_tokens": 643333269.0, "step": 16944 }, { "epoch": 2.155578170716194, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.668262481689453, "learning_rate": 1e-05, "loss": 0.4071, "mean_token_accuracy": 0.887468695640564, "num_tokens": 643374892.0, "step": 16945 }, { "epoch": 2.1557053809947844, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.549327850341797, "learning_rate": 1e-05, "loss": 0.3912, "mean_token_accuracy": 0.8910741209983826, "num_tokens": 643411774.0, "step": 16946 }, { "epoch": 2.155832591273375, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6023380756378174, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.8875693082809448, "num_tokens": 643450034.0, "step": 16947 }, { "epoch": 2.1559598015519654, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.483912944793701, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8809992074966431, "num_tokens": 643482901.0, "step": 16948 }, { "epoch": 2.156087011830556, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.570626735687256, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8754100203514099, "num_tokens": 643524918.0, "step": 16949 }, { "epoch": 2.1562142221091465, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4105114936828613, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8702032566070557, "num_tokens": 643563931.0, "step": 16950 }, { "epoch": 2.156341432387737, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.574892044067383, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8792926669120789, "num_tokens": 643602838.0, "step": 16951 }, { "epoch": 2.1564686426663275, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3520607948303223, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8803912997245789, "num_tokens": 643641128.0, "step": 16952 }, { "epoch": 2.156595852944918, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.623262643814087, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8836089372634888, "num_tokens": 643674853.0, "step": 16953 }, { "epoch": 2.1567230632235086, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.284653902053833, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8759958744049072, "num_tokens": 643709924.0, "step": 16954 }, { "epoch": 2.156850273502099, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7156567573547363, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8851583003997803, "num_tokens": 643744899.0, "step": 16955 }, { "epoch": 2.1569774837806897, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.205332040786743, "learning_rate": 1e-05, "loss": 0.3845, "mean_token_accuracy": 0.8916136026382446, "num_tokens": 643781316.0, "step": 16956 }, { "epoch": 2.15710469405928, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.879666805267334, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8824542164802551, "num_tokens": 643824267.0, "step": 16957 }, { "epoch": 2.1572319043378707, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.177225351333618, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8777820467948914, "num_tokens": 643867155.0, "step": 16958 }, { "epoch": 2.157359114616461, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.099257469177246, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.888286828994751, "num_tokens": 643907813.0, "step": 16959 }, { "epoch": 2.1574863248950513, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.282148599624634, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8861951231956482, "num_tokens": 643940387.0, "step": 16960 }, { "epoch": 2.157613535173642, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.255315780639648, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8719980716705322, "num_tokens": 643980387.0, "step": 16961 }, { "epoch": 2.1577407454522324, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4884142875671387, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8812791109085083, "num_tokens": 644020883.0, "step": 16962 }, { "epoch": 2.157867955730823, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.259697437286377, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8631458282470703, "num_tokens": 644054492.0, "step": 16963 }, { "epoch": 2.1579951660094134, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6831281185150146, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8684478998184204, "num_tokens": 644095009.0, "step": 16964 }, { "epoch": 2.158122376288004, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.07565975189209, "learning_rate": 1e-05, "loss": 0.3967, "mean_token_accuracy": 0.8908646106719971, "num_tokens": 644128469.0, "step": 16965 }, { "epoch": 2.1582495865665945, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002956390380859375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7167177200317383, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8626518845558167, "num_tokens": 644167102.0, "step": 16966 }, { "epoch": 2.158376796845185, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.875159502029419, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8774780035018921, "num_tokens": 644204319.0, "step": 16967 }, { "epoch": 2.1585040071237755, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.658432722091675, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8706520795822144, "num_tokens": 644246950.0, "step": 16968 }, { "epoch": 2.158631217402366, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7074968814849854, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8818126916885376, "num_tokens": 644288744.0, "step": 16969 }, { "epoch": 2.1587584276809566, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.591256856918335, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8811548948287964, "num_tokens": 644326719.0, "step": 16970 }, { "epoch": 2.158885637959547, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6110143661499023, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.876734733581543, "num_tokens": 644362027.0, "step": 16971 }, { "epoch": 2.1590128482381377, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5218729972839355, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8786368370056152, "num_tokens": 644401917.0, "step": 16972 }, { "epoch": 2.159140058516728, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5655159950256348, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8733123540878296, "num_tokens": 644441735.0, "step": 16973 }, { "epoch": 2.1592672687953187, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4532580375671387, "learning_rate": 1e-05, "loss": 0.3723, "mean_token_accuracy": 0.893653929233551, "num_tokens": 644475833.0, "step": 16974 }, { "epoch": 2.1593944790739092, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5488390922546387, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8798106908798218, "num_tokens": 644516506.0, "step": 16975 }, { "epoch": 2.1595216893524998, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.392624616622925, "learning_rate": 1e-05, "loss": 0.3968, "mean_token_accuracy": 0.886177659034729, "num_tokens": 644554081.0, "step": 16976 }, { "epoch": 2.1596488996310903, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.56307053565979, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8682303428649902, "num_tokens": 644599514.0, "step": 16977 }, { "epoch": 2.159776109909681, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3302292823791504, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8704013824462891, "num_tokens": 644639569.0, "step": 16978 }, { "epoch": 2.1599033201882714, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6151885986328125, "learning_rate": 1e-05, "loss": 0.3881, "mean_token_accuracy": 0.894844651222229, "num_tokens": 644685721.0, "step": 16979 }, { "epoch": 2.160030530466862, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2644948959350586, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8651134371757507, "num_tokens": 644729939.0, "step": 16980 }, { "epoch": 2.1601577407454524, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.728740930557251, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8717910647392273, "num_tokens": 644768181.0, "step": 16981 }, { "epoch": 2.160284951024043, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.57763671875e-05, "grad_norm": 3.1947896480560303, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8721576929092407, "num_tokens": 644806288.0, "step": 16982 }, { "epoch": 2.1604121613026335, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8970892429351807, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8857975006103516, "num_tokens": 644846116.0, "step": 16983 }, { "epoch": 2.1605393715812236, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1680545806884766, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8737350702285767, "num_tokens": 644885330.0, "step": 16984 }, { "epoch": 2.160666581859814, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.13368034362793, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8695935606956482, "num_tokens": 644923341.0, "step": 16985 }, { "epoch": 2.1607937921384046, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2698802947998047, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8729889392852783, "num_tokens": 644964165.0, "step": 16986 }, { "epoch": 2.160921002416995, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.310248374938965, "learning_rate": 1e-05, "loss": 0.3905, "mean_token_accuracy": 0.8928857445716858, "num_tokens": 644999319.0, "step": 16987 }, { "epoch": 2.1610482126955857, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.522369146347046, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8889418840408325, "num_tokens": 645040687.0, "step": 16988 }, { "epoch": 2.161175422974176, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.294896125793457, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8759061098098755, "num_tokens": 645075772.0, "step": 16989 }, { "epoch": 2.1613026332527667, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7438454627990723, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8764604330062866, "num_tokens": 645111856.0, "step": 16990 }, { "epoch": 2.1614298435313573, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.076389312744141, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8737030625343323, "num_tokens": 645155995.0, "step": 16991 }, { "epoch": 2.161557053809948, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.741180419921875, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8643391132354736, "num_tokens": 645200101.0, "step": 16992 }, { "epoch": 2.1616842640885383, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.897627592086792, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8916300535202026, "num_tokens": 645235243.0, "step": 16993 }, { "epoch": 2.161811474367129, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.691222906112671, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8907322287559509, "num_tokens": 645270369.0, "step": 16994 }, { "epoch": 2.1619386846457194, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7228281497955322, "learning_rate": 1e-05, "loss": 0.3881, "mean_token_accuracy": 0.8925579190254211, "num_tokens": 645304111.0, "step": 16995 }, { "epoch": 2.16206589492431, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5966222286224365, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8681393265724182, "num_tokens": 645339884.0, "step": 16996 }, { "epoch": 2.1621931052029004, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6126368045806885, "learning_rate": 1e-05, "loss": 0.3664, "mean_token_accuracy": 0.8954987525939941, "num_tokens": 645372465.0, "step": 16997 }, { "epoch": 2.162320315481491, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.513225555419922, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8834306001663208, "num_tokens": 645413022.0, "step": 16998 }, { "epoch": 2.1624475257600815, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5686569213867188, "learning_rate": 1e-05, "loss": 0.3861, "mean_token_accuracy": 0.8940684795379639, "num_tokens": 645451344.0, "step": 16999 }, { "epoch": 2.162574736038672, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.432800531387329, "learning_rate": 1e-05, "loss": 0.3617, "mean_token_accuracy": 0.8983080983161926, "num_tokens": 645493463.0, "step": 17000 }, { "epoch": 2.1627019463172625, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5677573680877686, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8875139951705933, "num_tokens": 645533508.0, "step": 17001 }, { "epoch": 2.162829156595853, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3787944316864014, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8831606507301331, "num_tokens": 645573866.0, "step": 17002 }, { "epoch": 2.1629563668744436, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.621783971786499, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.886243999004364, "num_tokens": 645609835.0, "step": 17003 }, { "epoch": 2.163083577153034, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3055200576782227, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8853635191917419, "num_tokens": 645644228.0, "step": 17004 }, { "epoch": 2.1632107874316246, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6994917392730713, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8821508884429932, "num_tokens": 645682619.0, "step": 17005 }, { "epoch": 2.163337997710215, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.226644992828369, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8842519521713257, "num_tokens": 645719506.0, "step": 17006 }, { "epoch": 2.1634652079888053, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8466603755950928, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8754316568374634, "num_tokens": 645761930.0, "step": 17007 }, { "epoch": 2.1635924182673962, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1847987174987793, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8689875602722168, "num_tokens": 645800322.0, "step": 17008 }, { "epoch": 2.1637196285459863, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.048671722412109, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8796886205673218, "num_tokens": 645842914.0, "step": 17009 }, { "epoch": 2.163846838824577, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.227736473083496, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.873543918132782, "num_tokens": 645885798.0, "step": 17010 }, { "epoch": 2.1639740491031674, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.266739845275879, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8782640695571899, "num_tokens": 645922187.0, "step": 17011 }, { "epoch": 2.164101259381758, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.4496145248413086, "learning_rate": 1e-05, "loss": 0.3842, "mean_token_accuracy": 0.8937225341796875, "num_tokens": 645955594.0, "step": 17012 }, { "epoch": 2.1642284696603484, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.294480323791504, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.889519453048706, "num_tokens": 645992579.0, "step": 17013 }, { "epoch": 2.164355679938939, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.682135820388794, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8765529990196228, "num_tokens": 646033243.0, "step": 17014 }, { "epoch": 2.1644828902175295, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.127558708190918, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8897978067398071, "num_tokens": 646069309.0, "step": 17015 }, { "epoch": 2.16461010049612, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7502479553222656, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8805394768714905, "num_tokens": 646100720.0, "step": 17016 }, { "epoch": 2.1647373107747105, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9024343490600586, "learning_rate": 1e-05, "loss": 0.3675, "mean_token_accuracy": 0.8998315930366516, "num_tokens": 646135917.0, "step": 17017 }, { "epoch": 2.164864521053301, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6802704334259033, "learning_rate": 1e-05, "loss": 0.3819, "mean_token_accuracy": 0.893698513507843, "num_tokens": 646173227.0, "step": 17018 }, { "epoch": 2.1649917313318916, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7449915409088135, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8726270198822021, "num_tokens": 646210023.0, "step": 17019 }, { "epoch": 2.165118941610482, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6202335357666016, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.8837780952453613, "num_tokens": 646241124.0, "step": 17020 }, { "epoch": 2.1652461518890727, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.630247116088867, "learning_rate": 1e-05, "loss": 0.4815, "mean_token_accuracy": 0.8637493848800659, "num_tokens": 646276892.0, "step": 17021 }, { "epoch": 2.165373362167663, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.533461809158325, "learning_rate": 1e-05, "loss": 0.3816, "mean_token_accuracy": 0.8925960659980774, "num_tokens": 646308959.0, "step": 17022 }, { "epoch": 2.1655005724462537, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5582776069641113, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.8858821392059326, "num_tokens": 646342614.0, "step": 17023 }, { "epoch": 2.1656277827248442, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4691412448883057, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8789927959442139, "num_tokens": 646380395.0, "step": 17024 }, { "epoch": 2.1657549930034348, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.552445888519287, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8703418970108032, "num_tokens": 646417068.0, "step": 17025 }, { "epoch": 2.1658822032820253, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3968019485473633, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8780734539031982, "num_tokens": 646456655.0, "step": 17026 }, { "epoch": 2.166009413560616, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.566908359527588, "learning_rate": 1e-05, "loss": 0.365, "mean_token_accuracy": 0.9019348621368408, "num_tokens": 646495035.0, "step": 17027 }, { "epoch": 2.1661366238392064, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3276498317718506, "learning_rate": 1e-05, "loss": 0.3724, "mean_token_accuracy": 0.8920719027519226, "num_tokens": 646529695.0, "step": 17028 }, { "epoch": 2.166263834117797, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6343905925750732, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8619690537452698, "num_tokens": 646566563.0, "step": 17029 }, { "epoch": 2.1663910443963874, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2609331607818604, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8770256042480469, "num_tokens": 646609515.0, "step": 17030 }, { "epoch": 2.166518254674978, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.747765302658081, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8871444463729858, "num_tokens": 646650024.0, "step": 17031 }, { "epoch": 2.166645464953568, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 4.839658260345459, "learning_rate": 1e-05, "loss": 0.3922, "mean_token_accuracy": 0.8892195820808411, "num_tokens": 646690547.0, "step": 17032 }, { "epoch": 2.1667726752321586, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.458160400390625, "learning_rate": 1e-05, "loss": 0.3969, "mean_token_accuracy": 0.8875579237937927, "num_tokens": 646722060.0, "step": 17033 }, { "epoch": 2.166899885510749, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.815324544906616, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.876883864402771, "num_tokens": 646761972.0, "step": 17034 }, { "epoch": 2.1670270957893396, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.131362199783325, "learning_rate": 1e-05, "loss": 0.4643, "mean_token_accuracy": 0.8683820366859436, "num_tokens": 646794847.0, "step": 17035 }, { "epoch": 2.16715430606793, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.188544273376465, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.889215350151062, "num_tokens": 646833744.0, "step": 17036 }, { "epoch": 2.1672815163465207, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.1799280643463135, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8855435252189636, "num_tokens": 646874257.0, "step": 17037 }, { "epoch": 2.167408726625111, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00029754638671875, "ewc_loss_parallel": 7.486343383789062e-05, "grad_norm": 4.513187885284424, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8921821117401123, "num_tokens": 646912439.0, "step": 17038 }, { "epoch": 2.1675359369037017, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.577399730682373, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.8897138833999634, "num_tokens": 646946311.0, "step": 17039 }, { "epoch": 2.1676631471822922, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 7.581710815429688e-05, "grad_norm": 4.566677570343018, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.868108868598938, "num_tokens": 646982426.0, "step": 17040 }, { "epoch": 2.1677903574608828, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 3.946347236633301, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.8935953378677368, "num_tokens": 647023244.0, "step": 17041 }, { "epoch": 2.1679175677394733, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.279725551605225, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8626916408538818, "num_tokens": 647062434.0, "step": 17042 }, { "epoch": 2.168044778018064, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.9667181968688965, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.875755786895752, "num_tokens": 647098021.0, "step": 17043 }, { "epoch": 2.1681719882966544, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.032415390014648, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8755061626434326, "num_tokens": 647134709.0, "step": 17044 }, { "epoch": 2.168299198575245, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.821169137954712, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.866484522819519, "num_tokens": 647169267.0, "step": 17045 }, { "epoch": 2.1684264088538354, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.8607735633850098, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8775573372840881, "num_tokens": 647207399.0, "step": 17046 }, { "epoch": 2.168553619132426, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6783812046051025, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.876555323600769, "num_tokens": 647244722.0, "step": 17047 }, { "epoch": 2.1686808294110165, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.773897409439087, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8857541680335999, "num_tokens": 647279051.0, "step": 17048 }, { "epoch": 2.168808039689607, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.531641960144043, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8767684698104858, "num_tokens": 647319201.0, "step": 17049 }, { "epoch": 2.1689352499681975, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.763502597808838, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8752776384353638, "num_tokens": 647358306.0, "step": 17050 }, { "epoch": 2.169062460246788, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.425657272338867, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8693609237670898, "num_tokens": 647393879.0, "step": 17051 }, { "epoch": 2.1691896705253786, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.80672287940979, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8724873065948486, "num_tokens": 647434848.0, "step": 17052 }, { "epoch": 2.169316880803969, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3374929428100586, "learning_rate": 1e-05, "loss": 0.3974, "mean_token_accuracy": 0.8872102499008179, "num_tokens": 647469711.0, "step": 17053 }, { "epoch": 2.1694440910825596, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.930492639541626, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.8640444874763489, "num_tokens": 647507788.0, "step": 17054 }, { "epoch": 2.16957130136115, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2856578826904297, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8803386688232422, "num_tokens": 647542193.0, "step": 17055 }, { "epoch": 2.1696985116397407, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.0913496017456055, "learning_rate": 1e-05, "loss": 0.3822, "mean_token_accuracy": 0.8979045152664185, "num_tokens": 647578692.0, "step": 17056 }, { "epoch": 2.169825721918331, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.358020544052124, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8849239349365234, "num_tokens": 647618158.0, "step": 17057 }, { "epoch": 2.1699529321969213, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.230096340179443, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8825477361679077, "num_tokens": 647660788.0, "step": 17058 }, { "epoch": 2.170080142475512, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5629115104675293, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8713974356651306, "num_tokens": 647700418.0, "step": 17059 }, { "epoch": 2.1702073527541024, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.158495903015137, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8867397308349609, "num_tokens": 647738248.0, "step": 17060 }, { "epoch": 2.170334563032693, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.711761951446533, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.8871515989303589, "num_tokens": 647777287.0, "step": 17061 }, { "epoch": 2.1704617733112834, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9368131160736084, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8774082064628601, "num_tokens": 647823213.0, "step": 17062 }, { "epoch": 2.170588983589874, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.71527099609375, "learning_rate": 1e-05, "loss": 0.397, "mean_token_accuracy": 0.894023060798645, "num_tokens": 647861506.0, "step": 17063 }, { "epoch": 2.1707161938684645, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7037863731384277, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.8910441398620605, "num_tokens": 647900039.0, "step": 17064 }, { "epoch": 2.170843404147055, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.691087007522583, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8771139979362488, "num_tokens": 647938368.0, "step": 17065 }, { "epoch": 2.1709706144256455, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.534017324447632, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8681302666664124, "num_tokens": 647972549.0, "step": 17066 }, { "epoch": 2.171097824704236, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6535911560058594, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8759979009628296, "num_tokens": 648016923.0, "step": 17067 }, { "epoch": 2.1712250349828266, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.392728567123413, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.885238766670227, "num_tokens": 648059880.0, "step": 17068 }, { "epoch": 2.171352245261417, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.688493490219116, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8748824596405029, "num_tokens": 648093657.0, "step": 17069 }, { "epoch": 2.1714794555400077, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3009088039398193, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8813307881355286, "num_tokens": 648130723.0, "step": 17070 }, { "epoch": 2.171606665818598, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7545619010925293, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8720777034759521, "num_tokens": 648176657.0, "step": 17071 }, { "epoch": 2.1717338760971887, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.227515935897827, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.885067343711853, "num_tokens": 648214505.0, "step": 17072 }, { "epoch": 2.1718610863757792, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8994312286376953, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8866224884986877, "num_tokens": 648247816.0, "step": 17073 }, { "epoch": 2.1719882966543698, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2315266132354736, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.882527232170105, "num_tokens": 648276418.0, "step": 17074 }, { "epoch": 2.1721155069329603, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.052090644836426, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.869247555732727, "num_tokens": 648314722.0, "step": 17075 }, { "epoch": 2.172242717211551, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.275362730026245, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8733044862747192, "num_tokens": 648354733.0, "step": 17076 }, { "epoch": 2.1723699274901414, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.211291313171387, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8772299289703369, "num_tokens": 648392164.0, "step": 17077 }, { "epoch": 2.172497137768732, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.454465627670288, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.884316086769104, "num_tokens": 648431728.0, "step": 17078 }, { "epoch": 2.1726243480473224, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2228875160217285, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8745739459991455, "num_tokens": 648471225.0, "step": 17079 }, { "epoch": 2.172751558325913, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6378469467163086, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8795182108879089, "num_tokens": 648507742.0, "step": 17080 }, { "epoch": 2.1728787686045035, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.078950881958008, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8765997886657715, "num_tokens": 648548641.0, "step": 17081 }, { "epoch": 2.1730059788830935, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.684971809387207, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8894878625869751, "num_tokens": 648583540.0, "step": 17082 }, { "epoch": 2.173133189161684, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8854379653930664, "learning_rate": 1e-05, "loss": 0.3928, "mean_token_accuracy": 0.8929119110107422, "num_tokens": 648627225.0, "step": 17083 }, { "epoch": 2.1732603994402746, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.647096872329712, "learning_rate": 1e-05, "loss": 0.3884, "mean_token_accuracy": 0.893589198589325, "num_tokens": 648662760.0, "step": 17084 }, { "epoch": 2.173387609718865, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.736846446990967, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8738707304000854, "num_tokens": 648697748.0, "step": 17085 }, { "epoch": 2.1735148199974557, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6187479496002197, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8737515211105347, "num_tokens": 648732981.0, "step": 17086 }, { "epoch": 2.173642030276046, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.617074728012085, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.88227379322052, "num_tokens": 648764356.0, "step": 17087 }, { "epoch": 2.1737692405546367, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5277955532073975, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8775547742843628, "num_tokens": 648809064.0, "step": 17088 }, { "epoch": 2.1738964508332272, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5654470920562744, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8884379267692566, "num_tokens": 648843299.0, "step": 17089 }, { "epoch": 2.1740236611118178, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.458082437515259, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8724420070648193, "num_tokens": 648881251.0, "step": 17090 }, { "epoch": 2.1741508713904083, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.549649715423584, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8763861060142517, "num_tokens": 648915837.0, "step": 17091 }, { "epoch": 2.174278081668999, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.4046380519866943, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8684999942779541, "num_tokens": 648953354.0, "step": 17092 }, { "epoch": 2.1744052919475894, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5537822246551514, "learning_rate": 1e-05, "loss": 0.4054, "mean_token_accuracy": 0.8844408392906189, "num_tokens": 648993224.0, "step": 17093 }, { "epoch": 2.17453250222618, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3558459281921387, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8776087760925293, "num_tokens": 649026429.0, "step": 17094 }, { "epoch": 2.1746597125047704, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5921902656555176, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8852656483650208, "num_tokens": 649064672.0, "step": 17095 }, { "epoch": 2.174786922783361, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.294522762298584, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8818958401679993, "num_tokens": 649097926.0, "step": 17096 }, { "epoch": 2.1749141330619515, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.681105375289917, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8774548768997192, "num_tokens": 649138502.0, "step": 17097 }, { "epoch": 2.175041343340542, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.218017101287842, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.872258186340332, "num_tokens": 649175867.0, "step": 17098 }, { "epoch": 2.1751685536191325, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8228211402893066, "learning_rate": 1e-05, "loss": 0.3897, "mean_token_accuracy": 0.8924140334129333, "num_tokens": 649214132.0, "step": 17099 }, { "epoch": 2.175295763897723, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1754581928253174, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8775423765182495, "num_tokens": 649262397.0, "step": 17100 }, { "epoch": 2.1754229741763136, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.036602020263672, "learning_rate": 1e-05, "loss": 0.4887, "mean_token_accuracy": 0.8644217252731323, "num_tokens": 649298165.0, "step": 17101 }, { "epoch": 2.175550184454904, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.209735155105591, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8765021562576294, "num_tokens": 649337802.0, "step": 17102 }, { "epoch": 2.1756773947334946, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.24562406539917, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8896212577819824, "num_tokens": 649378922.0, "step": 17103 }, { "epoch": 2.175804605012085, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.4153921604156494, "learning_rate": 1e-05, "loss": 0.3936, "mean_token_accuracy": 0.88953697681427, "num_tokens": 649414332.0, "step": 17104 }, { "epoch": 2.1759318152906753, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.313179969787598, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8881750106811523, "num_tokens": 649448965.0, "step": 17105 }, { "epoch": 2.1760590255692662, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6687309741973877, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8781044483184814, "num_tokens": 649487322.0, "step": 17106 }, { "epoch": 2.1761862358478563, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.1639275550842285, "learning_rate": 1e-05, "loss": 0.527, "mean_token_accuracy": 0.8519273996353149, "num_tokens": 649526524.0, "step": 17107 }, { "epoch": 2.176313446126447, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7500948905944824, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8724079132080078, "num_tokens": 649566169.0, "step": 17108 }, { "epoch": 2.1764406564050374, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9408934116363525, "learning_rate": 1e-05, "loss": 0.3968, "mean_token_accuracy": 0.8897760510444641, "num_tokens": 649603725.0, "step": 17109 }, { "epoch": 2.176567866683628, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.7002112865448, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8747159242630005, "num_tokens": 649644789.0, "step": 17110 }, { "epoch": 2.1766950769622184, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.764488935470581, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8735764622688293, "num_tokens": 649682051.0, "step": 17111 }, { "epoch": 2.176822287240809, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.623685598373413, "learning_rate": 1e-05, "loss": 0.3928, "mean_token_accuracy": 0.8926514387130737, "num_tokens": 649715730.0, "step": 17112 }, { "epoch": 2.1769494975193995, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6469101905822754, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8779827952384949, "num_tokens": 649750237.0, "step": 17113 }, { "epoch": 2.17707670779799, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.535292863845825, "learning_rate": 1e-05, "loss": 0.3911, "mean_token_accuracy": 0.8905969858169556, "num_tokens": 649791250.0, "step": 17114 }, { "epoch": 2.1772039180765805, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.583580732345581, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8816683888435364, "num_tokens": 649831750.0, "step": 17115 }, { "epoch": 2.177331128355171, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.465426445007324, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8753395080566406, "num_tokens": 649869245.0, "step": 17116 }, { "epoch": 2.1774583386337616, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5782577991485596, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8769779205322266, "num_tokens": 649903763.0, "step": 17117 }, { "epoch": 2.177585548912352, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.383230209350586, "learning_rate": 1e-05, "loss": 0.3838, "mean_token_accuracy": 0.8917922973632812, "num_tokens": 649946601.0, "step": 17118 }, { "epoch": 2.1777127591909426, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.613800525665283, "learning_rate": 1e-05, "loss": 0.3879, "mean_token_accuracy": 0.8907808065414429, "num_tokens": 649979308.0, "step": 17119 }, { "epoch": 2.177839969469533, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3112916946411133, "learning_rate": 1e-05, "loss": 0.393, "mean_token_accuracy": 0.8907334804534912, "num_tokens": 650018068.0, "step": 17120 }, { "epoch": 2.1779671797481237, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.690821409225464, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8733842968940735, "num_tokens": 650059559.0, "step": 17121 }, { "epoch": 2.1780943900267142, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.227769613265991, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8822171688079834, "num_tokens": 650103266.0, "step": 17122 }, { "epoch": 2.1782216003053048, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8499577045440674, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8750362396240234, "num_tokens": 650139525.0, "step": 17123 }, { "epoch": 2.1783488105838953, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.167818546295166, "learning_rate": 1e-05, "loss": 0.3772, "mean_token_accuracy": 0.8937597274780273, "num_tokens": 650179249.0, "step": 17124 }, { "epoch": 2.178476020862486, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.089389801025391, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8678489923477173, "num_tokens": 650217258.0, "step": 17125 }, { "epoch": 2.1786032311410763, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.225101947784424, "learning_rate": 1e-05, "loss": 0.3628, "mean_token_accuracy": 0.8995729684829712, "num_tokens": 650261923.0, "step": 17126 }, { "epoch": 2.178730441419667, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.309891700744629, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8872002363204956, "num_tokens": 650295562.0, "step": 17127 }, { "epoch": 2.1788576516982574, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5060184001922607, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8778197765350342, "num_tokens": 650331988.0, "step": 17128 }, { "epoch": 2.178984861976848, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2959370613098145, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8872596025466919, "num_tokens": 650370572.0, "step": 17129 }, { "epoch": 2.179112072255438, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7153708934783936, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8894252181053162, "num_tokens": 650412838.0, "step": 17130 }, { "epoch": 2.1792392825340285, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.102626800537109, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.880054235458374, "num_tokens": 650455588.0, "step": 17131 }, { "epoch": 2.179366492812619, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7527899742126465, "learning_rate": 1e-05, "loss": 0.4004, "mean_token_accuracy": 0.8889899253845215, "num_tokens": 650496369.0, "step": 17132 }, { "epoch": 2.1794937030912096, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8958733081817627, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8703516721725464, "num_tokens": 650535366.0, "step": 17133 }, { "epoch": 2.1796209133698, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.686516046524048, "learning_rate": 1e-05, "loss": 0.3997, "mean_token_accuracy": 0.8902722597122192, "num_tokens": 650574171.0, "step": 17134 }, { "epoch": 2.1797481236483907, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.715507745742798, "learning_rate": 1e-05, "loss": 0.3827, "mean_token_accuracy": 0.8939639329910278, "num_tokens": 650612132.0, "step": 17135 }, { "epoch": 2.179875333926981, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.608372449874878, "learning_rate": 1e-05, "loss": 0.3915, "mean_token_accuracy": 0.8908272981643677, "num_tokens": 650648813.0, "step": 17136 }, { "epoch": 2.1800025442055717, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6253411769866943, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8854618668556213, "num_tokens": 650678807.0, "step": 17137 }, { "epoch": 2.1801297544841622, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5421669483184814, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8729603290557861, "num_tokens": 650716011.0, "step": 17138 }, { "epoch": 2.1802569647627528, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5465800762176514, "learning_rate": 1e-05, "loss": 0.3718, "mean_token_accuracy": 0.8958566784858704, "num_tokens": 650749763.0, "step": 17139 }, { "epoch": 2.1803841750413433, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4776771068573, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8783484697341919, "num_tokens": 650786977.0, "step": 17140 }, { "epoch": 2.180511385319934, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.525754690170288, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8705927133560181, "num_tokens": 650823293.0, "step": 17141 }, { "epoch": 2.1806385955985244, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.424269437789917, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8781205415725708, "num_tokens": 650865037.0, "step": 17142 }, { "epoch": 2.180765805877115, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5285732746124268, "learning_rate": 1e-05, "loss": 0.4038, "mean_token_accuracy": 0.8854740262031555, "num_tokens": 650901803.0, "step": 17143 }, { "epoch": 2.1808930161557054, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.365926742553711, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8874987363815308, "num_tokens": 650942163.0, "step": 17144 }, { "epoch": 2.181020226434296, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5677850246429443, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8771547079086304, "num_tokens": 650977989.0, "step": 17145 }, { "epoch": 2.1811474367128865, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.319087505340576, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8837679624557495, "num_tokens": 651016831.0, "step": 17146 }, { "epoch": 2.181274646991477, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.634430408477783, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8740545511245728, "num_tokens": 651052493.0, "step": 17147 }, { "epoch": 2.1814018572700675, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.238513708114624, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8675734996795654, "num_tokens": 651093140.0, "step": 17148 }, { "epoch": 2.181529067548658, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.764240026473999, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8739500045776367, "num_tokens": 651129367.0, "step": 17149 }, { "epoch": 2.1816562778272486, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1623003482818604, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8862261176109314, "num_tokens": 651169573.0, "step": 17150 }, { "epoch": 2.181783488105839, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9631476402282715, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.890357494354248, "num_tokens": 651206771.0, "step": 17151 }, { "epoch": 2.1819106983844296, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1479451656341553, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8907880783081055, "num_tokens": 651251322.0, "step": 17152 }, { "epoch": 2.18203790866302, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.252045631408691, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8874842524528503, "num_tokens": 651287495.0, "step": 17153 }, { "epoch": 2.1821651189416107, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3445982933044434, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8678262233734131, "num_tokens": 651326005.0, "step": 17154 }, { "epoch": 2.1822923292202008, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.408047199249268, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8690981864929199, "num_tokens": 651364843.0, "step": 17155 }, { "epoch": 2.1824195394987913, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.647585153579712, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8871075510978699, "num_tokens": 651403981.0, "step": 17156 }, { "epoch": 2.182546749777382, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.293420314788818, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8881676197052002, "num_tokens": 651443751.0, "step": 17157 }, { "epoch": 2.1826739600559724, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.8106653690338135, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8750576376914978, "num_tokens": 651487248.0, "step": 17158 }, { "epoch": 2.182801170334563, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.08402156829834, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8705523014068604, "num_tokens": 651523055.0, "step": 17159 }, { "epoch": 2.1829283806131534, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.787411689758301, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8794546127319336, "num_tokens": 651554749.0, "step": 17160 }, { "epoch": 2.183055590891744, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8676514625549316, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8612102270126343, "num_tokens": 651597342.0, "step": 17161 }, { "epoch": 2.1831828011703345, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.678182601928711, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8797767162322998, "num_tokens": 651634578.0, "step": 17162 }, { "epoch": 2.183310011448925, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.723863363265991, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8894729614257812, "num_tokens": 651674561.0, "step": 17163 }, { "epoch": 2.1834372217275155, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.569246768951416, "learning_rate": 1e-05, "loss": 0.3938, "mean_token_accuracy": 0.8927163481712341, "num_tokens": 651713426.0, "step": 17164 }, { "epoch": 2.183564432006106, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6626408100128174, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8839453458786011, "num_tokens": 651753267.0, "step": 17165 }, { "epoch": 2.1836916422846966, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.470796585083008, "learning_rate": 1e-05, "loss": 0.4026, "mean_token_accuracy": 0.886080801486969, "num_tokens": 651797580.0, "step": 17166 }, { "epoch": 2.183818852563287, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6820778846740723, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.875575065612793, "num_tokens": 651828294.0, "step": 17167 }, { "epoch": 2.1839460628418776, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.396549940109253, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8640416860580444, "num_tokens": 651865572.0, "step": 17168 }, { "epoch": 2.184073273120468, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6986358165740967, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8823023438453674, "num_tokens": 651897453.0, "step": 17169 }, { "epoch": 2.1842004833990587, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.304529905319214, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8762197494506836, "num_tokens": 651933096.0, "step": 17170 }, { "epoch": 2.1843276936776492, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7900118827819824, "learning_rate": 1e-05, "loss": 0.3966, "mean_token_accuracy": 0.8871155381202698, "num_tokens": 651970042.0, "step": 17171 }, { "epoch": 2.1844549039562398, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2545533180236816, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8764108419418335, "num_tokens": 652006513.0, "step": 17172 }, { "epoch": 2.1845821142348303, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9410648345947266, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8899639844894409, "num_tokens": 652042706.0, "step": 17173 }, { "epoch": 2.184709324513421, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2358338832855225, "learning_rate": 1e-05, "loss": 0.3929, "mean_token_accuracy": 0.8901430368423462, "num_tokens": 652084219.0, "step": 17174 }, { "epoch": 2.1848365347920113, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.123746395111084, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8890348076820374, "num_tokens": 652130640.0, "step": 17175 }, { "epoch": 2.184963745070602, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.3648617267608643, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8748480677604675, "num_tokens": 652169812.0, "step": 17176 }, { "epoch": 2.1850909553491924, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2147345542907715, "learning_rate": 1e-05, "loss": 0.3971, "mean_token_accuracy": 0.8963963985443115, "num_tokens": 652206317.0, "step": 17177 }, { "epoch": 2.185218165627783, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5675578117370605, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8769851922988892, "num_tokens": 652241420.0, "step": 17178 }, { "epoch": 2.1853453759063735, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.100158214569092, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8843770027160645, "num_tokens": 652283951.0, "step": 17179 }, { "epoch": 2.1854725861849635, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.670095920562744, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8829585313796997, "num_tokens": 652320729.0, "step": 17180 }, { "epoch": 2.185599796463554, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9034335613250732, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8749366998672485, "num_tokens": 652362969.0, "step": 17181 }, { "epoch": 2.1857270067421446, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.673771381378174, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8763688206672668, "num_tokens": 652403120.0, "step": 17182 }, { "epoch": 2.185854217020735, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7208685874938965, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8645370006561279, "num_tokens": 652436904.0, "step": 17183 }, { "epoch": 2.1859814272993257, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.624547004699707, "learning_rate": 1e-05, "loss": 0.3912, "mean_token_accuracy": 0.8908442258834839, "num_tokens": 652474938.0, "step": 17184 }, { "epoch": 2.186108637577916, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6248557567596436, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8848423361778259, "num_tokens": 652503718.0, "step": 17185 }, { "epoch": 2.1862358478565067, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5637569427490234, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8837329149246216, "num_tokens": 652542235.0, "step": 17186 }, { "epoch": 2.1863630581350972, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.506056785583496, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8636544942855835, "num_tokens": 652584779.0, "step": 17187 }, { "epoch": 2.1864902684136878, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5108184814453125, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8729995489120483, "num_tokens": 652627132.0, "step": 17188 }, { "epoch": 2.1866174786922783, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.445000410079956, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8739548921585083, "num_tokens": 652670156.0, "step": 17189 }, { "epoch": 2.186744688970869, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5104758739471436, "learning_rate": 1e-05, "loss": 0.3901, "mean_token_accuracy": 0.89015132188797, "num_tokens": 652705246.0, "step": 17190 }, { "epoch": 2.1868718992494594, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.424898862838745, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8802942633628845, "num_tokens": 652742234.0, "step": 17191 }, { "epoch": 2.18699910952805, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4839394092559814, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8867451548576355, "num_tokens": 652779100.0, "step": 17192 }, { "epoch": 2.1871263198066404, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.386410713195801, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8841077089309692, "num_tokens": 652821174.0, "step": 17193 }, { "epoch": 2.187253530085231, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.500110387802124, "learning_rate": 1e-05, "loss": 0.385, "mean_token_accuracy": 0.8904670476913452, "num_tokens": 652856271.0, "step": 17194 }, { "epoch": 2.1873807403638215, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3544814586639404, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8779593706130981, "num_tokens": 652898737.0, "step": 17195 }, { "epoch": 2.187507950642412, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.528533935546875, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8821585178375244, "num_tokens": 652938778.0, "step": 17196 }, { "epoch": 2.1876351609210025, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3079450130462646, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8779270648956299, "num_tokens": 652982202.0, "step": 17197 }, { "epoch": 2.187762371199593, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5883333683013916, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8713783025741577, "num_tokens": 653024990.0, "step": 17198 }, { "epoch": 2.1878895814781836, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2585911750793457, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8724231719970703, "num_tokens": 653061913.0, "step": 17199 }, { "epoch": 2.188016791756774, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6822545528411865, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8787047863006592, "num_tokens": 653105944.0, "step": 17200 }, { "epoch": 2.1881440020353646, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1784896850585938, "learning_rate": 1e-05, "loss": 0.3816, "mean_token_accuracy": 0.8915305137634277, "num_tokens": 653139819.0, "step": 17201 }, { "epoch": 2.188271212313955, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.858215093612671, "learning_rate": 1e-05, "loss": 0.3986, "mean_token_accuracy": 0.8920440077781677, "num_tokens": 653184569.0, "step": 17202 }, { "epoch": 2.1883984225925452, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.115015983581543, "learning_rate": 1e-05, "loss": 0.3544, "mean_token_accuracy": 0.9024608135223389, "num_tokens": 653220346.0, "step": 17203 }, { "epoch": 2.188525632871136, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.139987468719482, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8787782788276672, "num_tokens": 653257288.0, "step": 17204 }, { "epoch": 2.1886528431497263, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.1808383464813232, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8746315240859985, "num_tokens": 653293055.0, "step": 17205 }, { "epoch": 2.188780053428317, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.4443793296813965, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8720124959945679, "num_tokens": 653329988.0, "step": 17206 }, { "epoch": 2.1889072637069074, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.516042947769165, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8939286470413208, "num_tokens": 653367704.0, "step": 17207 }, { "epoch": 2.189034473985498, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.491462230682373, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8858058452606201, "num_tokens": 653403915.0, "step": 17208 }, { "epoch": 2.1891616842640884, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.8577358722686768, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8878517150878906, "num_tokens": 653438045.0, "step": 17209 }, { "epoch": 2.189288894542679, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.264644622802734, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8798667192459106, "num_tokens": 653469552.0, "step": 17210 }, { "epoch": 2.1894161048212695, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.8778951168060303, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8864702582359314, "num_tokens": 653511809.0, "step": 17211 }, { "epoch": 2.18954331509986, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.045103073120117, "learning_rate": 1e-05, "loss": 0.3947, "mean_token_accuracy": 0.8923835754394531, "num_tokens": 653543812.0, "step": 17212 }, { "epoch": 2.1896705253784505, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7730965614318848, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8715600967407227, "num_tokens": 653586232.0, "step": 17213 }, { "epoch": 2.189797735657041, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8747332096099854, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8775268793106079, "num_tokens": 653626700.0, "step": 17214 }, { "epoch": 2.1899249459356316, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.644657850265503, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8764992356300354, "num_tokens": 653666648.0, "step": 17215 }, { "epoch": 2.190052156214222, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.793229103088379, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.888659656047821, "num_tokens": 653698991.0, "step": 17216 }, { "epoch": 2.1901793664928126, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5219175815582275, "learning_rate": 1e-05, "loss": 0.3889, "mean_token_accuracy": 0.8891929388046265, "num_tokens": 653737696.0, "step": 17217 }, { "epoch": 2.190306576771403, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.772355079650879, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8881485462188721, "num_tokens": 653776386.0, "step": 17218 }, { "epoch": 2.1904337870499937, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4113001823425293, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.891137421131134, "num_tokens": 653811346.0, "step": 17219 }, { "epoch": 2.1905609973285842, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8196709156036377, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8741753697395325, "num_tokens": 653847966.0, "step": 17220 }, { "epoch": 2.1906882076071748, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3313229084014893, "learning_rate": 1e-05, "loss": 0.365, "mean_token_accuracy": 0.8956711292266846, "num_tokens": 653882381.0, "step": 17221 }, { "epoch": 2.1908154178857653, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.932058811187744, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8860877156257629, "num_tokens": 653925026.0, "step": 17222 }, { "epoch": 2.190942628164356, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.302872657775879, "learning_rate": 1e-05, "loss": 0.4121, "mean_token_accuracy": 0.8837765455245972, "num_tokens": 653962210.0, "step": 17223 }, { "epoch": 2.1910698384429463, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.085925579071045, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8833034038543701, "num_tokens": 653998713.0, "step": 17224 }, { "epoch": 2.191197048721537, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.393550157546997, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8764466047286987, "num_tokens": 654033383.0, "step": 17225 }, { "epoch": 2.1913242590001274, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.152372360229492, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8787027597427368, "num_tokens": 654074712.0, "step": 17226 }, { "epoch": 2.191451469278718, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5611557960510254, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8756439089775085, "num_tokens": 654110650.0, "step": 17227 }, { "epoch": 2.191578679557308, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.052461624145508, "learning_rate": 1e-05, "loss": 0.3884, "mean_token_accuracy": 0.8938524723052979, "num_tokens": 654146628.0, "step": 17228 }, { "epoch": 2.1917058898358985, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6404759883880615, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8718178272247314, "num_tokens": 654192239.0, "step": 17229 }, { "epoch": 2.191833100114489, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8707375526428223, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8607962131500244, "num_tokens": 654232458.0, "step": 17230 }, { "epoch": 2.1919603103930796, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6618666648864746, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8614199757575989, "num_tokens": 654271266.0, "step": 17231 }, { "epoch": 2.19208752067167, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.660191535949707, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.87992262840271, "num_tokens": 654307801.0, "step": 17232 }, { "epoch": 2.1922147309502606, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0002994537353515625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6323494911193848, "learning_rate": 1e-05, "loss": 0.3841, "mean_token_accuracy": 0.8952227830886841, "num_tokens": 654348903.0, "step": 17233 }, { "epoch": 2.192341941228851, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4981045722961426, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8862831592559814, "num_tokens": 654388659.0, "step": 17234 }, { "epoch": 2.1924691515074417, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6330325603485107, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.8837590217590332, "num_tokens": 654425231.0, "step": 17235 }, { "epoch": 2.1925963617860322, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.39827561378479, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8864455223083496, "num_tokens": 654457560.0, "step": 17236 }, { "epoch": 2.1927235720646228, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6479313373565674, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.878354012966156, "num_tokens": 654491000.0, "step": 17237 }, { "epoch": 2.1928507823432133, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.308771848678589, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8783043026924133, "num_tokens": 654531142.0, "step": 17238 }, { "epoch": 2.192977992621804, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.699453115463257, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8788117170333862, "num_tokens": 654574133.0, "step": 17239 }, { "epoch": 2.1931052029003943, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.246741533279419, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8736659288406372, "num_tokens": 654612745.0, "step": 17240 }, { "epoch": 2.193232413178985, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.81964373588562, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.882543683052063, "num_tokens": 654646097.0, "step": 17241 }, { "epoch": 2.1933596234575754, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1934478282928467, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8780548572540283, "num_tokens": 654688216.0, "step": 17242 }, { "epoch": 2.193486833736166, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9811182022094727, "learning_rate": 1e-05, "loss": 0.3736, "mean_token_accuracy": 0.9017015099525452, "num_tokens": 654725697.0, "step": 17243 }, { "epoch": 2.1936140440147565, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2263357639312744, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8796511888504028, "num_tokens": 654762440.0, "step": 17244 }, { "epoch": 2.193741254293347, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.175382137298584, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8737868070602417, "num_tokens": 654795843.0, "step": 17245 }, { "epoch": 2.1938684645719375, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.362002372741699, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8834819793701172, "num_tokens": 654835212.0, "step": 17246 }, { "epoch": 2.193995674850528, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.277366638183594, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.872064471244812, "num_tokens": 654869092.0, "step": 17247 }, { "epoch": 2.1941228851291186, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5913138389587402, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8860040903091431, "num_tokens": 654907202.0, "step": 17248 }, { "epoch": 2.194250095407709, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.178504467010498, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8822226524353027, "num_tokens": 654938867.0, "step": 17249 }, { "epoch": 2.1943773056862996, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6987833976745605, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8761396408081055, "num_tokens": 654978838.0, "step": 17250 }, { "epoch": 2.19450451596489, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.977419376373291, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8876946568489075, "num_tokens": 655016987.0, "step": 17251 }, { "epoch": 2.1946317262434807, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.683656930923462, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8812448978424072, "num_tokens": 655056797.0, "step": 17252 }, { "epoch": 2.1947589365220708, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.796306610107422, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8850128650665283, "num_tokens": 655094959.0, "step": 17253 }, { "epoch": 2.1948861468006613, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6261889934539795, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8706332445144653, "num_tokens": 655136490.0, "step": 17254 }, { "epoch": 2.195013357079252, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.653346538543701, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8832579851150513, "num_tokens": 655175174.0, "step": 17255 }, { "epoch": 2.1951405673578424, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5882909297943115, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8736343383789062, "num_tokens": 655214753.0, "step": 17256 }, { "epoch": 2.195267777636433, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5536866188049316, "learning_rate": 1e-05, "loss": 0.3807, "mean_token_accuracy": 0.8943856954574585, "num_tokens": 655249737.0, "step": 17257 }, { "epoch": 2.1953949879150234, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.515507459640503, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8887198567390442, "num_tokens": 655286553.0, "step": 17258 }, { "epoch": 2.195522198193614, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.513113498687744, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8822340369224548, "num_tokens": 655320409.0, "step": 17259 }, { "epoch": 2.1956494084722045, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4689226150512695, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8805344700813293, "num_tokens": 655362436.0, "step": 17260 }, { "epoch": 2.195776618750795, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4857630729675293, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8766502141952515, "num_tokens": 655405059.0, "step": 17261 }, { "epoch": 2.1959038290293855, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4272260665893555, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.87738037109375, "num_tokens": 655448688.0, "step": 17262 }, { "epoch": 2.196031039307976, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.489980697631836, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8856785893440247, "num_tokens": 655486361.0, "step": 17263 }, { "epoch": 2.1961582495865666, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.401409387588501, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8827641010284424, "num_tokens": 655518610.0, "step": 17264 }, { "epoch": 2.196285459865157, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5167062282562256, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8771412372589111, "num_tokens": 655557860.0, "step": 17265 }, { "epoch": 2.1964126701437476, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3457741737365723, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8811883926391602, "num_tokens": 655597885.0, "step": 17266 }, { "epoch": 2.196539880422338, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5535078048706055, "learning_rate": 1e-05, "loss": 0.3875, "mean_token_accuracy": 0.8882689476013184, "num_tokens": 655632847.0, "step": 17267 }, { "epoch": 2.1966670907009287, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.306969165802002, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8772399425506592, "num_tokens": 655665374.0, "step": 17268 }, { "epoch": 2.196794300979519, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.612358570098877, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8855406641960144, "num_tokens": 655705926.0, "step": 17269 }, { "epoch": 2.1969215112581097, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.229250192642212, "learning_rate": 1e-05, "loss": 0.3734, "mean_token_accuracy": 0.8931324481964111, "num_tokens": 655743500.0, "step": 17270 }, { "epoch": 2.1970487215367003, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.735325574874878, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8708169460296631, "num_tokens": 655788063.0, "step": 17271 }, { "epoch": 2.197175931815291, "ewc_loss": 0.07568359375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.553794860839844e-05, "grad_norm": 3.150677442550659, "learning_rate": 1e-05, "loss": 0.382, "mean_token_accuracy": 0.8918319344520569, "num_tokens": 655828393.0, "step": 17272 }, { "epoch": 2.1973031420938813, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.949084997177124, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8863928914070129, "num_tokens": 655871666.0, "step": 17273 }, { "epoch": 2.197430352372472, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1307950019836426, "learning_rate": 1e-05, "loss": 0.3747, "mean_token_accuracy": 0.8957209587097168, "num_tokens": 655905396.0, "step": 17274 }, { "epoch": 2.1975575626510624, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.233891010284424, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.878373384475708, "num_tokens": 655944515.0, "step": 17275 }, { "epoch": 2.197684772929653, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.2885806560516357, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8705315589904785, "num_tokens": 655979773.0, "step": 17276 }, { "epoch": 2.1978119832082434, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.441863536834717, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8826416730880737, "num_tokens": 656017786.0, "step": 17277 }, { "epoch": 2.1979391934868335, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.632237195968628, "learning_rate": 1e-05, "loss": 0.5061, "mean_token_accuracy": 0.8578102588653564, "num_tokens": 656058776.0, "step": 17278 }, { "epoch": 2.198066403765424, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.358150482177734, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8855966925621033, "num_tokens": 656100231.0, "step": 17279 }, { "epoch": 2.1981936140440146, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.8399808406829834, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8807625770568848, "num_tokens": 656131786.0, "step": 17280 }, { "epoch": 2.198320824322605, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.122798442840576, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8844674825668335, "num_tokens": 656172041.0, "step": 17281 }, { "epoch": 2.1984480346011956, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.800126314163208, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8888416290283203, "num_tokens": 656203152.0, "step": 17282 }, { "epoch": 2.198575244879786, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9395670890808105, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.874737024307251, "num_tokens": 656241568.0, "step": 17283 }, { "epoch": 2.1987024551583767, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.686814546585083, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8682945966720581, "num_tokens": 656275221.0, "step": 17284 }, { "epoch": 2.1988296654369672, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8087902069091797, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8738807439804077, "num_tokens": 656317535.0, "step": 17285 }, { "epoch": 2.1989568757155578, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.574807643890381, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8857728242874146, "num_tokens": 656351440.0, "step": 17286 }, { "epoch": 2.1990840859941483, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.747349739074707, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8766101598739624, "num_tokens": 656390942.0, "step": 17287 }, { "epoch": 2.199211296272739, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4748172760009766, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8865745663642883, "num_tokens": 656427314.0, "step": 17288 }, { "epoch": 2.1993385065513293, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7480597496032715, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8718660473823547, "num_tokens": 656464014.0, "step": 17289 }, { "epoch": 2.19946571682992, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.381211042404175, "learning_rate": 1e-05, "loss": 0.4102, "mean_token_accuracy": 0.880395233631134, "num_tokens": 656497783.0, "step": 17290 }, { "epoch": 2.1995929271085104, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8076682090759277, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8774627447128296, "num_tokens": 656538160.0, "step": 17291 }, { "epoch": 2.199720137387101, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3088226318359375, "learning_rate": 1e-05, "loss": 0.3971, "mean_token_accuracy": 0.8876680731773376, "num_tokens": 656578296.0, "step": 17292 }, { "epoch": 2.1998473476656915, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.929267406463623, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8813624382019043, "num_tokens": 656620662.0, "step": 17293 }, { "epoch": 2.199974557944282, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2810094356536865, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8826202750205994, "num_tokens": 656665511.0, "step": 17294 }, { "epoch": 2.2001017682228725, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.083766460418701, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8806600570678711, "num_tokens": 656703079.0, "step": 17295 }, { "epoch": 2.200228978501463, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3785457611083984, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8733718395233154, "num_tokens": 656741430.0, "step": 17296 }, { "epoch": 2.2003561887800536, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.15012788772583, "learning_rate": 1e-05, "loss": 0.4063, "mean_token_accuracy": 0.8884046673774719, "num_tokens": 656778836.0, "step": 17297 }, { "epoch": 2.200483399058644, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.546107053756714, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8827602863311768, "num_tokens": 656811331.0, "step": 17298 }, { "epoch": 2.2006106093372346, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.063324451446533, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8891113996505737, "num_tokens": 656845579.0, "step": 17299 }, { "epoch": 2.200737819615825, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6454732418060303, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.877694308757782, "num_tokens": 656880718.0, "step": 17300 }, { "epoch": 2.2008650298944152, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8630053997039795, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.882181704044342, "num_tokens": 656918350.0, "step": 17301 }, { "epoch": 2.200992240173006, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6619009971618652, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8780677318572998, "num_tokens": 656954661.0, "step": 17302 }, { "epoch": 2.2011194504515963, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6545209884643555, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.883349597454071, "num_tokens": 656993233.0, "step": 17303 }, { "epoch": 2.201246660730187, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6417202949523926, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8777045011520386, "num_tokens": 657032843.0, "step": 17304 }, { "epoch": 2.2013738710087773, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4828479290008545, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8794429302215576, "num_tokens": 657078625.0, "step": 17305 }, { "epoch": 2.201501081287368, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6606545448303223, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8827109336853027, "num_tokens": 657113696.0, "step": 17306 }, { "epoch": 2.2016282915659584, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3710439205169678, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8779986500740051, "num_tokens": 657149224.0, "step": 17307 }, { "epoch": 2.201755501844549, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6992764472961426, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.87532639503479, "num_tokens": 657181946.0, "step": 17308 }, { "epoch": 2.2018827121231395, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3056323528289795, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.857474684715271, "num_tokens": 657214372.0, "step": 17309 }, { "epoch": 2.20200992240173, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7508766651153564, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8814421892166138, "num_tokens": 657255572.0, "step": 17310 }, { "epoch": 2.2021371326803205, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.23372220993042, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8813292384147644, "num_tokens": 657288336.0, "step": 17311 }, { "epoch": 2.202264342958911, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8849682807922363, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8782261610031128, "num_tokens": 657327954.0, "step": 17312 }, { "epoch": 2.2023915532375016, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.205839157104492, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8713201284408569, "num_tokens": 657366691.0, "step": 17313 }, { "epoch": 2.202518763516092, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.058496475219727, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8797574639320374, "num_tokens": 657405694.0, "step": 17314 }, { "epoch": 2.2026459737946826, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.26975679397583, "learning_rate": 1e-05, "loss": 0.3729, "mean_token_accuracy": 0.8962681293487549, "num_tokens": 657442461.0, "step": 17315 }, { "epoch": 2.202773184073273, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.21440315246582, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8861405253410339, "num_tokens": 657476384.0, "step": 17316 }, { "epoch": 2.2029003943518637, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4685065746307373, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8713042736053467, "num_tokens": 657511291.0, "step": 17317 }, { "epoch": 2.203027604630454, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.214119911193848, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8734419345855713, "num_tokens": 657546820.0, "step": 17318 }, { "epoch": 2.2031548149090447, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.635881185531616, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8847761154174805, "num_tokens": 657582513.0, "step": 17319 }, { "epoch": 2.2032820251876353, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.066074848175049, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8819400668144226, "num_tokens": 657617108.0, "step": 17320 }, { "epoch": 2.203409235466226, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6634795665740967, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8902272582054138, "num_tokens": 657656220.0, "step": 17321 }, { "epoch": 2.2035364457448163, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8921597003936768, "learning_rate": 1e-05, "loss": 0.4029, "mean_token_accuracy": 0.8899062871932983, "num_tokens": 657691386.0, "step": 17322 }, { "epoch": 2.203663656023407, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6452674865722656, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8671611547470093, "num_tokens": 657728118.0, "step": 17323 }, { "epoch": 2.2037908663019974, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7229201793670654, "learning_rate": 1e-05, "loss": 0.387, "mean_token_accuracy": 0.8941795825958252, "num_tokens": 657768832.0, "step": 17324 }, { "epoch": 2.203918076580588, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.589862823486328, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8854143619537354, "num_tokens": 657805511.0, "step": 17325 }, { "epoch": 2.204045286859178, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6152703762054443, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.879591703414917, "num_tokens": 657843250.0, "step": 17326 }, { "epoch": 2.2041724971377685, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5247957706451416, "learning_rate": 1e-05, "loss": 0.3854, "mean_token_accuracy": 0.8913311958312988, "num_tokens": 657878882.0, "step": 17327 }, { "epoch": 2.204299707416359, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5505316257476807, "learning_rate": 1e-05, "loss": 0.3636, "mean_token_accuracy": 0.901814341545105, "num_tokens": 657917007.0, "step": 17328 }, { "epoch": 2.2044269176949496, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4769351482391357, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.870481014251709, "num_tokens": 657956741.0, "step": 17329 }, { "epoch": 2.20455412797354, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000301361083984375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.526825428009033, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8833341598510742, "num_tokens": 657995050.0, "step": 17330 }, { "epoch": 2.2046813382521306, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4250845909118652, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8767880201339722, "num_tokens": 658032708.0, "step": 17331 }, { "epoch": 2.204808548530721, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5370121002197266, "learning_rate": 1e-05, "loss": 0.3762, "mean_token_accuracy": 0.8958951234817505, "num_tokens": 658065725.0, "step": 17332 }, { "epoch": 2.2049357588093117, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.359621286392212, "learning_rate": 1e-05, "loss": 0.3888, "mean_token_accuracy": 0.8897733688354492, "num_tokens": 658107290.0, "step": 17333 }, { "epoch": 2.2050629690879022, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5746572017669678, "learning_rate": 1e-05, "loss": 0.4065, "mean_token_accuracy": 0.8844550251960754, "num_tokens": 658147516.0, "step": 17334 }, { "epoch": 2.2051901793664928, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2974982261657715, "learning_rate": 1e-05, "loss": 0.3943, "mean_token_accuracy": 0.8870888948440552, "num_tokens": 658188477.0, "step": 17335 }, { "epoch": 2.2053173896450833, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.668527603149414, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8823509216308594, "num_tokens": 658223517.0, "step": 17336 }, { "epoch": 2.205444599923674, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2377688884735107, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.88669353723526, "num_tokens": 658259167.0, "step": 17337 }, { "epoch": 2.2055718102022643, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.807816505432129, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8827582001686096, "num_tokens": 658294825.0, "step": 17338 }, { "epoch": 2.205699020480855, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.170611619949341, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8740573525428772, "num_tokens": 658336898.0, "step": 17339 }, { "epoch": 2.2058262307594454, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.00278377532959, "learning_rate": 1e-05, "loss": 0.3826, "mean_token_accuracy": 0.8958536386489868, "num_tokens": 658373201.0, "step": 17340 }, { "epoch": 2.205953441038036, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.187612533569336, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8747290372848511, "num_tokens": 658413147.0, "step": 17341 }, { "epoch": 2.2060806513166265, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.246041774749756, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8695191144943237, "num_tokens": 658453832.0, "step": 17342 }, { "epoch": 2.206207861595217, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3670225143432617, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8729970455169678, "num_tokens": 658498302.0, "step": 17343 }, { "epoch": 2.2063350718738075, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.362739086151123, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8839102983474731, "num_tokens": 658535017.0, "step": 17344 }, { "epoch": 2.206462282152398, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6598217487335205, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8639813661575317, "num_tokens": 658575495.0, "step": 17345 }, { "epoch": 2.2065894924309886, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.213866710662842, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8889156579971313, "num_tokens": 658619493.0, "step": 17346 }, { "epoch": 2.206716702709579, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7740328311920166, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8801059126853943, "num_tokens": 658654371.0, "step": 17347 }, { "epoch": 2.2068439129881696, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.012584686279297, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8735145330429077, "num_tokens": 658694543.0, "step": 17348 }, { "epoch": 2.20697112326676, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7442996501922607, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8759116530418396, "num_tokens": 658724859.0, "step": 17349 }, { "epoch": 2.2070983335453507, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.7970595359802246, "learning_rate": 1e-05, "loss": 0.3924, "mean_token_accuracy": 0.8930941820144653, "num_tokens": 658759941.0, "step": 17350 }, { "epoch": 2.2072255438239408, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6424477100372314, "learning_rate": 1e-05, "loss": 0.3965, "mean_token_accuracy": 0.8885211944580078, "num_tokens": 658801254.0, "step": 17351 }, { "epoch": 2.2073527541025313, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7028541564941406, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8780299425125122, "num_tokens": 658835177.0, "step": 17352 }, { "epoch": 2.207479964381122, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.54478120803833, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8876157999038696, "num_tokens": 658874636.0, "step": 17353 }, { "epoch": 2.2076071746597123, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6265976428985596, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8762689232826233, "num_tokens": 658911388.0, "step": 17354 }, { "epoch": 2.207734384938303, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4548258781433105, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.8864127397537231, "num_tokens": 658954685.0, "step": 17355 }, { "epoch": 2.2078615952168934, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6293439865112305, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8841956257820129, "num_tokens": 658992573.0, "step": 17356 }, { "epoch": 2.207988805495484, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.390024185180664, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8741428852081299, "num_tokens": 659024460.0, "step": 17357 }, { "epoch": 2.2081160157740745, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.666229486465454, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8614468574523926, "num_tokens": 659066103.0, "step": 17358 }, { "epoch": 2.208243226052665, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3004653453826904, "learning_rate": 1e-05, "loss": 0.3891, "mean_token_accuracy": 0.889381468296051, "num_tokens": 659097293.0, "step": 17359 }, { "epoch": 2.2083704363312555, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.750494956970215, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8750656247138977, "num_tokens": 659141221.0, "step": 17360 }, { "epoch": 2.208497646609846, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2321207523345947, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8843272924423218, "num_tokens": 659181026.0, "step": 17361 }, { "epoch": 2.2086248568884366, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.911038398742676, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8767349720001221, "num_tokens": 659218503.0, "step": 17362 }, { "epoch": 2.208752067167027, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.216759443283081, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8864078521728516, "num_tokens": 659255940.0, "step": 17363 }, { "epoch": 2.2088792774456176, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.103127956390381, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8589640855789185, "num_tokens": 659296994.0, "step": 17364 }, { "epoch": 2.209006487724208, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.323681592941284, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8723348379135132, "num_tokens": 659335372.0, "step": 17365 }, { "epoch": 2.2091336980027987, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.213048934936523, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8770098090171814, "num_tokens": 659375954.0, "step": 17366 }, { "epoch": 2.209260908281389, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5287060737609863, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8787583112716675, "num_tokens": 659409995.0, "step": 17367 }, { "epoch": 2.2093881185599797, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.145429611206055, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8885510563850403, "num_tokens": 659445640.0, "step": 17368 }, { "epoch": 2.2095153288385703, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6515119075775146, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8786458969116211, "num_tokens": 659488609.0, "step": 17369 }, { "epoch": 2.209642539117161, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.953543186187744, "learning_rate": 1e-05, "loss": 0.3913, "mean_token_accuracy": 0.8926208019256592, "num_tokens": 659533077.0, "step": 17370 }, { "epoch": 2.2097697493957513, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.660714864730835, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.883026123046875, "num_tokens": 659574080.0, "step": 17371 }, { "epoch": 2.209896959674342, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.78005313873291, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8802385926246643, "num_tokens": 659610508.0, "step": 17372 }, { "epoch": 2.2100241699529324, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6338491439819336, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8782433271408081, "num_tokens": 659648378.0, "step": 17373 }, { "epoch": 2.210151380231523, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.631326198577881, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8725366592407227, "num_tokens": 659681639.0, "step": 17374 }, { "epoch": 2.2102785905101134, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5834271907806396, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8722367882728577, "num_tokens": 659721579.0, "step": 17375 }, { "epoch": 2.2104058007887035, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.509408712387085, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8782985210418701, "num_tokens": 659761577.0, "step": 17376 }, { "epoch": 2.210533011067294, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5527398586273193, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8796421885490417, "num_tokens": 659799393.0, "step": 17377 }, { "epoch": 2.2106602213458846, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4505746364593506, "learning_rate": 1e-05, "loss": 0.4768, "mean_token_accuracy": 0.8627124428749084, "num_tokens": 659839118.0, "step": 17378 }, { "epoch": 2.210787431624475, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.524533987045288, "learning_rate": 1e-05, "loss": 0.3875, "mean_token_accuracy": 0.8929048180580139, "num_tokens": 659874955.0, "step": 17379 }, { "epoch": 2.2109146419030656, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.429269790649414, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8728783130645752, "num_tokens": 659908590.0, "step": 17380 }, { "epoch": 2.211041852181656, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.505693197250366, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8778334259986877, "num_tokens": 659945753.0, "step": 17381 }, { "epoch": 2.2111690624602467, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.396190643310547, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.8893373012542725, "num_tokens": 659981284.0, "step": 17382 }, { "epoch": 2.211296272738837, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.509585380554199, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8785333633422852, "num_tokens": 660016160.0, "step": 17383 }, { "epoch": 2.2114234830174277, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3516459465026855, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.8850271701812744, "num_tokens": 660053755.0, "step": 17384 }, { "epoch": 2.2115506932960183, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.533294916152954, "learning_rate": 1e-05, "loss": 0.3564, "mean_token_accuracy": 0.902118444442749, "num_tokens": 660092808.0, "step": 17385 }, { "epoch": 2.211677903574609, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.316728115081787, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8721534013748169, "num_tokens": 660128996.0, "step": 17386 }, { "epoch": 2.2118051138531993, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5917298793792725, "learning_rate": 1e-05, "loss": 0.3715, "mean_token_accuracy": 0.8965871334075928, "num_tokens": 660163345.0, "step": 17387 }, { "epoch": 2.21193232413179, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.262362480163574, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8825348019599915, "num_tokens": 660196483.0, "step": 17388 }, { "epoch": 2.2120595344103804, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.702512741088867, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8758823871612549, "num_tokens": 660226039.0, "step": 17389 }, { "epoch": 2.212186744688971, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.1650609970092773, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8798404932022095, "num_tokens": 660274301.0, "step": 17390 }, { "epoch": 2.2123139549675614, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8930249214172363, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8726648688316345, "num_tokens": 660312893.0, "step": 17391 }, { "epoch": 2.212441165246152, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1164116859436035, "learning_rate": 1e-05, "loss": 0.3937, "mean_token_accuracy": 0.8878768682479858, "num_tokens": 660349396.0, "step": 17392 }, { "epoch": 2.2125683755247425, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.181440830230713, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8858223557472229, "num_tokens": 660385902.0, "step": 17393 }, { "epoch": 2.212695585803333, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.214726209640503, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8764016032218933, "num_tokens": 660421527.0, "step": 17394 }, { "epoch": 2.2128227960819236, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.462840557098389, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8840532898902893, "num_tokens": 660459687.0, "step": 17395 }, { "epoch": 2.212950006360514, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.5634989738464355, "learning_rate": 1e-05, "loss": 0.4862, "mean_token_accuracy": 0.864631712436676, "num_tokens": 660502415.0, "step": 17396 }, { "epoch": 2.2130772166391046, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.469759941101074, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8818533420562744, "num_tokens": 660542503.0, "step": 17397 }, { "epoch": 2.213204426917695, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.8723220825195312, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8746659159660339, "num_tokens": 660576482.0, "step": 17398 }, { "epoch": 2.2133316371962852, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.222772598266602, "learning_rate": 1e-05, "loss": 0.3854, "mean_token_accuracy": 0.8985317945480347, "num_tokens": 660615151.0, "step": 17399 }, { "epoch": 2.213458847474876, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.866408109664917, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8771765232086182, "num_tokens": 660656555.0, "step": 17400 }, { "epoch": 2.2135860577534663, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.008602619171143, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8683247566223145, "num_tokens": 660695991.0, "step": 17401 }, { "epoch": 2.213713268032057, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7674174308776855, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8809068202972412, "num_tokens": 660727840.0, "step": 17402 }, { "epoch": 2.2138404783106473, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.841128349304199, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8848812580108643, "num_tokens": 660770107.0, "step": 17403 }, { "epoch": 2.213967688589238, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6374988555908203, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8634095191955566, "num_tokens": 660809813.0, "step": 17404 }, { "epoch": 2.2140948988678284, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7554569244384766, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8698071241378784, "num_tokens": 660849427.0, "step": 17405 }, { "epoch": 2.214222109146419, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.509021759033203, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8887116312980652, "num_tokens": 660888456.0, "step": 17406 }, { "epoch": 2.2143493194250095, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7327752113342285, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.890681266784668, "num_tokens": 660931889.0, "step": 17407 }, { "epoch": 2.2144765297036, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.403242826461792, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8832818269729614, "num_tokens": 660972965.0, "step": 17408 }, { "epoch": 2.2146037399821905, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.793212890625, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8786135911941528, "num_tokens": 661007352.0, "step": 17409 }, { "epoch": 2.214730950260781, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.316520929336548, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8845096826553345, "num_tokens": 661044202.0, "step": 17410 }, { "epoch": 2.2148581605393716, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.908435583114624, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8789829015731812, "num_tokens": 661088542.0, "step": 17411 }, { "epoch": 2.214985370817962, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.278139114379883, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.871623158454895, "num_tokens": 661130060.0, "step": 17412 }, { "epoch": 2.2151125810965526, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.088653564453125, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8728776574134827, "num_tokens": 661170325.0, "step": 17413 }, { "epoch": 2.215239791375143, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.337125062942505, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8752368688583374, "num_tokens": 661216151.0, "step": 17414 }, { "epoch": 2.2153670016537337, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.208733558654785, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8851954340934753, "num_tokens": 661257857.0, "step": 17415 }, { "epoch": 2.215494211932324, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5488674640655518, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8820797204971313, "num_tokens": 661294113.0, "step": 17416 }, { "epoch": 2.2156214222109147, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.147276401519775, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8751082420349121, "num_tokens": 661332681.0, "step": 17417 }, { "epoch": 2.2157486324895053, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.687861680984497, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8755091428756714, "num_tokens": 661373567.0, "step": 17418 }, { "epoch": 2.215875842768096, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.932110548019409, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8716782927513123, "num_tokens": 661415727.0, "step": 17419 }, { "epoch": 2.2160030530466863, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.703427791595459, "learning_rate": 1e-05, "loss": 0.4102, "mean_token_accuracy": 0.8857834339141846, "num_tokens": 661454934.0, "step": 17420 }, { "epoch": 2.216130263325277, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7134101390838623, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8634072542190552, "num_tokens": 661499070.0, "step": 17421 }, { "epoch": 2.2162574736038674, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6788084506988525, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8845874071121216, "num_tokens": 661538595.0, "step": 17422 }, { "epoch": 2.216384683882458, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5216147899627686, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8716034293174744, "num_tokens": 661578472.0, "step": 17423 }, { "epoch": 2.216511894161048, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6625638008117676, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8745610117912292, "num_tokens": 661615127.0, "step": 17424 }, { "epoch": 2.2166391044396385, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.415926218032837, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8647481203079224, "num_tokens": 661655514.0, "step": 17425 }, { "epoch": 2.216766314718229, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6844029426574707, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8695236444473267, "num_tokens": 661690838.0, "step": 17426 }, { "epoch": 2.2168935249968196, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2998440265655518, "learning_rate": 1e-05, "loss": 0.4157, "mean_token_accuracy": 0.8803486227989197, "num_tokens": 661735697.0, "step": 17427 }, { "epoch": 2.21702073527541, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7742574214935303, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.8642016649246216, "num_tokens": 661766875.0, "step": 17428 }, { "epoch": 2.2171479455540006, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2469873428344727, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8795381188392639, "num_tokens": 661798059.0, "step": 17429 }, { "epoch": 2.217275155832591, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8901329040527344, "learning_rate": 1e-05, "loss": 0.4011, "mean_token_accuracy": 0.8881760835647583, "num_tokens": 661833128.0, "step": 17430 }, { "epoch": 2.2174023661111817, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2098100185394287, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8716267943382263, "num_tokens": 661872672.0, "step": 17431 }, { "epoch": 2.217529576389772, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.056656837463379, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8705493211746216, "num_tokens": 661919611.0, "step": 17432 }, { "epoch": 2.2176567866683627, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.2984628677368164, "learning_rate": 1e-05, "loss": 0.4886, "mean_token_accuracy": 0.8625115156173706, "num_tokens": 661955349.0, "step": 17433 }, { "epoch": 2.2177839969469533, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.194730281829834, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8860832452774048, "num_tokens": 661990178.0, "step": 17434 }, { "epoch": 2.217911207225544, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.442913293838501, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8835451602935791, "num_tokens": 662035855.0, "step": 17435 }, { "epoch": 2.2180384175041343, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.2174811363220215, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8702002763748169, "num_tokens": 662073959.0, "step": 17436 }, { "epoch": 2.218165627782725, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.622908592224121, "learning_rate": 1e-05, "loss": 0.3727, "mean_token_accuracy": 0.8987836241722107, "num_tokens": 662108598.0, "step": 17437 }, { "epoch": 2.2182928380613154, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.092776298522949, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8793303966522217, "num_tokens": 662136316.0, "step": 17438 }, { "epoch": 2.218420048339906, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.679163694381714, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8814173936843872, "num_tokens": 662175771.0, "step": 17439 }, { "epoch": 2.2185472586184964, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9015254974365234, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8759652376174927, "num_tokens": 662219031.0, "step": 17440 }, { "epoch": 2.218674468897087, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.635554552078247, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8852863311767578, "num_tokens": 662257734.0, "step": 17441 }, { "epoch": 2.2188016791756775, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8204774856567383, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8807913064956665, "num_tokens": 662297329.0, "step": 17442 }, { "epoch": 2.218928889454268, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.57846736907959, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8748936057090759, "num_tokens": 662334597.0, "step": 17443 }, { "epoch": 2.2190560997328586, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.672528028488159, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8753654360771179, "num_tokens": 662370369.0, "step": 17444 }, { "epoch": 2.219183310011449, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5120716094970703, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8730391263961792, "num_tokens": 662406193.0, "step": 17445 }, { "epoch": 2.2193105202900396, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6214990615844727, "learning_rate": 1e-05, "loss": 0.3856, "mean_token_accuracy": 0.8932595252990723, "num_tokens": 662439692.0, "step": 17446 }, { "epoch": 2.21943773056863, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4399986267089844, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8841069936752319, "num_tokens": 662479768.0, "step": 17447 }, { "epoch": 2.2195649408472207, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6051254272460938, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8838088512420654, "num_tokens": 662523591.0, "step": 17448 }, { "epoch": 2.2196921511258108, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.397874355316162, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8853764533996582, "num_tokens": 662557693.0, "step": 17449 }, { "epoch": 2.2198193614044013, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.619859457015991, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8789790868759155, "num_tokens": 662595404.0, "step": 17450 }, { "epoch": 2.219946571682992, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.327094078063965, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8783779144287109, "num_tokens": 662637033.0, "step": 17451 }, { "epoch": 2.2200737819615823, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.696970224380493, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.878493070602417, "num_tokens": 662675199.0, "step": 17452 }, { "epoch": 2.220200992240173, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2583625316619873, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.879363477230072, "num_tokens": 662720809.0, "step": 17453 }, { "epoch": 2.2203282025187634, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.817106246948242, "learning_rate": 1e-05, "loss": 0.3879, "mean_token_accuracy": 0.8928866982460022, "num_tokens": 662759894.0, "step": 17454 }, { "epoch": 2.220455412797354, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.235670328140259, "learning_rate": 1e-05, "loss": 0.3894, "mean_token_accuracy": 0.8881295919418335, "num_tokens": 662795373.0, "step": 17455 }, { "epoch": 2.2205826230759445, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.963963270187378, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8877202272415161, "num_tokens": 662832764.0, "step": 17456 }, { "epoch": 2.220709833354535, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.259042501449585, "learning_rate": 1e-05, "loss": 0.4823, "mean_token_accuracy": 0.8618068695068359, "num_tokens": 662872796.0, "step": 17457 }, { "epoch": 2.2208370436331255, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.121748924255371, "learning_rate": 1e-05, "loss": 0.486, "mean_token_accuracy": 0.8706055879592896, "num_tokens": 662914027.0, "step": 17458 }, { "epoch": 2.220964253911716, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3815810680389404, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8685407638549805, "num_tokens": 662953924.0, "step": 17459 }, { "epoch": 2.2210914641903066, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.176798343658447, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8763848543167114, "num_tokens": 662982792.0, "step": 17460 }, { "epoch": 2.221218674468897, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5473103523254395, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8836764097213745, "num_tokens": 663023674.0, "step": 17461 }, { "epoch": 2.2213458847474876, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.054948806762695, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8701021671295166, "num_tokens": 663064187.0, "step": 17462 }, { "epoch": 2.221473095026078, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6184933185577393, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8929408192634583, "num_tokens": 663109146.0, "step": 17463 }, { "epoch": 2.2216003053046687, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.8763773441314697, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8832446336746216, "num_tokens": 663147988.0, "step": 17464 }, { "epoch": 2.221727515583259, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6633148193359375, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8840847015380859, "num_tokens": 663190072.0, "step": 17465 }, { "epoch": 2.2218547258618497, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.687570571899414, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8801798820495605, "num_tokens": 663222896.0, "step": 17466 }, { "epoch": 2.2219819361404403, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.60787296295166, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8791488409042358, "num_tokens": 663261846.0, "step": 17467 }, { "epoch": 2.222109146419031, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5363833904266357, "learning_rate": 1e-05, "loss": 0.3697, "mean_token_accuracy": 0.8964358568191528, "num_tokens": 663295766.0, "step": 17468 }, { "epoch": 2.2222363566976213, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.582123041152954, "learning_rate": 1e-05, "loss": 0.3827, "mean_token_accuracy": 0.8913382887840271, "num_tokens": 663333218.0, "step": 17469 }, { "epoch": 2.222363566976212, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4394350051879883, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8805919885635376, "num_tokens": 663371491.0, "step": 17470 }, { "epoch": 2.2224907772548024, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.576669216156006, "learning_rate": 1e-05, "loss": 0.3889, "mean_token_accuracy": 0.8941580057144165, "num_tokens": 663410824.0, "step": 17471 }, { "epoch": 2.222617987533393, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3847568035125732, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8848150968551636, "num_tokens": 663445624.0, "step": 17472 }, { "epoch": 2.2227451978119834, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.59965181350708, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8685591220855713, "num_tokens": 663483223.0, "step": 17473 }, { "epoch": 2.2228724080905735, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3254942893981934, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8797265291213989, "num_tokens": 663520577.0, "step": 17474 }, { "epoch": 2.222999618369164, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.641362428665161, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8790759444236755, "num_tokens": 663554667.0, "step": 17475 }, { "epoch": 2.2231268286477546, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.260939121246338, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8867383003234863, "num_tokens": 663593475.0, "step": 17476 }, { "epoch": 2.223254038926345, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7278969287872314, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8870440721511841, "num_tokens": 663635957.0, "step": 17477 }, { "epoch": 2.2233812492049356, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.192901849746704, "learning_rate": 1e-05, "loss": 0.3825, "mean_token_accuracy": 0.8924302458763123, "num_tokens": 663678610.0, "step": 17478 }, { "epoch": 2.223508459483526, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.889777660369873, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8810074329376221, "num_tokens": 663718204.0, "step": 17479 }, { "epoch": 2.2236356697621167, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.174140691757202, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8817471265792847, "num_tokens": 663760501.0, "step": 17480 }, { "epoch": 2.223762880040707, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.099830150604248, "learning_rate": 1e-05, "loss": 0.4083, "mean_token_accuracy": 0.889606237411499, "num_tokens": 663802515.0, "step": 17481 }, { "epoch": 2.2238900903192977, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.25278639793396, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.872322678565979, "num_tokens": 663839095.0, "step": 17482 }, { "epoch": 2.2240173005978883, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.303406715393066, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8670681715011597, "num_tokens": 663875734.0, "step": 17483 }, { "epoch": 2.224144510876479, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.479213237762451, "learning_rate": 1e-05, "loss": 0.4039, "mean_token_accuracy": 0.886775016784668, "num_tokens": 663911321.0, "step": 17484 }, { "epoch": 2.2242717211550693, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.320280075073242, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8769217729568481, "num_tokens": 663942334.0, "step": 17485 }, { "epoch": 2.22439893143366, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.708465099334717, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8786963820457458, "num_tokens": 663976306.0, "step": 17486 }, { "epoch": 2.2245261417122504, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.136734485626221, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8841370344161987, "num_tokens": 664016234.0, "step": 17487 }, { "epoch": 2.224653351990841, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.737197160720825, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8752962946891785, "num_tokens": 664057012.0, "step": 17488 }, { "epoch": 2.2247805622694314, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.951068639755249, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.87932288646698, "num_tokens": 664092911.0, "step": 17489 }, { "epoch": 2.224907772548022, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.685717821121216, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8731212615966797, "num_tokens": 664133069.0, "step": 17490 }, { "epoch": 2.2250349828266125, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7924492359161377, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8822391033172607, "num_tokens": 664166059.0, "step": 17491 }, { "epoch": 2.225162193105203, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6079509258270264, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8730742335319519, "num_tokens": 664204918.0, "step": 17492 }, { "epoch": 2.2252894033837936, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.681243658065796, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8824382424354553, "num_tokens": 664240561.0, "step": 17493 }, { "epoch": 2.225416613662384, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5353376865386963, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8734136819839478, "num_tokens": 664274905.0, "step": 17494 }, { "epoch": 2.2255438239409746, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6214101314544678, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8743076324462891, "num_tokens": 664316666.0, "step": 17495 }, { "epoch": 2.225671034219565, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.450993537902832, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8871569633483887, "num_tokens": 664358590.0, "step": 17496 }, { "epoch": 2.225798244498155, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.614468812942505, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8809848427772522, "num_tokens": 664394764.0, "step": 17497 }, { "epoch": 2.225925454776746, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3777685165405273, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8848369121551514, "num_tokens": 664433730.0, "step": 17498 }, { "epoch": 2.2260526650553363, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6551942825317383, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8789220452308655, "num_tokens": 664469926.0, "step": 17499 }, { "epoch": 2.226179875333927, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.307121515274048, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8715916275978088, "num_tokens": 664511204.0, "step": 17500 }, { "epoch": 2.2263070856125173, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7535316944122314, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8616733551025391, "num_tokens": 664547474.0, "step": 17501 }, { "epoch": 2.226434295891108, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.251359701156616, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8790457844734192, "num_tokens": 664579252.0, "step": 17502 }, { "epoch": 2.2265615061696984, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8988566398620605, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8708059787750244, "num_tokens": 664612129.0, "step": 17503 }, { "epoch": 2.226688716448289, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.209040641784668, "learning_rate": 1e-05, "loss": 0.378, "mean_token_accuracy": 0.8949862122535706, "num_tokens": 664650847.0, "step": 17504 }, { "epoch": 2.2268159267268794, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.08999490737915, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8715461492538452, "num_tokens": 664694056.0, "step": 17505 }, { "epoch": 2.22694313700547, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.311997413635254, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.873346745967865, "num_tokens": 664732868.0, "step": 17506 }, { "epoch": 2.2270703472840605, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.214095592498779, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8881890773773193, "num_tokens": 664772345.0, "step": 17507 }, { "epoch": 2.227197557562651, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5063114166259766, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8742162585258484, "num_tokens": 664811398.0, "step": 17508 }, { "epoch": 2.2273247678412416, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.16748046875, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8761842250823975, "num_tokens": 664846773.0, "step": 17509 }, { "epoch": 2.227451978119832, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.653607130050659, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8847436904907227, "num_tokens": 664887027.0, "step": 17510 }, { "epoch": 2.2275791883984226, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.985435724258423, "learning_rate": 1e-05, "loss": 0.3926, "mean_token_accuracy": 0.8939762115478516, "num_tokens": 664922798.0, "step": 17511 }, { "epoch": 2.227706398677013, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.673393726348877, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8825926184654236, "num_tokens": 664965862.0, "step": 17512 }, { "epoch": 2.2278336089556037, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.799339532852173, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8627477288246155, "num_tokens": 665006053.0, "step": 17513 }, { "epoch": 2.227960819234194, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6526498794555664, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8783712983131409, "num_tokens": 665040003.0, "step": 17514 }, { "epoch": 2.2280880295127847, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.620638608932495, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8740164637565613, "num_tokens": 665079989.0, "step": 17515 }, { "epoch": 2.2282152397913753, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6106762886047363, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8856660723686218, "num_tokens": 665114009.0, "step": 17516 }, { "epoch": 2.228342450069966, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5070998668670654, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8814656138420105, "num_tokens": 665153259.0, "step": 17517 }, { "epoch": 2.2284696603485563, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.56579327583313, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8895143270492554, "num_tokens": 665195024.0, "step": 17518 }, { "epoch": 2.228596870627147, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4373183250427246, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8805387020111084, "num_tokens": 665236000.0, "step": 17519 }, { "epoch": 2.2287240809057374, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.557871103286743, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8595483303070068, "num_tokens": 665277837.0, "step": 17520 }, { "epoch": 2.228851291184328, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.390043258666992, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8836522102355957, "num_tokens": 665312927.0, "step": 17521 }, { "epoch": 2.228978501462918, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5712928771972656, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.883981466293335, "num_tokens": 665346962.0, "step": 17522 }, { "epoch": 2.2291057117415085, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3343424797058105, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8833849430084229, "num_tokens": 665385684.0, "step": 17523 }, { "epoch": 2.229232922020099, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.6042587757110596, "learning_rate": 1e-05, "loss": 0.3616, "mean_token_accuracy": 0.9014180898666382, "num_tokens": 665421194.0, "step": 17524 }, { "epoch": 2.2293601322986896, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2844386100769043, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8771532773971558, "num_tokens": 665459731.0, "step": 17525 }, { "epoch": 2.22948734257728, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6929421424865723, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8825032711029053, "num_tokens": 665490310.0, "step": 17526 }, { "epoch": 2.2296145528558706, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.222470998764038, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8740313053131104, "num_tokens": 665527543.0, "step": 17527 }, { "epoch": 2.229741763134461, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.819227933883667, "learning_rate": 1e-05, "loss": 0.3863, "mean_token_accuracy": 0.8885474801063538, "num_tokens": 665562555.0, "step": 17528 }, { "epoch": 2.2298689734130517, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1692955493927, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8820457458496094, "num_tokens": 665598165.0, "step": 17529 }, { "epoch": 2.229996183691642, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.025737285614014, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8664448261260986, "num_tokens": 665637225.0, "step": 17530 }, { "epoch": 2.2301233939702327, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.186316967010498, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8731551170349121, "num_tokens": 665676126.0, "step": 17531 }, { "epoch": 2.2302506042488233, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.272151470184326, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8842623233795166, "num_tokens": 665713570.0, "step": 17532 }, { "epoch": 2.230377814527414, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.38607120513916, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8884462714195251, "num_tokens": 665743648.0, "step": 17533 }, { "epoch": 2.2305050248060043, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.383923053741455, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8810553550720215, "num_tokens": 665780249.0, "step": 17534 }, { "epoch": 2.230632235084595, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.657747507095337, "learning_rate": 1e-05, "loss": 0.3964, "mean_token_accuracy": 0.8941656351089478, "num_tokens": 665821175.0, "step": 17535 }, { "epoch": 2.2307594453631854, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.268065929412842, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8920249938964844, "num_tokens": 665854598.0, "step": 17536 }, { "epoch": 2.230886655641776, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.790485382080078, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8909642696380615, "num_tokens": 665892294.0, "step": 17537 }, { "epoch": 2.2310138659203664, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.055897235870361, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8725765943527222, "num_tokens": 665928575.0, "step": 17538 }, { "epoch": 2.231141076198957, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7520716190338135, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.8928049802780151, "num_tokens": 665963136.0, "step": 17539 }, { "epoch": 2.2312682864775475, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8569440841674805, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8812010884284973, "num_tokens": 666008556.0, "step": 17540 }, { "epoch": 2.231395496756138, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6623940467834473, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8796054720878601, "num_tokens": 666047066.0, "step": 17541 }, { "epoch": 2.2315227070347285, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.731675386428833, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8832906484603882, "num_tokens": 666084998.0, "step": 17542 }, { "epoch": 2.231649917313319, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.573596954345703, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8761656880378723, "num_tokens": 666124935.0, "step": 17543 }, { "epoch": 2.2317771275919096, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.663896322250366, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.864890456199646, "num_tokens": 666165580.0, "step": 17544 }, { "epoch": 2.2319043378705, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.475325584411621, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8713657855987549, "num_tokens": 666206020.0, "step": 17545 }, { "epoch": 2.2320315481490907, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6601312160491943, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8694113492965698, "num_tokens": 666239757.0, "step": 17546 }, { "epoch": 2.2321587584276807, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.374753475189209, "learning_rate": 1e-05, "loss": 0.3784, "mean_token_accuracy": 0.8944355845451355, "num_tokens": 666278426.0, "step": 17547 }, { "epoch": 2.2322859687062713, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.711040735244751, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8839564919471741, "num_tokens": 666315784.0, "step": 17548 }, { "epoch": 2.232413178984862, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.309296131134033, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8709723949432373, "num_tokens": 666353851.0, "step": 17549 }, { "epoch": 2.2325403892634523, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8134472370147705, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8750763535499573, "num_tokens": 666391208.0, "step": 17550 }, { "epoch": 2.232667599542043, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2442498207092285, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.877682089805603, "num_tokens": 666434477.0, "step": 17551 }, { "epoch": 2.2327948098206334, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9665584564208984, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8773367404937744, "num_tokens": 666472481.0, "step": 17552 }, { "epoch": 2.232922020099224, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2776317596435547, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.878609836101532, "num_tokens": 666502858.0, "step": 17553 }, { "epoch": 2.2330492303778144, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.116524696350098, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.8641794919967651, "num_tokens": 666547311.0, "step": 17554 }, { "epoch": 2.233176440656405, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.3730921745300293, "learning_rate": 1e-05, "loss": 0.4029, "mean_token_accuracy": 0.8894254565238953, "num_tokens": 666587818.0, "step": 17555 }, { "epoch": 2.2333036509349955, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.198037147521973, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8805959820747375, "num_tokens": 666626071.0, "step": 17556 }, { "epoch": 2.233430861213586, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.568880796432495, "learning_rate": 1e-05, "loss": 0.3812, "mean_token_accuracy": 0.8953378796577454, "num_tokens": 666659221.0, "step": 17557 }, { "epoch": 2.2335580714921766, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.084067344665527, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8737722039222717, "num_tokens": 666696189.0, "step": 17558 }, { "epoch": 2.233685281770767, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.664451837539673, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8707873821258545, "num_tokens": 666738495.0, "step": 17559 }, { "epoch": 2.2338124920493576, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8958990573883057, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8848056793212891, "num_tokens": 666772848.0, "step": 17560 }, { "epoch": 2.233939702327948, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6554694175720215, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.882404088973999, "num_tokens": 666813849.0, "step": 17561 }, { "epoch": 2.2340669126065387, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7007124423980713, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8813222646713257, "num_tokens": 666848564.0, "step": 17562 }, { "epoch": 2.234194122885129, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.661832332611084, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8817731738090515, "num_tokens": 666877381.0, "step": 17563 }, { "epoch": 2.2343213331637197, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.519911050796509, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8710898160934448, "num_tokens": 666920452.0, "step": 17564 }, { "epoch": 2.2344485434423103, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.626523017883301, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8857063055038452, "num_tokens": 666958799.0, "step": 17565 }, { "epoch": 2.234575753720901, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.413473606109619, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8826277852058411, "num_tokens": 666993899.0, "step": 17566 }, { "epoch": 2.2347029639994913, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6390862464904785, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8759334683418274, "num_tokens": 667035154.0, "step": 17567 }, { "epoch": 2.234830174278082, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.333969831466675, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8718119859695435, "num_tokens": 667077047.0, "step": 17568 }, { "epoch": 2.2349573845566724, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.68318247795105, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8827128410339355, "num_tokens": 667114298.0, "step": 17569 }, { "epoch": 2.235084594835263, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2631192207336426, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8842161893844604, "num_tokens": 667157942.0, "step": 17570 }, { "epoch": 2.2352118051138534, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.805697202682495, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8710165023803711, "num_tokens": 667195547.0, "step": 17571 }, { "epoch": 2.2353390153924435, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2060043811798096, "learning_rate": 1e-05, "loss": 0.3794, "mean_token_accuracy": 0.8928216695785522, "num_tokens": 667232484.0, "step": 17572 }, { "epoch": 2.235466225671034, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.977586507797241, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8706083297729492, "num_tokens": 667267661.0, "step": 17573 }, { "epoch": 2.2355934359496246, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2127809524536133, "learning_rate": 1e-05, "loss": 0.3831, "mean_token_accuracy": 0.8922630548477173, "num_tokens": 667303963.0, "step": 17574 }, { "epoch": 2.235720646228215, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.168478488922119, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.87865149974823, "num_tokens": 667344103.0, "step": 17575 }, { "epoch": 2.2358478565068056, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3555662631988525, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8804481029510498, "num_tokens": 667377782.0, "step": 17576 }, { "epoch": 2.235975066785396, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.265392303466797, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8804422616958618, "num_tokens": 667419015.0, "step": 17577 }, { "epoch": 2.2361022770639867, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5602266788482666, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8820136189460754, "num_tokens": 667457465.0, "step": 17578 }, { "epoch": 2.236229487342577, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1999688148498535, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.870918869972229, "num_tokens": 667497671.0, "step": 17579 }, { "epoch": 2.2363566976211677, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6834514141082764, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.8897019624710083, "num_tokens": 667534258.0, "step": 17580 }, { "epoch": 2.2364839078997583, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0348968505859375, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8807315230369568, "num_tokens": 667572689.0, "step": 17581 }, { "epoch": 2.236611118178349, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.694911479949951, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8746849894523621, "num_tokens": 667612425.0, "step": 17582 }, { "epoch": 2.2367383284569393, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8499655723571777, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8732415437698364, "num_tokens": 667653309.0, "step": 17583 }, { "epoch": 2.23686553873553, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6258840560913086, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8825435638427734, "num_tokens": 667699174.0, "step": 17584 }, { "epoch": 2.2369927490141204, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7169384956359863, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8826121091842651, "num_tokens": 667738987.0, "step": 17585 }, { "epoch": 2.237119959292711, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5577657222747803, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8833789825439453, "num_tokens": 667780678.0, "step": 17586 }, { "epoch": 2.2372471695713014, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.644558906555176, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8854630589485168, "num_tokens": 667816298.0, "step": 17587 }, { "epoch": 2.237374379849892, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.495068311691284, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8689799308776855, "num_tokens": 667855723.0, "step": 17588 }, { "epoch": 2.2375015901284825, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.628793239593506, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8712528944015503, "num_tokens": 667893895.0, "step": 17589 }, { "epoch": 2.237628800407073, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4072649478912354, "learning_rate": 1e-05, "loss": 0.3764, "mean_token_accuracy": 0.8964053392410278, "num_tokens": 667932061.0, "step": 17590 }, { "epoch": 2.2377560106856635, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.653759002685547, "learning_rate": 1e-05, "loss": 0.4023, "mean_token_accuracy": 0.887925386428833, "num_tokens": 667965723.0, "step": 17591 }, { "epoch": 2.237883220964254, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3320696353912354, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8762305378913879, "num_tokens": 668002310.0, "step": 17592 }, { "epoch": 2.2380104312428446, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.720574378967285, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8668386936187744, "num_tokens": 668044970.0, "step": 17593 }, { "epoch": 2.238137641521435, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2695226669311523, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8700410723686218, "num_tokens": 668081898.0, "step": 17594 }, { "epoch": 2.238264851800025, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8548412322998047, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8792482614517212, "num_tokens": 668119072.0, "step": 17595 }, { "epoch": 2.238392062078616, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2269227504730225, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.883683443069458, "num_tokens": 668157063.0, "step": 17596 }, { "epoch": 2.2385192723572063, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.038026332855225, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8583823442459106, "num_tokens": 668196985.0, "step": 17597 }, { "epoch": 2.238646482635797, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.277860164642334, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8874999284744263, "num_tokens": 668235954.0, "step": 17598 }, { "epoch": 2.2387736929143873, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.179513454437256, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8794615268707275, "num_tokens": 668275442.0, "step": 17599 }, { "epoch": 2.238900903192978, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.470158576965332, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8803598880767822, "num_tokens": 668311413.0, "step": 17600 }, { "epoch": 2.2390281134715684, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.160121440887451, "learning_rate": 1e-05, "loss": 0.3929, "mean_token_accuracy": 0.8935948610305786, "num_tokens": 668346928.0, "step": 17601 }, { "epoch": 2.239155323750159, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.612187623977661, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8885520696640015, "num_tokens": 668388566.0, "step": 17602 }, { "epoch": 2.2392825340287494, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9978044033050537, "learning_rate": 1e-05, "loss": 0.3956, "mean_token_accuracy": 0.8924657702445984, "num_tokens": 668430253.0, "step": 17603 }, { "epoch": 2.23940974430734, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6614365577697754, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8708318471908569, "num_tokens": 668470561.0, "step": 17604 }, { "epoch": 2.2395369545859305, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8070785999298096, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8794841170310974, "num_tokens": 668509718.0, "step": 17605 }, { "epoch": 2.239664164864521, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.64394474029541, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8752751350402832, "num_tokens": 668548367.0, "step": 17606 }, { "epoch": 2.2397913751431116, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6345162391662598, "learning_rate": 1e-05, "loss": 0.4878, "mean_token_accuracy": 0.8599410057067871, "num_tokens": 668586977.0, "step": 17607 }, { "epoch": 2.239918585421702, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6240787506103516, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8810216188430786, "num_tokens": 668617801.0, "step": 17608 }, { "epoch": 2.2400457957002926, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.499256134033203, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8723180294036865, "num_tokens": 668656074.0, "step": 17609 }, { "epoch": 2.240173005978883, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5846920013427734, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8886180520057678, "num_tokens": 668698354.0, "step": 17610 }, { "epoch": 2.2403002162574737, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.416616201400757, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8864197731018066, "num_tokens": 668734822.0, "step": 17611 }, { "epoch": 2.240427426536064, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5713460445404053, "learning_rate": 1e-05, "loss": 0.3575, "mean_token_accuracy": 0.9009538888931274, "num_tokens": 668773136.0, "step": 17612 }, { "epoch": 2.2405546368146547, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3695852756500244, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.878990113735199, "num_tokens": 668811580.0, "step": 17613 }, { "epoch": 2.2406818470932452, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.598733425140381, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8814993500709534, "num_tokens": 668846743.0, "step": 17614 }, { "epoch": 2.2408090573718358, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.305147171020508, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8869513273239136, "num_tokens": 668885634.0, "step": 17615 }, { "epoch": 2.2409362676504263, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6681463718414307, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8816748857498169, "num_tokens": 668920319.0, "step": 17616 }, { "epoch": 2.241063477929017, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.250230312347412, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8819615840911865, "num_tokens": 668959741.0, "step": 17617 }, { "epoch": 2.2411906882076074, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.79270339012146, "learning_rate": 1e-05, "loss": 0.5032, "mean_token_accuracy": 0.8578774333000183, "num_tokens": 668998469.0, "step": 17618 }, { "epoch": 2.241317898486198, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2014002799987793, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8672252893447876, "num_tokens": 669030018.0, "step": 17619 }, { "epoch": 2.241445108764788, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9426345825195312, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.8904451727867126, "num_tokens": 669066393.0, "step": 17620 }, { "epoch": 2.2415723190433785, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.165518045425415, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8881441354751587, "num_tokens": 669108321.0, "step": 17621 }, { "epoch": 2.241699529321969, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.179937839508057, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8819607496261597, "num_tokens": 669147878.0, "step": 17622 }, { "epoch": 2.2418267396005596, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3003294467926025, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8692748546600342, "num_tokens": 669192348.0, "step": 17623 }, { "epoch": 2.24195394987915, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.346930503845215, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.870485782623291, "num_tokens": 669233763.0, "step": 17624 }, { "epoch": 2.2420811601577406, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.5737574100494385, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8862941265106201, "num_tokens": 669269609.0, "step": 17625 }, { "epoch": 2.242208370436331, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.3055877685546875, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8785594701766968, "num_tokens": 669308649.0, "step": 17626 }, { "epoch": 2.2423355807149217, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7545711994171143, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8796602487564087, "num_tokens": 669350550.0, "step": 17627 }, { "epoch": 2.242462790993512, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.083409309387207, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.87871253490448, "num_tokens": 669392795.0, "step": 17628 }, { "epoch": 2.2425900012721027, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7643203735351562, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8741157054901123, "num_tokens": 669426928.0, "step": 17629 }, { "epoch": 2.2427172115506933, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8712923526763916, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.886755108833313, "num_tokens": 669467064.0, "step": 17630 }, { "epoch": 2.242844421829284, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.679072141647339, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.871269166469574, "num_tokens": 669506656.0, "step": 17631 }, { "epoch": 2.2429716321078743, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7324302196502686, "learning_rate": 1e-05, "loss": 0.5219, "mean_token_accuracy": 0.8511948585510254, "num_tokens": 669545787.0, "step": 17632 }, { "epoch": 2.243098842386465, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6051037311553955, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8782444000244141, "num_tokens": 669581242.0, "step": 17633 }, { "epoch": 2.2432260526650554, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6233901977539062, "learning_rate": 1e-05, "loss": 0.3841, "mean_token_accuracy": 0.8923700451850891, "num_tokens": 669621586.0, "step": 17634 }, { "epoch": 2.243353262943646, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5168914794921875, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8813736438751221, "num_tokens": 669662903.0, "step": 17635 }, { "epoch": 2.2434804732222364, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.589257001876831, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8816386461257935, "num_tokens": 669698409.0, "step": 17636 }, { "epoch": 2.243607683500827, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4378607273101807, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.864429235458374, "num_tokens": 669741344.0, "step": 17637 }, { "epoch": 2.2437348937794175, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.582798480987549, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8867765665054321, "num_tokens": 669780042.0, "step": 17638 }, { "epoch": 2.243862104058008, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3768458366394043, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8622363805770874, "num_tokens": 669819340.0, "step": 17639 }, { "epoch": 2.2439893143365985, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.614532470703125, "learning_rate": 1e-05, "loss": 0.3897, "mean_token_accuracy": 0.8909623622894287, "num_tokens": 669859269.0, "step": 17640 }, { "epoch": 2.244116524615189, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2985458374023438, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8824417591094971, "num_tokens": 669903085.0, "step": 17641 }, { "epoch": 2.2442437348937796, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.708798885345459, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.8872345089912415, "num_tokens": 669945329.0, "step": 17642 }, { "epoch": 2.24437094517237, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.241245746612549, "learning_rate": 1e-05, "loss": 0.4024, "mean_token_accuracy": 0.8862529993057251, "num_tokens": 669985141.0, "step": 17643 }, { "epoch": 2.2444981554509607, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8626179695129395, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8693459033966064, "num_tokens": 670024034.0, "step": 17644 }, { "epoch": 2.2446253657295507, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1938650608062744, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8836732506752014, "num_tokens": 670062098.0, "step": 17645 }, { "epoch": 2.2447525760081413, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.073454856872559, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8767595887184143, "num_tokens": 670100801.0, "step": 17646 }, { "epoch": 2.244879786286732, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2424514293670654, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8752070665359497, "num_tokens": 670142221.0, "step": 17647 }, { "epoch": 2.2450069965653223, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.265150547027588, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8855330348014832, "num_tokens": 670183131.0, "step": 17648 }, { "epoch": 2.245134206843913, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4652490615844727, "learning_rate": 1e-05, "loss": 0.3725, "mean_token_accuracy": 0.894882082939148, "num_tokens": 670219207.0, "step": 17649 }, { "epoch": 2.2452614171225034, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.290227890014648, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8788278102874756, "num_tokens": 670254085.0, "step": 17650 }, { "epoch": 2.245388627401094, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.692042112350464, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8698822259902954, "num_tokens": 670299669.0, "step": 17651 }, { "epoch": 2.2455158376796844, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.112107753753662, "learning_rate": 1e-05, "loss": 0.3894, "mean_token_accuracy": 0.8947246074676514, "num_tokens": 670332898.0, "step": 17652 }, { "epoch": 2.245643047958275, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.748999834060669, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8801860809326172, "num_tokens": 670371635.0, "step": 17653 }, { "epoch": 2.2457702582368655, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.888040542602539, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8766992092132568, "num_tokens": 670412568.0, "step": 17654 }, { "epoch": 2.245897468515456, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6784563064575195, "learning_rate": 1e-05, "loss": 0.3976, "mean_token_accuracy": 0.8906379342079163, "num_tokens": 670454338.0, "step": 17655 }, { "epoch": 2.2460246787940465, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7175145149230957, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.886411190032959, "num_tokens": 670490826.0, "step": 17656 }, { "epoch": 2.246151889072637, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6231260299682617, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8692439198493958, "num_tokens": 670528806.0, "step": 17657 }, { "epoch": 2.2462790993512276, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6028616428375244, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8771477937698364, "num_tokens": 670563349.0, "step": 17658 }, { "epoch": 2.246406309629818, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5481693744659424, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8744730949401855, "num_tokens": 670604682.0, "step": 17659 }, { "epoch": 2.2465335199084087, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.545165538787842, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8685364127159119, "num_tokens": 670644807.0, "step": 17660 }, { "epoch": 2.246660730186999, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4842076301574707, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8839631080627441, "num_tokens": 670683788.0, "step": 17661 }, { "epoch": 2.2467879404655897, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5091893672943115, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8855679035186768, "num_tokens": 670728270.0, "step": 17662 }, { "epoch": 2.2469151507441802, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4427144527435303, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8810550570487976, "num_tokens": 670763203.0, "step": 17663 }, { "epoch": 2.2470423610227708, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5037553310394287, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8735886812210083, "num_tokens": 670804548.0, "step": 17664 }, { "epoch": 2.2471695713013613, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.395158052444458, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8771647214889526, "num_tokens": 670844922.0, "step": 17665 }, { "epoch": 2.247296781579952, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5207083225250244, "learning_rate": 1e-05, "loss": 0.3883, "mean_token_accuracy": 0.890168309211731, "num_tokens": 670882369.0, "step": 17666 }, { "epoch": 2.2474239918585424, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.347198963165283, "learning_rate": 1e-05, "loss": 0.3872, "mean_token_accuracy": 0.8921077847480774, "num_tokens": 670922716.0, "step": 17667 }, { "epoch": 2.247551202137133, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.57306170463562, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8739991188049316, "num_tokens": 670964618.0, "step": 17668 }, { "epoch": 2.2476784124157234, "ewc_loss": 0.07666015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2871406078338623, "learning_rate": 1e-05, "loss": 0.3656, "mean_token_accuracy": 0.8979186415672302, "num_tokens": 670997915.0, "step": 17669 }, { "epoch": 2.2478056226943135, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6674787998199463, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8717174530029297, "num_tokens": 671031675.0, "step": 17670 }, { "epoch": 2.247932832972904, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.601478576660156e-05, "grad_norm": 3.224181890487671, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8708094358444214, "num_tokens": 671068060.0, "step": 17671 }, { "epoch": 2.2480600432514946, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7989871501922607, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8824120759963989, "num_tokens": 671109014.0, "step": 17672 }, { "epoch": 2.248187253530085, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.163299798965454, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.882610023021698, "num_tokens": 671140173.0, "step": 17673 }, { "epoch": 2.2483144638086756, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.021444320678711, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.875209629535675, "num_tokens": 671177644.0, "step": 17674 }, { "epoch": 2.248441674087266, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1531662940979004, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.878057599067688, "num_tokens": 671221860.0, "step": 17675 }, { "epoch": 2.2485688843658567, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.302635192871094, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8803497552871704, "num_tokens": 671262567.0, "step": 17676 }, { "epoch": 2.248696094644447, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.3764474391937256, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8743909001350403, "num_tokens": 671301195.0, "step": 17677 }, { "epoch": 2.2488233049230377, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.43927001953125, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8773109912872314, "num_tokens": 671346028.0, "step": 17678 }, { "epoch": 2.2489505152016283, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.6969780921936035, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8846265077590942, "num_tokens": 671386927.0, "step": 17679 }, { "epoch": 2.249077725480219, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.311307430267334, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8741784691810608, "num_tokens": 671422447.0, "step": 17680 }, { "epoch": 2.2492049357588093, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.84370493888855, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.875361979007721, "num_tokens": 671461390.0, "step": 17681 }, { "epoch": 2.2493321460374, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.074072360992432, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8801376223564148, "num_tokens": 671499150.0, "step": 17682 }, { "epoch": 2.2494593563159904, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7767016887664795, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8796008825302124, "num_tokens": 671537529.0, "step": 17683 }, { "epoch": 2.249586566594581, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.8954765796661377, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8743605613708496, "num_tokens": 671577661.0, "step": 17684 }, { "epoch": 2.2497137768731714, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6745026111602783, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8848581314086914, "num_tokens": 671611375.0, "step": 17685 }, { "epoch": 2.249840987151762, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7589378356933594, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8697806000709534, "num_tokens": 671655430.0, "step": 17686 }, { "epoch": 2.2499681974303525, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.556863784790039, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8848199844360352, "num_tokens": 671696451.0, "step": 17687 }, { "epoch": 2.250095407708943, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7127838134765625, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8738257884979248, "num_tokens": 671736576.0, "step": 17688 }, { "epoch": 2.2502226179875335, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4607248306274414, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8808552622795105, "num_tokens": 671774638.0, "step": 17689 }, { "epoch": 2.250349828266124, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7158443927764893, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8768293857574463, "num_tokens": 671812626.0, "step": 17690 }, { "epoch": 2.2504770385447146, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3875062465667725, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8828125596046448, "num_tokens": 671844752.0, "step": 17691 }, { "epoch": 2.250604248823305, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7684521675109863, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.876983642578125, "num_tokens": 671883161.0, "step": 17692 }, { "epoch": 2.250731459101895, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.295469284057617, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8847920894622803, "num_tokens": 671924386.0, "step": 17693 }, { "epoch": 2.250858669380486, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.9019012451171875, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8788571357727051, "num_tokens": 671958452.0, "step": 17694 }, { "epoch": 2.2509858796590763, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2786781787872314, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.874563992023468, "num_tokens": 671995335.0, "step": 17695 }, { "epoch": 2.251113089937667, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.071213722229004, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8745309114456177, "num_tokens": 672025640.0, "step": 17696 }, { "epoch": 2.2512403002162573, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.343024253845215, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8722467422485352, "num_tokens": 672066013.0, "step": 17697 }, { "epoch": 2.251367510494848, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.170309543609619, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8794164657592773, "num_tokens": 672107243.0, "step": 17698 }, { "epoch": 2.2514947207734384, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.506324529647827, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8809230327606201, "num_tokens": 672145753.0, "step": 17699 }, { "epoch": 2.251621931052029, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.128015518188477, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8816457390785217, "num_tokens": 672185489.0, "step": 17700 }, { "epoch": 2.2517491413306194, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6377205848693848, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8833260536193848, "num_tokens": 672225702.0, "step": 17701 }, { "epoch": 2.25187635160921, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.955477237701416, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8742773532867432, "num_tokens": 672265901.0, "step": 17702 }, { "epoch": 2.2520035618878005, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6768133640289307, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8694384694099426, "num_tokens": 672304081.0, "step": 17703 }, { "epoch": 2.252130772166391, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7465546131134033, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8829121589660645, "num_tokens": 672343425.0, "step": 17704 }, { "epoch": 2.2522579824449815, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6500234603881836, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8845453262329102, "num_tokens": 672377745.0, "step": 17705 }, { "epoch": 2.252385192723572, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5786540508270264, "learning_rate": 1e-05, "loss": 0.4004, "mean_token_accuracy": 0.8874726295471191, "num_tokens": 672413811.0, "step": 17706 }, { "epoch": 2.2525124030021626, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.628636121749878, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8764537572860718, "num_tokens": 672455083.0, "step": 17707 }, { "epoch": 2.252639613280753, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.429471969604492, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8875976800918579, "num_tokens": 672502401.0, "step": 17708 }, { "epoch": 2.2527668235593437, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6366050243377686, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8827904462814331, "num_tokens": 672540504.0, "step": 17709 }, { "epoch": 2.252894033837934, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3532376289367676, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8820040225982666, "num_tokens": 672578157.0, "step": 17710 }, { "epoch": 2.2530212441165247, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.685577630996704, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8793954253196716, "num_tokens": 672611859.0, "step": 17711 }, { "epoch": 2.2531484543951152, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2763469219207764, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8769374489784241, "num_tokens": 672651779.0, "step": 17712 }, { "epoch": 2.2532756646737058, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.77272891998291, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8763065338134766, "num_tokens": 672691559.0, "step": 17713 }, { "epoch": 2.2534028749522963, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.221146583557129, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8820316791534424, "num_tokens": 672731353.0, "step": 17714 }, { "epoch": 2.253530085230887, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.918262243270874, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8843705654144287, "num_tokens": 672770427.0, "step": 17715 }, { "epoch": 2.2536572955094774, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2174651622772217, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8775878548622131, "num_tokens": 672808817.0, "step": 17716 }, { "epoch": 2.253784505788068, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.089975833892822, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8798021078109741, "num_tokens": 672848368.0, "step": 17717 }, { "epoch": 2.253911716066658, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.2881131172180176, "learning_rate": 1e-05, "loss": 0.3885, "mean_token_accuracy": 0.8914190530776978, "num_tokens": 672885074.0, "step": 17718 }, { "epoch": 2.254038926345249, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.239584922790527, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8811130523681641, "num_tokens": 672924866.0, "step": 17719 }, { "epoch": 2.254166136623839, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.496670722961426, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.884786069393158, "num_tokens": 672962051.0, "step": 17720 }, { "epoch": 2.2542933469024296, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.219570636749268, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8842365145683289, "num_tokens": 672999954.0, "step": 17721 }, { "epoch": 2.25442055718102, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6595001220703125, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8647712469100952, "num_tokens": 673037825.0, "step": 17722 }, { "epoch": 2.2545477674596106, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.061947822570801, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.892786979675293, "num_tokens": 673072036.0, "step": 17723 }, { "epoch": 2.254674977738201, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6868536472320557, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8681315183639526, "num_tokens": 673110733.0, "step": 17724 }, { "epoch": 2.2548021880167917, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8732857704162598, "learning_rate": 1e-05, "loss": 0.4026, "mean_token_accuracy": 0.8908491134643555, "num_tokens": 673146611.0, "step": 17725 }, { "epoch": 2.254929398295382, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.631817579269409, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8785772323608398, "num_tokens": 673189965.0, "step": 17726 }, { "epoch": 2.2550566085739727, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.743901014328003, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8691632151603699, "num_tokens": 673226990.0, "step": 17727 }, { "epoch": 2.2551838188525632, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.583350658416748, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8756208419799805, "num_tokens": 673263109.0, "step": 17728 }, { "epoch": 2.2553110291311538, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.630235195159912, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8860291242599487, "num_tokens": 673304081.0, "step": 17729 }, { "epoch": 2.2554382394097443, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.503262996673584, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8883132934570312, "num_tokens": 673345510.0, "step": 17730 }, { "epoch": 2.255565449688335, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.604362726211548, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8758813142776489, "num_tokens": 673383607.0, "step": 17731 }, { "epoch": 2.2556926599669254, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.431933879852295, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8772150874137878, "num_tokens": 673422676.0, "step": 17732 }, { "epoch": 2.255819870245516, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.610342502593994, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8809152841567993, "num_tokens": 673466135.0, "step": 17733 }, { "epoch": 2.2559470805241064, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.361635446548462, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.870404839515686, "num_tokens": 673509003.0, "step": 17734 }, { "epoch": 2.256074290802697, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.66434907913208, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8793792128562927, "num_tokens": 673546231.0, "step": 17735 }, { "epoch": 2.2562015010812875, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2942776679992676, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8759153485298157, "num_tokens": 673585812.0, "step": 17736 }, { "epoch": 2.256328711359878, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.761108875274658, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.8906147480010986, "num_tokens": 673621566.0, "step": 17737 }, { "epoch": 2.2564559216384685, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.21283221244812, "learning_rate": 1e-05, "loss": 0.3613, "mean_token_accuracy": 0.8984212279319763, "num_tokens": 673665659.0, "step": 17738 }, { "epoch": 2.256583131917059, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9392080307006836, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8880953788757324, "num_tokens": 673705936.0, "step": 17739 }, { "epoch": 2.2567103421956496, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.225186586380005, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8740544319152832, "num_tokens": 673740741.0, "step": 17740 }, { "epoch": 2.2568375524742397, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.126179218292236, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.883476972579956, "num_tokens": 673778304.0, "step": 17741 }, { "epoch": 2.2569647627528306, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.333369016647339, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8822759389877319, "num_tokens": 673815136.0, "step": 17742 }, { "epoch": 2.2570919730314207, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.24561071395874, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.883478045463562, "num_tokens": 673847227.0, "step": 17743 }, { "epoch": 2.2572191833100113, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.5523316860198975, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8915359973907471, "num_tokens": 673885110.0, "step": 17744 }, { "epoch": 2.257346393588602, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.144965171813965, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8705628514289856, "num_tokens": 673929534.0, "step": 17745 }, { "epoch": 2.2574736038671923, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.669752359390259, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8708096742630005, "num_tokens": 673972353.0, "step": 17746 }, { "epoch": 2.257600814145783, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.955972909927368, "learning_rate": 1e-05, "loss": 0.3922, "mean_token_accuracy": 0.8882644176483154, "num_tokens": 674009441.0, "step": 17747 }, { "epoch": 2.2577280244243734, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.674116849899292, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.890240490436554, "num_tokens": 674045397.0, "step": 17748 }, { "epoch": 2.257855234702964, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7617969512939453, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8722051382064819, "num_tokens": 674084896.0, "step": 17749 }, { "epoch": 2.2579824449815544, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6384851932525635, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8696141839027405, "num_tokens": 674121223.0, "step": 17750 }, { "epoch": 2.258109655260145, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6352760791778564, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8782906532287598, "num_tokens": 674157395.0, "step": 17751 }, { "epoch": 2.2582368655387355, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5796444416046143, "learning_rate": 1e-05, "loss": 0.4935, "mean_token_accuracy": 0.8605136871337891, "num_tokens": 674200157.0, "step": 17752 }, { "epoch": 2.258364075817326, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.532672643661499, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8653176426887512, "num_tokens": 674234696.0, "step": 17753 }, { "epoch": 2.2584912860959165, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5338358879089355, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8778029680252075, "num_tokens": 674277456.0, "step": 17754 }, { "epoch": 2.258618496374507, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.459254026412964, "learning_rate": 1e-05, "loss": 0.3896, "mean_token_accuracy": 0.8917236328125, "num_tokens": 674316913.0, "step": 17755 }, { "epoch": 2.2587457066530976, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5192127227783203, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8812727928161621, "num_tokens": 674355539.0, "step": 17756 }, { "epoch": 2.258872916931688, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4382762908935547, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8714962005615234, "num_tokens": 674386759.0, "step": 17757 }, { "epoch": 2.2590001272102787, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4985196590423584, "learning_rate": 1e-05, "loss": 0.3951, "mean_token_accuracy": 0.8861565589904785, "num_tokens": 674427098.0, "step": 17758 }, { "epoch": 2.259127337488869, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.406653642654419, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8859362602233887, "num_tokens": 674464766.0, "step": 17759 }, { "epoch": 2.2592545477674597, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4932141304016113, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8761082887649536, "num_tokens": 674506260.0, "step": 17760 }, { "epoch": 2.2593817580460502, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3837337493896484, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8811428546905518, "num_tokens": 674544255.0, "step": 17761 }, { "epoch": 2.2595089683246408, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.500953197479248, "learning_rate": 1e-05, "loss": 0.3822, "mean_token_accuracy": 0.889060378074646, "num_tokens": 674579806.0, "step": 17762 }, { "epoch": 2.2596361786032313, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.361402750015259, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8677960634231567, "num_tokens": 674620734.0, "step": 17763 }, { "epoch": 2.259763388881822, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.600376844406128, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8845727443695068, "num_tokens": 674661213.0, "step": 17764 }, { "epoch": 2.2598905991604123, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2771551609039307, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8874632120132446, "num_tokens": 674696323.0, "step": 17765 }, { "epoch": 2.2600178094390024, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.678631067276001, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8767530918121338, "num_tokens": 674728041.0, "step": 17766 }, { "epoch": 2.2601450197175934, "ewc_loss": 0.076171875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.173898220062256, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8852500915527344, "num_tokens": 674767518.0, "step": 17767 }, { "epoch": 2.2602722299961835, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.85752534866333, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.881585955619812, "num_tokens": 674807510.0, "step": 17768 }, { "epoch": 2.260399440274774, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.105358123779297, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8799722194671631, "num_tokens": 674844851.0, "step": 17769 }, { "epoch": 2.2605266505533645, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.147635459899902, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.8938244581222534, "num_tokens": 674881839.0, "step": 17770 }, { "epoch": 2.260653860831955, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.1718857288360596, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8777522444725037, "num_tokens": 674920450.0, "step": 17771 }, { "epoch": 2.2607810711105456, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.469656944274902, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.881850004196167, "num_tokens": 674960031.0, "step": 17772 }, { "epoch": 2.260908281389136, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.525102138519287, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8873609900474548, "num_tokens": 674993354.0, "step": 17773 }, { "epoch": 2.2610354916677267, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.528892993927002, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8837159276008606, "num_tokens": 675029186.0, "step": 17774 }, { "epoch": 2.261162701946317, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 3.8768773078918457, "learning_rate": 1e-05, "loss": 0.4046, "mean_token_accuracy": 0.8898674249649048, "num_tokens": 675065170.0, "step": 17775 }, { "epoch": 2.2612899122249077, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.27676248550415, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8927834033966064, "num_tokens": 675102774.0, "step": 17776 }, { "epoch": 2.2614171225034982, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.9174931049346924, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8728346824645996, "num_tokens": 675136890.0, "step": 17777 }, { "epoch": 2.2615443327820888, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.055123329162598, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8874017000198364, "num_tokens": 675171959.0, "step": 17778 }, { "epoch": 2.2616715430606793, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.795978546142578, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8844515085220337, "num_tokens": 675203931.0, "step": 17779 }, { "epoch": 2.26179875333927, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.884409189224243, "learning_rate": 1e-05, "loss": 0.3853, "mean_token_accuracy": 0.8927375078201294, "num_tokens": 675243645.0, "step": 17780 }, { "epoch": 2.2619259636178604, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6523959636688232, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8841997385025024, "num_tokens": 675280318.0, "step": 17781 }, { "epoch": 2.262053173896451, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8106529712677, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8758589625358582, "num_tokens": 675321442.0, "step": 17782 }, { "epoch": 2.2621803841750414, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.542358636856079, "learning_rate": 1e-05, "loss": 0.3775, "mean_token_accuracy": 0.8960126638412476, "num_tokens": 675351039.0, "step": 17783 }, { "epoch": 2.262307594453632, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7804629802703857, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.881827712059021, "num_tokens": 675385756.0, "step": 17784 }, { "epoch": 2.2624348047322225, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.417598247528076, "learning_rate": 1e-05, "loss": 0.3736, "mean_token_accuracy": 0.8959041833877563, "num_tokens": 675429282.0, "step": 17785 }, { "epoch": 2.262562015010813, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.825878620147705, "learning_rate": 1e-05, "loss": 0.3942, "mean_token_accuracy": 0.8896814584732056, "num_tokens": 675466661.0, "step": 17786 }, { "epoch": 2.2626892252894035, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.339582920074463, "learning_rate": 1e-05, "loss": 0.3873, "mean_token_accuracy": 0.8889822363853455, "num_tokens": 675501885.0, "step": 17787 }, { "epoch": 2.262816435567994, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9376797676086426, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8690894842147827, "num_tokens": 675544423.0, "step": 17788 }, { "epoch": 2.2629436458465846, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.308690071105957, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8780434727668762, "num_tokens": 675585301.0, "step": 17789 }, { "epoch": 2.263070856125175, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.087448596954346, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8786495923995972, "num_tokens": 675625244.0, "step": 17790 }, { "epoch": 2.263198066403765, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3912787437438965, "learning_rate": 1e-05, "loss": 0.3845, "mean_token_accuracy": 0.8929187655448914, "num_tokens": 675657310.0, "step": 17791 }, { "epoch": 2.263325276682356, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.184062957763672, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8801303505897522, "num_tokens": 675690702.0, "step": 17792 }, { "epoch": 2.2634524869609463, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.552938938140869, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.8894409537315369, "num_tokens": 675725211.0, "step": 17793 }, { "epoch": 2.263579697239537, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.094151496887207, "learning_rate": 1e-05, "loss": 0.3983, "mean_token_accuracy": 0.8929654359817505, "num_tokens": 675762418.0, "step": 17794 }, { "epoch": 2.2637069075181273, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6645514965057373, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8672060966491699, "num_tokens": 675807640.0, "step": 17795 }, { "epoch": 2.263834117796718, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.911187171936035, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8810721635818481, "num_tokens": 675843032.0, "step": 17796 }, { "epoch": 2.2639613280753084, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.7008163928985596, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8826303482055664, "num_tokens": 675875150.0, "step": 17797 }, { "epoch": 2.264088538353899, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6882355213165283, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.865221381187439, "num_tokens": 675917635.0, "step": 17798 }, { "epoch": 2.2642157486324894, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6666669845581055, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8759210109710693, "num_tokens": 675955577.0, "step": 17799 }, { "epoch": 2.26434295891108, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5629477500915527, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8652541041374207, "num_tokens": 675996326.0, "step": 17800 }, { "epoch": 2.2644701691896705, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6320672035217285, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8650640249252319, "num_tokens": 676040857.0, "step": 17801 }, { "epoch": 2.264597379468261, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4246420860290527, "learning_rate": 1e-05, "loss": 0.3773, "mean_token_accuracy": 0.8911036252975464, "num_tokens": 676077601.0, "step": 17802 }, { "epoch": 2.2647245897468515, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.640411138534546, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8773506879806519, "num_tokens": 676115558.0, "step": 17803 }, { "epoch": 2.264851800025442, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3364920616149902, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8784610033035278, "num_tokens": 676157828.0, "step": 17804 }, { "epoch": 2.2649790103040326, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.692037343978882, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8696644306182861, "num_tokens": 676199147.0, "step": 17805 }, { "epoch": 2.265106220582623, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003032684326171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2586820125579834, "learning_rate": 1e-05, "loss": 0.38, "mean_token_accuracy": 0.8915265202522278, "num_tokens": 676241271.0, "step": 17806 }, { "epoch": 2.2652334308612136, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8012051582336426, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8808948993682861, "num_tokens": 676276990.0, "step": 17807 }, { "epoch": 2.265360641139804, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.228344202041626, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8782994151115417, "num_tokens": 676313441.0, "step": 17808 }, { "epoch": 2.2654878514183947, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.95827317237854, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8776199221611023, "num_tokens": 676350322.0, "step": 17809 }, { "epoch": 2.2656150616969852, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.212942361831665, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8859863877296448, "num_tokens": 676386113.0, "step": 17810 }, { "epoch": 2.2657422719755758, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.143208980560303, "learning_rate": 1e-05, "loss": 0.3882, "mean_token_accuracy": 0.8932533264160156, "num_tokens": 676421995.0, "step": 17811 }, { "epoch": 2.2658694822541663, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.3298110961914062, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8800449967384338, "num_tokens": 676459761.0, "step": 17812 }, { "epoch": 2.265996692532757, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.267724990844727, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8780838251113892, "num_tokens": 676503870.0, "step": 17813 }, { "epoch": 2.2661239028113473, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.544426679611206, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8904699683189392, "num_tokens": 676544173.0, "step": 17814 }, { "epoch": 2.266251113089938, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.219823360443115, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8815715312957764, "num_tokens": 676580896.0, "step": 17815 }, { "epoch": 2.266378323368528, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6956660747528076, "learning_rate": 1e-05, "loss": 0.4125, "mean_token_accuracy": 0.8852569460868835, "num_tokens": 676620127.0, "step": 17816 }, { "epoch": 2.266505533647119, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.040165901184082, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8884361982345581, "num_tokens": 676655988.0, "step": 17817 }, { "epoch": 2.266632743925709, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.7111377716064453, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8793978095054626, "num_tokens": 676689125.0, "step": 17818 }, { "epoch": 2.2667599542042995, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8413426876068115, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8771482110023499, "num_tokens": 676729231.0, "step": 17819 }, { "epoch": 2.26688716448289, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.648390769958496, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8753859996795654, "num_tokens": 676770130.0, "step": 17820 }, { "epoch": 2.2670143747614806, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.696129322052002, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8844446539878845, "num_tokens": 676807847.0, "step": 17821 }, { "epoch": 2.267141585040071, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.571211814880371, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8891129493713379, "num_tokens": 676841050.0, "step": 17822 }, { "epoch": 2.2672687953186617, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.618013620376587, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8768966197967529, "num_tokens": 676877114.0, "step": 17823 }, { "epoch": 2.267396005597252, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4909818172454834, "learning_rate": 1e-05, "loss": 0.369, "mean_token_accuracy": 0.8983539342880249, "num_tokens": 676917913.0, "step": 17824 }, { "epoch": 2.2675232158758427, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5704915523529053, "learning_rate": 1e-05, "loss": 0.379, "mean_token_accuracy": 0.8924144506454468, "num_tokens": 676955974.0, "step": 17825 }, { "epoch": 2.2676504261544332, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4305686950683594, "learning_rate": 1e-05, "loss": 0.3871, "mean_token_accuracy": 0.8914105892181396, "num_tokens": 676991285.0, "step": 17826 }, { "epoch": 2.2677776364330238, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.59488582611084, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8750332593917847, "num_tokens": 677025472.0, "step": 17827 }, { "epoch": 2.2679048467116143, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3566198348999023, "learning_rate": 1e-05, "loss": 0.4028, "mean_token_accuracy": 0.8850575685501099, "num_tokens": 677065047.0, "step": 17828 }, { "epoch": 2.268032056990205, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6416430473327637, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.874801754951477, "num_tokens": 677103027.0, "step": 17829 }, { "epoch": 2.2681592672687954, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2803242206573486, "learning_rate": 1e-05, "loss": 0.3809, "mean_token_accuracy": 0.8930552005767822, "num_tokens": 677142302.0, "step": 17830 }, { "epoch": 2.268286477547386, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.746260643005371, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8754168152809143, "num_tokens": 677186963.0, "step": 17831 }, { "epoch": 2.2684136878259764, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.222660541534424, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.860310435295105, "num_tokens": 677232227.0, "step": 17832 }, { "epoch": 2.268540898104567, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.908735513687134, "learning_rate": 1e-05, "loss": 0.384, "mean_token_accuracy": 0.8969227075576782, "num_tokens": 677274213.0, "step": 17833 }, { "epoch": 2.2686681083831575, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.197819232940674, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8862698674201965, "num_tokens": 677313870.0, "step": 17834 }, { "epoch": 2.268795318661748, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.123456954956055, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8833163380622864, "num_tokens": 677350478.0, "step": 17835 }, { "epoch": 2.2689225289403385, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3050599098205566, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8748878240585327, "num_tokens": 677392253.0, "step": 17836 }, { "epoch": 2.269049739218929, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.27427339553833, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8878556489944458, "num_tokens": 677429024.0, "step": 17837 }, { "epoch": 2.2691769494975196, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5418941974639893, "learning_rate": 1e-05, "loss": 0.3896, "mean_token_accuracy": 0.8948417901992798, "num_tokens": 677461705.0, "step": 17838 }, { "epoch": 2.2693041597761097, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.207157135009766, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8814690113067627, "num_tokens": 677505586.0, "step": 17839 }, { "epoch": 2.2694313700547006, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7412328720092773, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8822394013404846, "num_tokens": 677544892.0, "step": 17840 }, { "epoch": 2.2695585803332907, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.983522653579712, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8785216808319092, "num_tokens": 677586721.0, "step": 17841 }, { "epoch": 2.2696857906118812, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.712491512298584, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8725128173828125, "num_tokens": 677627657.0, "step": 17842 }, { "epoch": 2.2698130008904718, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.766340970993042, "learning_rate": 1e-05, "loss": 0.3692, "mean_token_accuracy": 0.8982636332511902, "num_tokens": 677667513.0, "step": 17843 }, { "epoch": 2.2699402111690623, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.644974946975708, "learning_rate": 1e-05, "loss": 0.4125, "mean_token_accuracy": 0.8844506740570068, "num_tokens": 677702120.0, "step": 17844 }, { "epoch": 2.270067421447653, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6366970539093018, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8688082695007324, "num_tokens": 677739353.0, "step": 17845 }, { "epoch": 2.2701946317262434, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.584397315979004, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8675752878189087, "num_tokens": 677778464.0, "step": 17846 }, { "epoch": 2.270321842004834, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5637238025665283, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8790731430053711, "num_tokens": 677815169.0, "step": 17847 }, { "epoch": 2.2704490522834244, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5069479942321777, "learning_rate": 1e-05, "loss": 0.3833, "mean_token_accuracy": 0.8919066190719604, "num_tokens": 677852824.0, "step": 17848 }, { "epoch": 2.270576262562015, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5065078735351562, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8713887929916382, "num_tokens": 677895274.0, "step": 17849 }, { "epoch": 2.2707034728406055, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4674034118652344, "learning_rate": 1e-05, "loss": 0.3902, "mean_token_accuracy": 0.8902292251586914, "num_tokens": 677929956.0, "step": 17850 }, { "epoch": 2.270830683119196, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5058610439300537, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8721369504928589, "num_tokens": 677965921.0, "step": 17851 }, { "epoch": 2.2709578933977865, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4140632152557373, "learning_rate": 1e-05, "loss": 0.3847, "mean_token_accuracy": 0.8903082609176636, "num_tokens": 678005701.0, "step": 17852 }, { "epoch": 2.271085103676377, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5224218368530273, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8801704049110413, "num_tokens": 678040275.0, "step": 17853 }, { "epoch": 2.2712123139549676, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.383661985397339, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8827764987945557, "num_tokens": 678071963.0, "step": 17854 }, { "epoch": 2.271339524233558, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.553992509841919, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8782148361206055, "num_tokens": 678103422.0, "step": 17855 }, { "epoch": 2.2714667345121486, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3172495365142822, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.881857693195343, "num_tokens": 678144288.0, "step": 17856 }, { "epoch": 2.271593944790739, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00030517578125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6207103729248047, "learning_rate": 1e-05, "loss": 0.3768, "mean_token_accuracy": 0.8933475017547607, "num_tokens": 678175575.0, "step": 17857 }, { "epoch": 2.2717211550693297, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2605605125427246, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8681781888008118, "num_tokens": 678218089.0, "step": 17858 }, { "epoch": 2.2718483653479202, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7199923992156982, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8726093769073486, "num_tokens": 678258777.0, "step": 17859 }, { "epoch": 2.2719755756265108, "ewc_loss": 0.0771484375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1822478771209717, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8747316598892212, "num_tokens": 678300354.0, "step": 17860 }, { "epoch": 2.2721027859051013, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9093940258026123, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8773491382598877, "num_tokens": 678343579.0, "step": 17861 }, { "epoch": 2.272229996183692, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1453704833984375, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8822587728500366, "num_tokens": 678377059.0, "step": 17862 }, { "epoch": 2.2723572064622823, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.171658039093018, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8730306625366211, "num_tokens": 678417608.0, "step": 17863 }, { "epoch": 2.2724844167408724, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.2446885108947754, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8795093297958374, "num_tokens": 678455473.0, "step": 17864 }, { "epoch": 2.2726116270194634, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.3951735496521, "learning_rate": 1e-05, "loss": 0.4967, "mean_token_accuracy": 0.8680000305175781, "num_tokens": 678500258.0, "step": 17865 }, { "epoch": 2.2727388372980535, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.544027328491211, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8755719065666199, "num_tokens": 678542219.0, "step": 17866 }, { "epoch": 2.272866047576644, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.395970344543457, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8784593343734741, "num_tokens": 678579364.0, "step": 17867 }, { "epoch": 2.2729932578552345, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.803838014602661, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8667235374450684, "num_tokens": 678616549.0, "step": 17868 }, { "epoch": 2.273120468133825, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.179239273071289, "learning_rate": 1e-05, "loss": 0.3912, "mean_token_accuracy": 0.8933875560760498, "num_tokens": 678652324.0, "step": 17869 }, { "epoch": 2.2732476784124156, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.8081183433532715, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8772724866867065, "num_tokens": 678693574.0, "step": 17870 }, { "epoch": 2.273374888691006, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9947104454040527, "learning_rate": 1e-05, "loss": 0.4987, "mean_token_accuracy": 0.8637546300888062, "num_tokens": 678730097.0, "step": 17871 }, { "epoch": 2.2735020989695967, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.729498863220215, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8791764974594116, "num_tokens": 678768392.0, "step": 17872 }, { "epoch": 2.273629309248187, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8308353424072266, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8737647533416748, "num_tokens": 678804290.0, "step": 17873 }, { "epoch": 2.2737565195267777, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6149492263793945, "learning_rate": 1e-05, "loss": 0.3727, "mean_token_accuracy": 0.8992213010787964, "num_tokens": 678839535.0, "step": 17874 }, { "epoch": 2.2738837298053682, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7326927185058594, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8824396133422852, "num_tokens": 678879179.0, "step": 17875 }, { "epoch": 2.2740109400839588, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5047106742858887, "learning_rate": 1e-05, "loss": 0.3935, "mean_token_accuracy": 0.889116644859314, "num_tokens": 678916804.0, "step": 17876 }, { "epoch": 2.2741381503625493, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.724114179611206, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8797942399978638, "num_tokens": 678950386.0, "step": 17877 }, { "epoch": 2.27426536064114, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4126107692718506, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8801174163818359, "num_tokens": 678988367.0, "step": 17878 }, { "epoch": 2.2743925709197303, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7526988983154297, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.890724241733551, "num_tokens": 679022577.0, "step": 17879 }, { "epoch": 2.274519781198321, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3270342350006104, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8808099031448364, "num_tokens": 679059613.0, "step": 17880 }, { "epoch": 2.2746469914769114, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.851287841796875, "learning_rate": 1e-05, "loss": 0.3852, "mean_token_accuracy": 0.8932682871818542, "num_tokens": 679092956.0, "step": 17881 }, { "epoch": 2.274774201755502, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2853317260742188, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8829033970832825, "num_tokens": 679131124.0, "step": 17882 }, { "epoch": 2.2749014120340925, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9977335929870605, "learning_rate": 1e-05, "loss": 0.3862, "mean_token_accuracy": 0.89404296875, "num_tokens": 679166077.0, "step": 17883 }, { "epoch": 2.275028622312683, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.309786796569824, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8794083595275879, "num_tokens": 679204427.0, "step": 17884 }, { "epoch": 2.2751558325912735, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.145710468292236, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8821290731430054, "num_tokens": 679236003.0, "step": 17885 }, { "epoch": 2.275283042869864, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4406747817993164, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8822547197341919, "num_tokens": 679276347.0, "step": 17886 }, { "epoch": 2.2754102531484546, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.154901027679443, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8834353089332581, "num_tokens": 679318335.0, "step": 17887 }, { "epoch": 2.275537463427045, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.608354330062866, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8650593757629395, "num_tokens": 679361253.0, "step": 17888 }, { "epoch": 2.275664673705635, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.013611793518066, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8741217255592346, "num_tokens": 679400234.0, "step": 17889 }, { "epoch": 2.275791883984226, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6751556396484375, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8792542815208435, "num_tokens": 679438538.0, "step": 17890 }, { "epoch": 2.2759190942628162, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7983031272888184, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8769940137863159, "num_tokens": 679475425.0, "step": 17891 }, { "epoch": 2.2760463045414068, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.667191505432129, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8851234912872314, "num_tokens": 679517447.0, "step": 17892 }, { "epoch": 2.2761735148199973, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.609520196914673, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8828011751174927, "num_tokens": 679549933.0, "step": 17893 }, { "epoch": 2.276300725098588, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.647196054458618, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8859401345252991, "num_tokens": 679590498.0, "step": 17894 }, { "epoch": 2.2764279353771784, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.453531265258789, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8772204518318176, "num_tokens": 679634582.0, "step": 17895 }, { "epoch": 2.276555145655769, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6546473503112793, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8843651413917542, "num_tokens": 679672324.0, "step": 17896 }, { "epoch": 2.2766823559343594, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003070831298828125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.364205837249756, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8826805353164673, "num_tokens": 679706426.0, "step": 17897 }, { "epoch": 2.27680956621295, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6830389499664307, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.8937900066375732, "num_tokens": 679744706.0, "step": 17898 }, { "epoch": 2.2769367764915405, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.300043821334839, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8806337118148804, "num_tokens": 679779939.0, "step": 17899 }, { "epoch": 2.277063986770131, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7643442153930664, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8830339908599854, "num_tokens": 679820286.0, "step": 17900 }, { "epoch": 2.2771911970487215, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2421436309814453, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8752622008323669, "num_tokens": 679857985.0, "step": 17901 }, { "epoch": 2.277318407327312, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8867201805114746, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.8907398581504822, "num_tokens": 679896379.0, "step": 17902 }, { "epoch": 2.2774456176059026, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2205541133880615, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8723049163818359, "num_tokens": 679933162.0, "step": 17903 }, { "epoch": 2.277572827884493, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.054295539855957, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8650850057601929, "num_tokens": 679975828.0, "step": 17904 }, { "epoch": 2.2777000381630836, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.2904410362243652, "learning_rate": 1e-05, "loss": 0.4877, "mean_token_accuracy": 0.8588119745254517, "num_tokens": 680014678.0, "step": 17905 }, { "epoch": 2.277827248441674, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.188226222991943, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8816960453987122, "num_tokens": 680050912.0, "step": 17906 }, { "epoch": 2.2779544587202647, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.444502830505371, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8863892555236816, "num_tokens": 680082591.0, "step": 17907 }, { "epoch": 2.2780816689988552, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.19901704788208, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8778729438781738, "num_tokens": 680122100.0, "step": 17908 }, { "epoch": 2.2782088792774458, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5929172039031982, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8712731599807739, "num_tokens": 680160926.0, "step": 17909 }, { "epoch": 2.2783360895560363, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.086802959442139, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.877219557762146, "num_tokens": 680202621.0, "step": 17910 }, { "epoch": 2.278463299834627, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.645317554473877, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8788211941719055, "num_tokens": 680241129.0, "step": 17911 }, { "epoch": 2.2785905101132173, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.928232431411743, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8807971477508545, "num_tokens": 680279832.0, "step": 17912 }, { "epoch": 2.278717720391808, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6534061431884766, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8701175451278687, "num_tokens": 680313541.0, "step": 17913 }, { "epoch": 2.278844930670398, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7559778690338135, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8732588291168213, "num_tokens": 680355756.0, "step": 17914 }, { "epoch": 2.278972140948989, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5840320587158203, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.875650942325592, "num_tokens": 680396276.0, "step": 17915 }, { "epoch": 2.279099351227579, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.656747579574585, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.874106228351593, "num_tokens": 680435196.0, "step": 17916 }, { "epoch": 2.2792265615061695, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5145671367645264, "learning_rate": 1e-05, "loss": 0.3899, "mean_token_accuracy": 0.8911845684051514, "num_tokens": 680473088.0, "step": 17917 }, { "epoch": 2.27935377178476, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6099507808685303, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8741440773010254, "num_tokens": 680506269.0, "step": 17918 }, { "epoch": 2.2794809820633506, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4616196155548096, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8653333783149719, "num_tokens": 680543464.0, "step": 17919 }, { "epoch": 2.279608192341941, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.58598256111145, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8744646906852722, "num_tokens": 680582730.0, "step": 17920 }, { "epoch": 2.2797354026205316, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3924646377563477, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8832987546920776, "num_tokens": 680625773.0, "step": 17921 }, { "epoch": 2.279862612899122, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6381256580352783, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8776049613952637, "num_tokens": 680655890.0, "step": 17922 }, { "epoch": 2.2799898231777127, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3251900672912598, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8818569183349609, "num_tokens": 680689921.0, "step": 17923 }, { "epoch": 2.2801170334563032, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.704948663711548, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8683299422264099, "num_tokens": 680735744.0, "step": 17924 }, { "epoch": 2.2802442437348938, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.276552677154541, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8774772882461548, "num_tokens": 680766925.0, "step": 17925 }, { "epoch": 2.2803714540134843, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.815312147140503, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8839212656021118, "num_tokens": 680804324.0, "step": 17926 }, { "epoch": 2.280498664292075, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2137887477874756, "learning_rate": 1e-05, "loss": 0.3942, "mean_token_accuracy": 0.8871090412139893, "num_tokens": 680842005.0, "step": 17927 }, { "epoch": 2.2806258745706653, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.008062839508057, "learning_rate": 1e-05, "loss": 0.4102, "mean_token_accuracy": 0.885577917098999, "num_tokens": 680873475.0, "step": 17928 }, { "epoch": 2.280753084849256, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.244513988494873, "learning_rate": 1e-05, "loss": 0.4033, "mean_token_accuracy": 0.8854950070381165, "num_tokens": 680911322.0, "step": 17929 }, { "epoch": 2.2808802951278464, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.186474323272705, "learning_rate": 1e-05, "loss": 0.4969, "mean_token_accuracy": 0.8574854135513306, "num_tokens": 680949778.0, "step": 17930 }, { "epoch": 2.281007505406437, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.4009640216827393, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8758155703544617, "num_tokens": 680992795.0, "step": 17931 }, { "epoch": 2.2811347156850275, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.246910572052002, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8678051233291626, "num_tokens": 681026430.0, "step": 17932 }, { "epoch": 2.281261925963618, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.5935144424438477, "learning_rate": 1e-05, "loss": 0.3286, "mean_token_accuracy": 0.9141465425491333, "num_tokens": 681062998.0, "step": 17933 }, { "epoch": 2.2813891362422085, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.118006706237793, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8768694400787354, "num_tokens": 681100956.0, "step": 17934 }, { "epoch": 2.281516346520799, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.699944496154785, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8847613334655762, "num_tokens": 681136406.0, "step": 17935 }, { "epoch": 2.2816435567993896, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.915357828140259, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8667256832122803, "num_tokens": 681174899.0, "step": 17936 }, { "epoch": 2.2817707670779797, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6889209747314453, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8734793066978455, "num_tokens": 681207662.0, "step": 17937 }, { "epoch": 2.2818979773565706, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.712850332260132, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.871620774269104, "num_tokens": 681248819.0, "step": 17938 }, { "epoch": 2.2820251876351607, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6291162967681885, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8799076080322266, "num_tokens": 681286416.0, "step": 17939 }, { "epoch": 2.2821523979137512, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5752389430999756, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8637250065803528, "num_tokens": 681325629.0, "step": 17940 }, { "epoch": 2.2822796081923418, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.593848705291748, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8654708862304688, "num_tokens": 681362310.0, "step": 17941 }, { "epoch": 2.2824068184709323, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.482663631439209, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8814578652381897, "num_tokens": 681401123.0, "step": 17942 }, { "epoch": 2.282534028749523, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5556206703186035, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8801202774047852, "num_tokens": 681439801.0, "step": 17943 }, { "epoch": 2.2826612390281134, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.434844970703125, "learning_rate": 1e-05, "loss": 0.3746, "mean_token_accuracy": 0.8933873176574707, "num_tokens": 681470032.0, "step": 17944 }, { "epoch": 2.282788449306704, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5406296253204346, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8853358626365662, "num_tokens": 681512421.0, "step": 17945 }, { "epoch": 2.2829156595852944, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.398362398147583, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.87786465883255, "num_tokens": 681550226.0, "step": 17946 }, { "epoch": 2.283042869863885, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.54982590675354, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8856663703918457, "num_tokens": 681587885.0, "step": 17947 }, { "epoch": 2.2831700801424755, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.347515344619751, "learning_rate": 1e-05, "loss": 0.3703, "mean_token_accuracy": 0.8946930170059204, "num_tokens": 681623230.0, "step": 17948 }, { "epoch": 2.283297290421066, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5994150638580322, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8872416019439697, "num_tokens": 681659419.0, "step": 17949 }, { "epoch": 2.2834245006996565, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2884130477905273, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8774316906929016, "num_tokens": 681699845.0, "step": 17950 }, { "epoch": 2.283551710978247, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6971170902252197, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.869330883026123, "num_tokens": 681737579.0, "step": 17951 }, { "epoch": 2.2836789212568376, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.225896120071411, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8717966079711914, "num_tokens": 681771174.0, "step": 17952 }, { "epoch": 2.283806131535428, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.84446120262146, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8732307553291321, "num_tokens": 681810641.0, "step": 17953 }, { "epoch": 2.2839333418140186, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.165778636932373, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8704686164855957, "num_tokens": 681849273.0, "step": 17954 }, { "epoch": 2.284060552092609, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.069882392883301, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8867275714874268, "num_tokens": 681893105.0, "step": 17955 }, { "epoch": 2.2841877623711997, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.182753801345825, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8845564723014832, "num_tokens": 681931364.0, "step": 17956 }, { "epoch": 2.28431497264979, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.328423976898193, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8832758665084839, "num_tokens": 681970809.0, "step": 17957 }, { "epoch": 2.2844421829283807, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4236631393432617, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8826795816421509, "num_tokens": 682004581.0, "step": 17958 }, { "epoch": 2.2845693932069713, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.434542179107666, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.884585976600647, "num_tokens": 682041809.0, "step": 17959 }, { "epoch": 2.284696603485562, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.732224702835083, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.872560977935791, "num_tokens": 682078058.0, "step": 17960 }, { "epoch": 2.2848238137641523, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2810959815979, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8890910148620605, "num_tokens": 682111743.0, "step": 17961 }, { "epoch": 2.2849510240427424, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.8341245651245117, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8746013641357422, "num_tokens": 682147710.0, "step": 17962 }, { "epoch": 2.2850782343213334, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.054479122161865, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8747480511665344, "num_tokens": 682187139.0, "step": 17963 }, { "epoch": 2.2852054445999235, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.77044677734375, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.887902021408081, "num_tokens": 682226345.0, "step": 17964 }, { "epoch": 2.285332654878514, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.879183292388916, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8841569423675537, "num_tokens": 682262479.0, "step": 17965 }, { "epoch": 2.2854598651571045, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6714658737182617, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8762578964233398, "num_tokens": 682300062.0, "step": 17966 }, { "epoch": 2.285587075435695, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7565250396728516, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8834840655326843, "num_tokens": 682338075.0, "step": 17967 }, { "epoch": 2.2857142857142856, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5660881996154785, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8787703514099121, "num_tokens": 682373036.0, "step": 17968 }, { "epoch": 2.285841495992876, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7058982849121094, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8681236505508423, "num_tokens": 682411131.0, "step": 17969 }, { "epoch": 2.2859687062714666, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4730920791625977, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8595178723335266, "num_tokens": 682449142.0, "step": 17970 }, { "epoch": 2.286095916550057, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.689276933670044, "learning_rate": 1e-05, "loss": 0.3734, "mean_token_accuracy": 0.8990991115570068, "num_tokens": 682486369.0, "step": 17971 }, { "epoch": 2.2862231268286477, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3851051330566406, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8823880553245544, "num_tokens": 682519756.0, "step": 17972 }, { "epoch": 2.2863503371072382, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7478601932525635, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8760585784912109, "num_tokens": 682559930.0, "step": 17973 }, { "epoch": 2.2864775473858288, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3051836490631104, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8716610074043274, "num_tokens": 682601379.0, "step": 17974 }, { "epoch": 2.2866047576644193, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.863685131072998, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8799288272857666, "num_tokens": 682639070.0, "step": 17975 }, { "epoch": 2.28673196794301, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2577524185180664, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8865358233451843, "num_tokens": 682675684.0, "step": 17976 }, { "epoch": 2.2868591782216003, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.042954444885254, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.883582592010498, "num_tokens": 682711026.0, "step": 17977 }, { "epoch": 2.286986388500191, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3016722202301025, "learning_rate": 1e-05, "loss": 0.3699, "mean_token_accuracy": 0.8942217826843262, "num_tokens": 682743074.0, "step": 17978 }, { "epoch": 2.2871135987787814, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.197058200836182, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8823704719543457, "num_tokens": 682778875.0, "step": 17979 }, { "epoch": 2.287240809057372, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4733622074127197, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.883711576461792, "num_tokens": 682817699.0, "step": 17980 }, { "epoch": 2.2873680193359625, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.192337989807129, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8701451420783997, "num_tokens": 682852742.0, "step": 17981 }, { "epoch": 2.287495229614553, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6487438678741455, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8742316365242004, "num_tokens": 682894911.0, "step": 17982 }, { "epoch": 2.2876224398931435, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.019135475158691, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8758230209350586, "num_tokens": 682932387.0, "step": 17983 }, { "epoch": 2.287749650171734, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.7002503871917725, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.886737585067749, "num_tokens": 682965072.0, "step": 17984 }, { "epoch": 2.2878768604503246, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8011271953582764, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8845966458320618, "num_tokens": 683006008.0, "step": 17985 }, { "epoch": 2.288004070728915, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6861324310302734, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8821207880973816, "num_tokens": 683042043.0, "step": 17986 }, { "epoch": 2.288131281007505, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.61057710647583, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8764653205871582, "num_tokens": 683079429.0, "step": 17987 }, { "epoch": 2.288258491286096, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6557512283325195, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8737977743148804, "num_tokens": 683115686.0, "step": 17988 }, { "epoch": 2.2883857015646862, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4610602855682373, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8899444341659546, "num_tokens": 683156715.0, "step": 17989 }, { "epoch": 2.2885129118432768, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6516191959381104, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8806637525558472, "num_tokens": 683191758.0, "step": 17990 }, { "epoch": 2.2886401221218673, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.377180814743042, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8801263570785522, "num_tokens": 683225146.0, "step": 17991 }, { "epoch": 2.288767332400458, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.717390298843384, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8890628814697266, "num_tokens": 683259552.0, "step": 17992 }, { "epoch": 2.2888945426790483, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2959647178649902, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8910083770751953, "num_tokens": 683296925.0, "step": 17993 }, { "epoch": 2.289021752957639, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.780740261077881, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8814317584037781, "num_tokens": 683328836.0, "step": 17994 }, { "epoch": 2.2891489632362294, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2423269748687744, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8868429064750671, "num_tokens": 683359306.0, "step": 17995 }, { "epoch": 2.28927617351482, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.911402940750122, "learning_rate": 1e-05, "loss": 0.3773, "mean_token_accuracy": 0.8973204493522644, "num_tokens": 683395743.0, "step": 17996 }, { "epoch": 2.2894033837934105, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.211060047149658, "learning_rate": 1e-05, "loss": 0.407, "mean_token_accuracy": 0.8848752379417419, "num_tokens": 683434528.0, "step": 17997 }, { "epoch": 2.289530594072001, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.087818145751953, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8750301599502563, "num_tokens": 683470540.0, "step": 17998 }, { "epoch": 2.2896578043505915, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3113324642181396, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8769455552101135, "num_tokens": 683502805.0, "step": 17999 }, { "epoch": 2.289785014629182, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.219582557678223, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8814160823822021, "num_tokens": 683540306.0, "step": 18000 }, { "epoch": 2.2899122249077726, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.486088514328003, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8797163963317871, "num_tokens": 683573089.0, "step": 18001 }, { "epoch": 2.290039435186363, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.202924728393555, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8741894960403442, "num_tokens": 683613540.0, "step": 18002 }, { "epoch": 2.2901666454649536, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6218008995056152, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8752639889717102, "num_tokens": 683655016.0, "step": 18003 }, { "epoch": 2.290293855743544, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.061986923217773, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8906829953193665, "num_tokens": 683692607.0, "step": 18004 }, { "epoch": 2.2904210660221347, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6744577884674072, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8833552598953247, "num_tokens": 683726825.0, "step": 18005 }, { "epoch": 2.290548276300725, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.897195339202881, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8651758432388306, "num_tokens": 683763905.0, "step": 18006 }, { "epoch": 2.2906754865793157, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6403188705444336, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8697223663330078, "num_tokens": 683806926.0, "step": 18007 }, { "epoch": 2.2908026968579063, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7453773021698, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.875613808631897, "num_tokens": 683846069.0, "step": 18008 }, { "epoch": 2.290929907136497, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.579190731048584, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8778708577156067, "num_tokens": 683887530.0, "step": 18009 }, { "epoch": 2.2910571174150873, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.644695281982422, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8751544952392578, "num_tokens": 683925721.0, "step": 18010 }, { "epoch": 2.291184327693678, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.510347366333008, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.883308470249176, "num_tokens": 683967867.0, "step": 18011 }, { "epoch": 2.291311537972268, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6014509201049805, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8771036267280579, "num_tokens": 684006360.0, "step": 18012 }, { "epoch": 2.291438748250859, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4393668174743652, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8857446908950806, "num_tokens": 684047559.0, "step": 18013 }, { "epoch": 2.291565958529449, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6103355884552, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8748973608016968, "num_tokens": 684092028.0, "step": 18014 }, { "epoch": 2.2916931688080395, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3785488605499268, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.871738076210022, "num_tokens": 684131441.0, "step": 18015 }, { "epoch": 2.29182037908663, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.654918670654297, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8831402063369751, "num_tokens": 684163864.0, "step": 18016 }, { "epoch": 2.2919475893652206, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3078315258026123, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.882285475730896, "num_tokens": 684200962.0, "step": 18017 }, { "epoch": 2.292074799643811, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7398738861083984, "learning_rate": 1e-05, "loss": 0.3969, "mean_token_accuracy": 0.8893043994903564, "num_tokens": 684237060.0, "step": 18018 }, { "epoch": 2.2922020099224016, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.25821590423584, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8830457925796509, "num_tokens": 684271299.0, "step": 18019 }, { "epoch": 2.292329220200992, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.870424509048462, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8718985319137573, "num_tokens": 684310008.0, "step": 18020 }, { "epoch": 2.2924564304795827, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.232604742050171, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.880544900894165, "num_tokens": 684348477.0, "step": 18021 }, { "epoch": 2.2925836407581732, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0483551025390625, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8708209991455078, "num_tokens": 684383090.0, "step": 18022 }, { "epoch": 2.2927108510367638, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.308988571166992, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8636599779129028, "num_tokens": 684415142.0, "step": 18023 }, { "epoch": 2.2928380613153543, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.159140586853027, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8807690143585205, "num_tokens": 684455878.0, "step": 18024 }, { "epoch": 2.292965271593945, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.4416630268096924, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8735024929046631, "num_tokens": 684501140.0, "step": 18025 }, { "epoch": 2.2930924818725353, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1586127281188965, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8753666281700134, "num_tokens": 684542682.0, "step": 18026 }, { "epoch": 2.293219692151126, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5974113941192627, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.871423065662384, "num_tokens": 684581301.0, "step": 18027 }, { "epoch": 2.2933469024297164, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0108466148376465, "learning_rate": 1e-05, "loss": 0.4071, "mean_token_accuracy": 0.8885191082954407, "num_tokens": 684620124.0, "step": 18028 }, { "epoch": 2.293474112708307, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6519930362701416, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.873096227645874, "num_tokens": 684657237.0, "step": 18029 }, { "epoch": 2.2936013229868975, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8204524517059326, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8857651352882385, "num_tokens": 684696717.0, "step": 18030 }, { "epoch": 2.293728533265488, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.633592367172241, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8793632984161377, "num_tokens": 684734823.0, "step": 18031 }, { "epoch": 2.2938557435440785, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6589720249176025, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8767051100730896, "num_tokens": 684773036.0, "step": 18032 }, { "epoch": 2.293982953822669, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5958473682403564, "learning_rate": 1e-05, "loss": 0.4102, "mean_token_accuracy": 0.8842820525169373, "num_tokens": 684811210.0, "step": 18033 }, { "epoch": 2.2941101641012596, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5365958213806152, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8801310658454895, "num_tokens": 684850656.0, "step": 18034 }, { "epoch": 2.2942373743798496, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.567283868789673, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8778300881385803, "num_tokens": 684890113.0, "step": 18035 }, { "epoch": 2.2943645846584406, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.455967664718628, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8858407735824585, "num_tokens": 684930418.0, "step": 18036 }, { "epoch": 2.2944917949370307, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.560220241546631, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8797844648361206, "num_tokens": 684967068.0, "step": 18037 }, { "epoch": 2.2946190052156212, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4020473957061768, "learning_rate": 1e-05, "loss": 0.4945, "mean_token_accuracy": 0.8572459816932678, "num_tokens": 685007602.0, "step": 18038 }, { "epoch": 2.2947462154942118, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.57391619682312, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8825386166572571, "num_tokens": 685043332.0, "step": 18039 }, { "epoch": 2.2948734257728023, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3430638313293457, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8843801617622375, "num_tokens": 685078327.0, "step": 18040 }, { "epoch": 2.295000636051393, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.617323398590088, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.880301833152771, "num_tokens": 685119065.0, "step": 18041 }, { "epoch": 2.2951278463299833, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.308505058288574, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.8557025194168091, "num_tokens": 685154988.0, "step": 18042 }, { "epoch": 2.295255056608574, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6828629970550537, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8768332004547119, "num_tokens": 685194739.0, "step": 18043 }, { "epoch": 2.2953822668871644, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2239155769348145, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8665822744369507, "num_tokens": 685235619.0, "step": 18044 }, { "epoch": 2.295509477165755, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.807987689971924, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8761317133903503, "num_tokens": 685275189.0, "step": 18045 }, { "epoch": 2.2956366874443455, "ewc_loss": 0.07763671875, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1660947799682617, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8731136918067932, "num_tokens": 685319928.0, "step": 18046 }, { "epoch": 2.295763897722936, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000308990478515625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.016213417053223, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8871859312057495, "num_tokens": 685355698.0, "step": 18047 }, { "epoch": 2.2958911080015265, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1830334663391113, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8806185126304626, "num_tokens": 685394873.0, "step": 18048 }, { "epoch": 2.296018318280117, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2634596824646, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8842594027519226, "num_tokens": 685433896.0, "step": 18049 }, { "epoch": 2.2961455285587076, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.3683929443359375, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8762946128845215, "num_tokens": 685476036.0, "step": 18050 }, { "epoch": 2.296272738837298, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.388612270355225, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8820019960403442, "num_tokens": 685514020.0, "step": 18051 }, { "epoch": 2.2963999491158886, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6654210090637207, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8871942162513733, "num_tokens": 685547644.0, "step": 18052 }, { "epoch": 2.296527159394479, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.277675628662109, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8725736737251282, "num_tokens": 685584147.0, "step": 18053 }, { "epoch": 2.2966543696730697, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.7990405559539795, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8804633617401123, "num_tokens": 685623093.0, "step": 18054 }, { "epoch": 2.29678157995166, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.058359622955322, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8873838782310486, "num_tokens": 685657837.0, "step": 18055 }, { "epoch": 2.2969087902302507, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.766176700592041, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8781485557556152, "num_tokens": 685691077.0, "step": 18056 }, { "epoch": 2.2970360005088413, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.871046781539917, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8746703863143921, "num_tokens": 685725011.0, "step": 18057 }, { "epoch": 2.297163210787432, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6756913661956787, "learning_rate": 1e-05, "loss": 0.4867, "mean_token_accuracy": 0.862949550151825, "num_tokens": 685764212.0, "step": 18058 }, { "epoch": 2.2972904210660223, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7409915924072266, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8597327470779419, "num_tokens": 685800780.0, "step": 18059 }, { "epoch": 2.2974176313446124, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.576174736022949, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8779576420783997, "num_tokens": 685837728.0, "step": 18060 }, { "epoch": 2.2975448416232034, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.696812868118286, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8622602224349976, "num_tokens": 685869349.0, "step": 18061 }, { "epoch": 2.2976720519017935, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4653091430664062, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.8866212964057922, "num_tokens": 685908114.0, "step": 18062 }, { "epoch": 2.297799262180384, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6831893920898438, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8691524267196655, "num_tokens": 685944980.0, "step": 18063 }, { "epoch": 2.2979264724589745, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.37650465965271, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8847293257713318, "num_tokens": 685982729.0, "step": 18064 }, { "epoch": 2.298053682737565, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.740154504776001, "learning_rate": 1e-05, "loss": 0.3817, "mean_token_accuracy": 0.8956270813941956, "num_tokens": 686014090.0, "step": 18065 }, { "epoch": 2.2981808930161556, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3039839267730713, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8845672607421875, "num_tokens": 686047345.0, "step": 18066 }, { "epoch": 2.298308103294746, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.845745325088501, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8722209334373474, "num_tokens": 686088662.0, "step": 18067 }, { "epoch": 2.2984353135733366, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2671854496002197, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.86921226978302, "num_tokens": 686124210.0, "step": 18068 }, { "epoch": 2.298562523851927, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0092339515686035, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8795171976089478, "num_tokens": 686157759.0, "step": 18069 }, { "epoch": 2.2986897341305177, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.290652275085449, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.873785674571991, "num_tokens": 686198486.0, "step": 18070 }, { "epoch": 2.298816944409108, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.162893772125244, "learning_rate": 1e-05, "loss": 0.4951, "mean_token_accuracy": 0.8593641519546509, "num_tokens": 686235859.0, "step": 18071 }, { "epoch": 2.2989441546876987, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.435516834259033, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8853079080581665, "num_tokens": 686274185.0, "step": 18072 }, { "epoch": 2.2990713649662893, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.179459571838379, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.8923047780990601, "num_tokens": 686308706.0, "step": 18073 }, { "epoch": 2.29919857524488, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6022226810455322, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8891448378562927, "num_tokens": 686347871.0, "step": 18074 }, { "epoch": 2.2993257855234703, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.042013645172119, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8683012127876282, "num_tokens": 686395284.0, "step": 18075 }, { "epoch": 2.299452995802061, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6791563034057617, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8704115152359009, "num_tokens": 686436566.0, "step": 18076 }, { "epoch": 2.2995802060806514, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.831713914871216, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8858986496925354, "num_tokens": 686477332.0, "step": 18077 }, { "epoch": 2.299707416359242, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6690759658813477, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.892675518989563, "num_tokens": 686512000.0, "step": 18078 }, { "epoch": 2.2998346266378324, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.653945207595825, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8818924427032471, "num_tokens": 686544732.0, "step": 18079 }, { "epoch": 2.299961836916423, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6427457332611084, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.879015326499939, "num_tokens": 686583100.0, "step": 18080 }, { "epoch": 2.3000890471950135, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5075013637542725, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8781329393386841, "num_tokens": 686620824.0, "step": 18081 }, { "epoch": 2.300216257473604, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.618785858154297, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8854537010192871, "num_tokens": 686660410.0, "step": 18082 }, { "epoch": 2.3003434677521946, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.412869453430176, "learning_rate": 1e-05, "loss": 0.3865, "mean_token_accuracy": 0.8898338079452515, "num_tokens": 686694385.0, "step": 18083 }, { "epoch": 2.300470678030785, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6300716400146484, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8827944993972778, "num_tokens": 686736844.0, "step": 18084 }, { "epoch": 2.300597888309375, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.329488515853882, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.872714638710022, "num_tokens": 686781104.0, "step": 18085 }, { "epoch": 2.300725098587966, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6898887157440186, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.8920191526412964, "num_tokens": 686815244.0, "step": 18086 }, { "epoch": 2.3008523088665562, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.263542652130127, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8791066408157349, "num_tokens": 686858133.0, "step": 18087 }, { "epoch": 2.3009795191451468, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7895452976226807, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8746880292892456, "num_tokens": 686901250.0, "step": 18088 }, { "epoch": 2.3011067294237373, "ewc_loss": 0.078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.208050489425659, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.886594295501709, "num_tokens": 686936408.0, "step": 18089 }, { "epoch": 2.301233939702328, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9528117179870605, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8790740966796875, "num_tokens": 686981924.0, "step": 18090 }, { "epoch": 2.3013611499809183, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2214040756225586, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8760462999343872, "num_tokens": 687016052.0, "step": 18091 }, { "epoch": 2.301488360259509, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.13533878326416, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8853176236152649, "num_tokens": 687059107.0, "step": 18092 }, { "epoch": 2.3016155705380994, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.303208351135254, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8849633932113647, "num_tokens": 687104015.0, "step": 18093 }, { "epoch": 2.30174278081669, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.293691635131836, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8731697797775269, "num_tokens": 687143719.0, "step": 18094 }, { "epoch": 2.3018699910952805, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5436956882476807, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8821588754653931, "num_tokens": 687179732.0, "step": 18095 }, { "epoch": 2.301997201373871, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.24511194229126, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8741874694824219, "num_tokens": 687222309.0, "step": 18096 }, { "epoch": 2.3021244116524615, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.702183723449707, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8876560926437378, "num_tokens": 687259114.0, "step": 18097 }, { "epoch": 2.302251621931052, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0680646896362305, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8852071166038513, "num_tokens": 687293052.0, "step": 18098 }, { "epoch": 2.3023788322096426, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7077531814575195, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8863504528999329, "num_tokens": 687333245.0, "step": 18099 }, { "epoch": 2.302506042488233, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.882132053375244, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8764397501945496, "num_tokens": 687371740.0, "step": 18100 }, { "epoch": 2.3026332527668236, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6745963096618652, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8574913740158081, "num_tokens": 687407298.0, "step": 18101 }, { "epoch": 2.302760463045414, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7104721069335938, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8852171897888184, "num_tokens": 687447425.0, "step": 18102 }, { "epoch": 2.3028876733240047, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6000399589538574, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.880164384841919, "num_tokens": 687488528.0, "step": 18103 }, { "epoch": 2.303014883602595, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.627296209335327, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8771743774414062, "num_tokens": 687521086.0, "step": 18104 }, { "epoch": 2.3031420938811857, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.533953905105591, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8736696243286133, "num_tokens": 687558882.0, "step": 18105 }, { "epoch": 2.3032693041597763, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5662896633148193, "learning_rate": 1e-05, "loss": 0.3886, "mean_token_accuracy": 0.8918810486793518, "num_tokens": 687592058.0, "step": 18106 }, { "epoch": 2.303396514438367, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4708609580993652, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8779191970825195, "num_tokens": 687629715.0, "step": 18107 }, { "epoch": 2.3035237247169573, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5553529262542725, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8762862682342529, "num_tokens": 687666201.0, "step": 18108 }, { "epoch": 2.303650934995548, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.406073808670044, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8846042156219482, "num_tokens": 687706897.0, "step": 18109 }, { "epoch": 2.303778145274138, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5793728828430176, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8802071809768677, "num_tokens": 687741449.0, "step": 18110 }, { "epoch": 2.303905355552729, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3537395000457764, "learning_rate": 1e-05, "loss": 0.3949, "mean_token_accuracy": 0.8917306661605835, "num_tokens": 687776758.0, "step": 18111 }, { "epoch": 2.304032565831319, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6279711723327637, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8643412590026855, "num_tokens": 687814636.0, "step": 18112 }, { "epoch": 2.3041597761099095, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3147311210632324, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8610715866088867, "num_tokens": 687843685.0, "step": 18113 }, { "epoch": 2.3042869863885, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7103769779205322, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8870863318443298, "num_tokens": 687875040.0, "step": 18114 }, { "epoch": 2.3044141966670906, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.221719264984131, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8737878203392029, "num_tokens": 687912952.0, "step": 18115 }, { "epoch": 2.304541406945681, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8498942852020264, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8815675973892212, "num_tokens": 687951366.0, "step": 18116 }, { "epoch": 2.3046686172242716, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.167344093322754, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8805388808250427, "num_tokens": 687993593.0, "step": 18117 }, { "epoch": 2.304795827502862, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.080747604370117, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8750290870666504, "num_tokens": 688033449.0, "step": 18118 }, { "epoch": 2.3049230377814527, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.586505651473999, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8766963481903076, "num_tokens": 688073239.0, "step": 18119 }, { "epoch": 2.305050248060043, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.94931960105896, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8748492002487183, "num_tokens": 688112760.0, "step": 18120 }, { "epoch": 2.3051774583386337, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3244550228118896, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8705146312713623, "num_tokens": 688150553.0, "step": 18121 }, { "epoch": 2.3053046686172243, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.97621488571167, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8757854700088501, "num_tokens": 688190582.0, "step": 18122 }, { "epoch": 2.305431878895815, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3687126636505127, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8781934976577759, "num_tokens": 688226512.0, "step": 18123 }, { "epoch": 2.3055590891744053, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.014217376708984, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8792611360549927, "num_tokens": 688264891.0, "step": 18124 }, { "epoch": 2.305686299452996, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.424760341644287, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8768616318702698, "num_tokens": 688301280.0, "step": 18125 }, { "epoch": 2.3058135097315864, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.9947543144226074, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8810845017433167, "num_tokens": 688337178.0, "step": 18126 }, { "epoch": 2.305940720010177, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.485408306121826, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8867383003234863, "num_tokens": 688375839.0, "step": 18127 }, { "epoch": 2.3060679302887674, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9298946857452393, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8748936057090759, "num_tokens": 688416395.0, "step": 18128 }, { "epoch": 2.306195140567358, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.5035207271575928, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.889923095703125, "num_tokens": 688454801.0, "step": 18129 }, { "epoch": 2.3063223508459485, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8517184257507324, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8787401914596558, "num_tokens": 688489265.0, "step": 18130 }, { "epoch": 2.306449561124539, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5276427268981934, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8729556798934937, "num_tokens": 688525392.0, "step": 18131 }, { "epoch": 2.3065767714031296, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.748427391052246, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8628669381141663, "num_tokens": 688562389.0, "step": 18132 }, { "epoch": 2.3067039816817196, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5137133598327637, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8803826570510864, "num_tokens": 688597728.0, "step": 18133 }, { "epoch": 2.3068311919603106, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6551613807678223, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.876201331615448, "num_tokens": 688633065.0, "step": 18134 }, { "epoch": 2.3069584022389007, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.4751384258270264, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8800547122955322, "num_tokens": 688677095.0, "step": 18135 }, { "epoch": 2.3070856125174912, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5967531204223633, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8832395672798157, "num_tokens": 688710421.0, "step": 18136 }, { "epoch": 2.3072128227960818, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4599359035491943, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8790141344070435, "num_tokens": 688747300.0, "step": 18137 }, { "epoch": 2.3073400330746723, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5527842044830322, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.887177586555481, "num_tokens": 688781867.0, "step": 18138 }, { "epoch": 2.307467243353263, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.426133394241333, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8807837963104248, "num_tokens": 688822777.0, "step": 18139 }, { "epoch": 2.3075944536318533, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5479655265808105, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8710049390792847, "num_tokens": 688862295.0, "step": 18140 }, { "epoch": 2.307721663910444, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3844237327575684, "learning_rate": 1e-05, "loss": 0.383, "mean_token_accuracy": 0.893143892288208, "num_tokens": 688904814.0, "step": 18141 }, { "epoch": 2.3078488741890344, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5696513652801514, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8700166940689087, "num_tokens": 688940403.0, "step": 18142 }, { "epoch": 2.307976084467625, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3512609004974365, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8609068393707275, "num_tokens": 688980256.0, "step": 18143 }, { "epoch": 2.3081032947462155, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.604856014251709, "learning_rate": 1e-05, "loss": 0.4861, "mean_token_accuracy": 0.8611676692962646, "num_tokens": 689020629.0, "step": 18144 }, { "epoch": 2.308230505024806, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.335193157196045, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8643300533294678, "num_tokens": 689051257.0, "step": 18145 }, { "epoch": 2.3083577153033965, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.650259256362915, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.882292628288269, "num_tokens": 689083723.0, "step": 18146 }, { "epoch": 2.308484925581987, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2682249546051025, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8611843585968018, "num_tokens": 689123785.0, "step": 18147 }, { "epoch": 2.3086121358605776, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.737161874771118, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8756444454193115, "num_tokens": 689164053.0, "step": 18148 }, { "epoch": 2.308739346139168, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2203240394592285, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8827961683273315, "num_tokens": 689198221.0, "step": 18149 }, { "epoch": 2.3088665564177586, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.906521797180176, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8770767450332642, "num_tokens": 689232330.0, "step": 18150 }, { "epoch": 2.308993766696349, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.191739797592163, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8692096471786499, "num_tokens": 689277663.0, "step": 18151 }, { "epoch": 2.3091209769749397, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.103060722351074, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8867078423500061, "num_tokens": 689308563.0, "step": 18152 }, { "epoch": 2.30924818725353, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.261883497238159, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8772103190422058, "num_tokens": 689346897.0, "step": 18153 }, { "epoch": 2.3093753975321207, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.282829284667969, "learning_rate": 1e-05, "loss": 0.4797, "mean_token_accuracy": 0.8677825927734375, "num_tokens": 689380689.0, "step": 18154 }, { "epoch": 2.3095026078107113, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.48681640625, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8788416385650635, "num_tokens": 689418587.0, "step": 18155 }, { "epoch": 2.309629818089302, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.2768425941467285, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8826950788497925, "num_tokens": 689455742.0, "step": 18156 }, { "epoch": 2.3097570283678923, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6601574420928955, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8783137202262878, "num_tokens": 689495618.0, "step": 18157 }, { "epoch": 2.3098842386464824, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.151294708251953, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8600118160247803, "num_tokens": 689529856.0, "step": 18158 }, { "epoch": 2.3100114489250734, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00031280517578125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7210404872894287, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.874542236328125, "num_tokens": 689566785.0, "step": 18159 }, { "epoch": 2.3101386592036635, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.9612817764282227, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8748600482940674, "num_tokens": 689604332.0, "step": 18160 }, { "epoch": 2.310265869482254, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.661552667617798, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8867663741111755, "num_tokens": 689639159.0, "step": 18161 }, { "epoch": 2.3103930797608445, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.817201614379883, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8714393377304077, "num_tokens": 689681660.0, "step": 18162 }, { "epoch": 2.310520290039435, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.585753917694092, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.8882623314857483, "num_tokens": 689717070.0, "step": 18163 }, { "epoch": 2.3106475003180256, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7136878967285156, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8814795613288879, "num_tokens": 689758885.0, "step": 18164 }, { "epoch": 2.310774710596616, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.496206045150757, "learning_rate": 1e-05, "loss": 0.3818, "mean_token_accuracy": 0.8936699628829956, "num_tokens": 689794943.0, "step": 18165 }, { "epoch": 2.3109019208752066, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.680257797241211, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8818297386169434, "num_tokens": 689836943.0, "step": 18166 }, { "epoch": 2.311029131153797, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4320504665374756, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8812891244888306, "num_tokens": 689870787.0, "step": 18167 }, { "epoch": 2.3111563414323877, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.689563274383545, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8770074844360352, "num_tokens": 689908302.0, "step": 18168 }, { "epoch": 2.311283551710978, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.385530710220337, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8710533976554871, "num_tokens": 689950031.0, "step": 18169 }, { "epoch": 2.3114107619895687, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7218551635742188, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8841410875320435, "num_tokens": 689994674.0, "step": 18170 }, { "epoch": 2.3115379722681593, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2906582355499268, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8843929767608643, "num_tokens": 690038697.0, "step": 18171 }, { "epoch": 2.31166518254675, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8296971321105957, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8731886148452759, "num_tokens": 690077449.0, "step": 18172 }, { "epoch": 2.3117923928253403, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2666006088256836, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8740941286087036, "num_tokens": 690113659.0, "step": 18173 }, { "epoch": 2.311919603103931, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9563748836517334, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8730411529541016, "num_tokens": 690153928.0, "step": 18174 }, { "epoch": 2.3120468133825214, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.285439968109131, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8813538551330566, "num_tokens": 690189526.0, "step": 18175 }, { "epoch": 2.312174023661112, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.088079929351807, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.8743628859519958, "num_tokens": 690230220.0, "step": 18176 }, { "epoch": 2.3123012339397024, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.3996362686157227, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8768589496612549, "num_tokens": 690264998.0, "step": 18177 }, { "epoch": 2.312428444218293, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1111907958984375, "learning_rate": 1e-05, "loss": 0.3742, "mean_token_accuracy": 0.9044991135597229, "num_tokens": 690304323.0, "step": 18178 }, { "epoch": 2.3125556544968835, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.523630142211914, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8775533437728882, "num_tokens": 690346832.0, "step": 18179 }, { "epoch": 2.312682864775474, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.022361755371094, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8824886679649353, "num_tokens": 690385829.0, "step": 18180 }, { "epoch": 2.3128100750540646, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.596541166305542, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8819891214370728, "num_tokens": 690429949.0, "step": 18181 }, { "epoch": 2.312937285332655, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8467180728912354, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8819396495819092, "num_tokens": 690473252.0, "step": 18182 }, { "epoch": 2.313064495611245, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.613476037979126, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8848934173583984, "num_tokens": 690508765.0, "step": 18183 }, { "epoch": 2.313191705889836, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6755762100219727, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8794256448745728, "num_tokens": 690546749.0, "step": 18184 }, { "epoch": 2.313318916168426, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.615111827850342, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8787815570831299, "num_tokens": 690579808.0, "step": 18185 }, { "epoch": 2.3134461264470167, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5384836196899414, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8778701424598694, "num_tokens": 690611667.0, "step": 18186 }, { "epoch": 2.3135733367256073, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5921237468719482, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8710854649543762, "num_tokens": 690654991.0, "step": 18187 }, { "epoch": 2.313700547004198, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4417285919189453, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8640153408050537, "num_tokens": 690692344.0, "step": 18188 }, { "epoch": 2.3138277572827883, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5930728912353516, "learning_rate": 1e-05, "loss": 0.3962, "mean_token_accuracy": 0.8890958428382874, "num_tokens": 690726812.0, "step": 18189 }, { "epoch": 2.313954967561379, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3639891147613525, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8754351139068604, "num_tokens": 690768229.0, "step": 18190 }, { "epoch": 2.3140821778399694, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.61944842338562, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.881292998790741, "num_tokens": 690810614.0, "step": 18191 }, { "epoch": 2.31420938811856, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.304032802581787, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8657269477844238, "num_tokens": 690855422.0, "step": 18192 }, { "epoch": 2.3143365983971504, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.695910930633545, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8870731592178345, "num_tokens": 690893751.0, "step": 18193 }, { "epoch": 2.314463808675741, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2271270751953125, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8806039094924927, "num_tokens": 690936734.0, "step": 18194 }, { "epoch": 2.3145910189543315, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8358547687530518, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8843823671340942, "num_tokens": 690975005.0, "step": 18195 }, { "epoch": 2.314718229232922, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.18542218208313, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8827633261680603, "num_tokens": 691011607.0, "step": 18196 }, { "epoch": 2.3148454395115126, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.03234338760376, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8834697008132935, "num_tokens": 691047635.0, "step": 18197 }, { "epoch": 2.314972649790103, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.1969738006591797, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8727838397026062, "num_tokens": 691095290.0, "step": 18198 }, { "epoch": 2.3150998600686936, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.258822917938232, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8792842626571655, "num_tokens": 691140641.0, "step": 18199 }, { "epoch": 2.315227070347284, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.394749164581299, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8787552118301392, "num_tokens": 691181744.0, "step": 18200 }, { "epoch": 2.3153542806258747, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.354614734649658, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.880400538444519, "num_tokens": 691221279.0, "step": 18201 }, { "epoch": 2.315481490904465, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.662799835205078, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8843072652816772, "num_tokens": 691258519.0, "step": 18202 }, { "epoch": 2.3156087011830557, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.220699310302734, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8755931258201599, "num_tokens": 691297063.0, "step": 18203 }, { "epoch": 2.3157359114616463, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7601637840270996, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8800280094146729, "num_tokens": 691337146.0, "step": 18204 }, { "epoch": 2.315863121740237, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.013427257537842, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.885650634765625, "num_tokens": 691373475.0, "step": 18205 }, { "epoch": 2.3159903320188273, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7212977409362793, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.87702876329422, "num_tokens": 691413286.0, "step": 18206 }, { "epoch": 2.316117542297418, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.840013265609741, "learning_rate": 1e-05, "loss": 0.3937, "mean_token_accuracy": 0.8924673795700073, "num_tokens": 691444950.0, "step": 18207 }, { "epoch": 2.316244752576008, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6415956020355225, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8759030103683472, "num_tokens": 691486767.0, "step": 18208 }, { "epoch": 2.316371962854599, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7036855220794678, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8743272423744202, "num_tokens": 691527149.0, "step": 18209 }, { "epoch": 2.316499173133189, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.566610097885132, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8694984912872314, "num_tokens": 691563963.0, "step": 18210 }, { "epoch": 2.3166263834117795, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.650702476501465, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8726156949996948, "num_tokens": 691596423.0, "step": 18211 }, { "epoch": 2.31675359369037, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.475890874862671, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8764454126358032, "num_tokens": 691636928.0, "step": 18212 }, { "epoch": 2.3168808039689606, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6282520294189453, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8774537444114685, "num_tokens": 691673427.0, "step": 18213 }, { "epoch": 2.317008014247551, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4078567028045654, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.87410569190979, "num_tokens": 691710427.0, "step": 18214 }, { "epoch": 2.3171352245261416, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6631109714508057, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.869722843170166, "num_tokens": 691746809.0, "step": 18215 }, { "epoch": 2.317262434804732, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.33107590675354, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8831135034561157, "num_tokens": 691786037.0, "step": 18216 }, { "epoch": 2.3173896450833227, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7369887828826904, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.874731183052063, "num_tokens": 691822270.0, "step": 18217 }, { "epoch": 2.317516855361913, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.25368332862854, "learning_rate": 1e-05, "loss": 0.3675, "mean_token_accuracy": 0.8986725807189941, "num_tokens": 691858538.0, "step": 18218 }, { "epoch": 2.3176440656405037, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.879098653793335, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8838589787483215, "num_tokens": 691898810.0, "step": 18219 }, { "epoch": 2.3177712759190943, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2286789417266846, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8809474110603333, "num_tokens": 691933708.0, "step": 18220 }, { "epoch": 2.317898486197685, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.067655563354492, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8917502760887146, "num_tokens": 691968293.0, "step": 18221 }, { "epoch": 2.3180256964762753, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.296417713165283, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8684893846511841, "num_tokens": 692004532.0, "step": 18222 }, { "epoch": 2.318152906754866, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.222139358520508, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8807998895645142, "num_tokens": 692042323.0, "step": 18223 }, { "epoch": 2.3182801170334564, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.4857635498046875, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8761762380599976, "num_tokens": 692080970.0, "step": 18224 }, { "epoch": 2.318407327312047, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.197066783905029, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8839993476867676, "num_tokens": 692115027.0, "step": 18225 }, { "epoch": 2.3185345375906374, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6632113456726074, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8770836591720581, "num_tokens": 692146602.0, "step": 18226 }, { "epoch": 2.318661747869228, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.010812282562256, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8625689744949341, "num_tokens": 692189721.0, "step": 18227 }, { "epoch": 2.3187889581478185, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6759164333343506, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.871448814868927, "num_tokens": 692233032.0, "step": 18228 }, { "epoch": 2.318916168426409, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8270211219787598, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8706017732620239, "num_tokens": 692269033.0, "step": 18229 }, { "epoch": 2.3190433787049995, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.646172046661377, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8761261701583862, "num_tokens": 692307415.0, "step": 18230 }, { "epoch": 2.3191705889835896, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6573615074157715, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8849198818206787, "num_tokens": 692347003.0, "step": 18231 }, { "epoch": 2.3192977992621806, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.605081796646118, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8779768347740173, "num_tokens": 692385141.0, "step": 18232 }, { "epoch": 2.3194250095407707, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5366570949554443, "learning_rate": 1e-05, "loss": 0.3853, "mean_token_accuracy": 0.8923463821411133, "num_tokens": 692421731.0, "step": 18233 }, { "epoch": 2.319552219819361, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5802531242370605, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8779416084289551, "num_tokens": 692466400.0, "step": 18234 }, { "epoch": 2.3196794300979517, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.449390172958374, "learning_rate": 1e-05, "loss": 0.3751, "mean_token_accuracy": 0.8947693109512329, "num_tokens": 692501201.0, "step": 18235 }, { "epoch": 2.3198066403765423, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.563157558441162, "learning_rate": 1e-05, "loss": 0.3942, "mean_token_accuracy": 0.8914719223976135, "num_tokens": 692541549.0, "step": 18236 }, { "epoch": 2.319933850655133, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.401721239089966, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8836966156959534, "num_tokens": 692582156.0, "step": 18237 }, { "epoch": 2.3200610609337233, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5807547569274902, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8656553030014038, "num_tokens": 692620522.0, "step": 18238 }, { "epoch": 2.320188271212314, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.352867603302002, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8744758367538452, "num_tokens": 692658372.0, "step": 18239 }, { "epoch": 2.3203154814909044, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.626441240310669, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8827013373374939, "num_tokens": 692687338.0, "step": 18240 }, { "epoch": 2.320442691769495, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.286562204360962, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8766416311264038, "num_tokens": 692724417.0, "step": 18241 }, { "epoch": 2.3205699020480854, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7045445442199707, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8785025477409363, "num_tokens": 692767881.0, "step": 18242 }, { "epoch": 2.320697112326676, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.222001791000366, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8743285536766052, "num_tokens": 692803154.0, "step": 18243 }, { "epoch": 2.3208243226052665, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.843568801879883, "learning_rate": 1e-05, "loss": 0.3876, "mean_token_accuracy": 0.8957598209381104, "num_tokens": 692842435.0, "step": 18244 }, { "epoch": 2.320951532883857, "ewc_loss": 0.078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1628880500793457, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8709836006164551, "num_tokens": 692884491.0, "step": 18245 }, { "epoch": 2.3210787431624476, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.073136806488037, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8734632134437561, "num_tokens": 692920825.0, "step": 18246 }, { "epoch": 2.321205953441038, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2098441123962402, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8842689990997314, "num_tokens": 692956413.0, "step": 18247 }, { "epoch": 2.3213331637196286, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.321520805358887, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8703957200050354, "num_tokens": 692995860.0, "step": 18248 }, { "epoch": 2.321460373998219, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.4377355575561523, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8968119025230408, "num_tokens": 693031077.0, "step": 18249 }, { "epoch": 2.3215875842768097, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.402803421020508, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8654104471206665, "num_tokens": 693067181.0, "step": 18250 }, { "epoch": 2.3217147945554, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.743710994720459, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.882217288017273, "num_tokens": 693103919.0, "step": 18251 }, { "epoch": 2.3218420048339907, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.227873802185059, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.863246500492096, "num_tokens": 693138471.0, "step": 18252 }, { "epoch": 2.3219692151125813, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.804475784301758, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8699966669082642, "num_tokens": 693174220.0, "step": 18253 }, { "epoch": 2.322096425391172, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.998821496963501, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8880882859230042, "num_tokens": 693217222.0, "step": 18254 }, { "epoch": 2.3222236356697623, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.726621627807617, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.892317533493042, "num_tokens": 693260052.0, "step": 18255 }, { "epoch": 2.3223508459483524, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8470535278320312, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8684638738632202, "num_tokens": 693293775.0, "step": 18256 }, { "epoch": 2.3224780562269434, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6329309940338135, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8831923007965088, "num_tokens": 693335876.0, "step": 18257 }, { "epoch": 2.3226052665055335, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7259650230407715, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8689432144165039, "num_tokens": 693376886.0, "step": 18258 }, { "epoch": 2.322732476784124, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.558631181716919, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8687529563903809, "num_tokens": 693416386.0, "step": 18259 }, { "epoch": 2.3228596870627145, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.663259506225586, "learning_rate": 1e-05, "loss": 0.3968, "mean_token_accuracy": 0.8898875713348389, "num_tokens": 693454926.0, "step": 18260 }, { "epoch": 2.322986897341305, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.469881057739258, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8777235746383667, "num_tokens": 693492926.0, "step": 18261 }, { "epoch": 2.3231141076198956, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6467013359069824, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.880764365196228, "num_tokens": 693534684.0, "step": 18262 }, { "epoch": 2.323241317898486, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.392972946166992, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8659883737564087, "num_tokens": 693570856.0, "step": 18263 }, { "epoch": 2.3233685281770766, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6842479705810547, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8689010143280029, "num_tokens": 693613286.0, "step": 18264 }, { "epoch": 2.323495738455667, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.318845272064209, "learning_rate": 1e-05, "loss": 0.3962, "mean_token_accuracy": 0.8854355812072754, "num_tokens": 693648748.0, "step": 18265 }, { "epoch": 2.3236229487342577, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.769548177719116, "learning_rate": 1e-05, "loss": 0.3903, "mean_token_accuracy": 0.8943346738815308, "num_tokens": 693688997.0, "step": 18266 }, { "epoch": 2.323750159012848, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2526907920837402, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8755015134811401, "num_tokens": 693727653.0, "step": 18267 }, { "epoch": 2.3238773692914387, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9186770915985107, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8770223259925842, "num_tokens": 693771146.0, "step": 18268 }, { "epoch": 2.3240045795700293, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2511847019195557, "learning_rate": 1e-05, "loss": 0.5086, "mean_token_accuracy": 0.8524658679962158, "num_tokens": 693809135.0, "step": 18269 }, { "epoch": 2.32413178984862, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.084527015686035, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8870704174041748, "num_tokens": 693847127.0, "step": 18270 }, { "epoch": 2.3242590001272103, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3757572174072266, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8816815614700317, "num_tokens": 693887186.0, "step": 18271 }, { "epoch": 2.324386210405801, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.150747776031494, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8863058090209961, "num_tokens": 693925943.0, "step": 18272 }, { "epoch": 2.3245134206843914, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4755117893218994, "learning_rate": 1e-05, "loss": 0.3797, "mean_token_accuracy": 0.8952915668487549, "num_tokens": 693958825.0, "step": 18273 }, { "epoch": 2.324640630962982, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.112935543060303, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8672361373901367, "num_tokens": 694003787.0, "step": 18274 }, { "epoch": 2.3247678412415724, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.595963478088379, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8797634243965149, "num_tokens": 694047253.0, "step": 18275 }, { "epoch": 2.324895051520163, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.9802894592285156, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8736851215362549, "num_tokens": 694085913.0, "step": 18276 }, { "epoch": 2.3250222617987535, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6509664058685303, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8648656606674194, "num_tokens": 694122848.0, "step": 18277 }, { "epoch": 2.325149472077344, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.7866909503936768, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8744544982910156, "num_tokens": 694160545.0, "step": 18278 }, { "epoch": 2.3252766823559345, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6354355812072754, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8802039623260498, "num_tokens": 694194272.0, "step": 18279 }, { "epoch": 2.325403892634525, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.627265453338623, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8826999664306641, "num_tokens": 694227164.0, "step": 18280 }, { "epoch": 2.325531102913115, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5928337574005127, "learning_rate": 1e-05, "loss": 0.3938, "mean_token_accuracy": 0.8916405439376831, "num_tokens": 694266589.0, "step": 18281 }, { "epoch": 2.325658313191706, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5166592597961426, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.870029091835022, "num_tokens": 694301558.0, "step": 18282 }, { "epoch": 2.325785523470296, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5586860179901123, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.866198718547821, "num_tokens": 694343332.0, "step": 18283 }, { "epoch": 2.3259127337488867, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4455366134643555, "learning_rate": 1e-05, "loss": 0.3906, "mean_token_accuracy": 0.8893818855285645, "num_tokens": 694383602.0, "step": 18284 }, { "epoch": 2.3260399440274773, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5627248287200928, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8675825595855713, "num_tokens": 694417664.0, "step": 18285 }, { "epoch": 2.326167154306068, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3916239738464355, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8631215691566467, "num_tokens": 694456176.0, "step": 18286 }, { "epoch": 2.3262943645846583, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5728039741516113, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8727381229400635, "num_tokens": 694497114.0, "step": 18287 }, { "epoch": 2.326421574863249, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3241324424743652, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.886177122592926, "num_tokens": 694540832.0, "step": 18288 }, { "epoch": 2.3265487851418394, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.620816469192505, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8830329179763794, "num_tokens": 694578673.0, "step": 18289 }, { "epoch": 2.32667599542043, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000316619873046875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.285945415496826, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8613786101341248, "num_tokens": 694624736.0, "step": 18290 }, { "epoch": 2.3268032056990204, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7045068740844727, "learning_rate": 1e-05, "loss": 0.3818, "mean_token_accuracy": 0.8944975733757019, "num_tokens": 694659123.0, "step": 18291 }, { "epoch": 2.326930415977611, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.2217886447906494, "learning_rate": 1e-05, "loss": 0.3955, "mean_token_accuracy": 0.889916718006134, "num_tokens": 694695310.0, "step": 18292 }, { "epoch": 2.3270576262562015, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8517611026763916, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8733024597167969, "num_tokens": 694732470.0, "step": 18293 }, { "epoch": 2.327184836534792, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1906650066375732, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8743724822998047, "num_tokens": 694763693.0, "step": 18294 }, { "epoch": 2.3273120468133826, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.050662517547607, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8885132074356079, "num_tokens": 694798127.0, "step": 18295 }, { "epoch": 2.327439257091973, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2078757286071777, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8818010091781616, "num_tokens": 694833462.0, "step": 18296 }, { "epoch": 2.3275664673705636, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2786359786987305, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8752450942993164, "num_tokens": 694870257.0, "step": 18297 }, { "epoch": 2.327693677649154, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.3929381370544434, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.870155930519104, "num_tokens": 694910455.0, "step": 18298 }, { "epoch": 2.3278208879277447, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.394372463226318, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8803359270095825, "num_tokens": 694947040.0, "step": 18299 }, { "epoch": 2.327948098206335, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6770474910736084, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8751683235168457, "num_tokens": 694989043.0, "step": 18300 }, { "epoch": 2.3280753084849257, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.258402347564697, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8934683203697205, "num_tokens": 695027094.0, "step": 18301 }, { "epoch": 2.3282025187635162, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.816333532333374, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8810229301452637, "num_tokens": 695056419.0, "step": 18302 }, { "epoch": 2.3283297290421068, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.029131889343262, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8747745752334595, "num_tokens": 695101430.0, "step": 18303 }, { "epoch": 2.3284569393206973, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7498507499694824, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8663609027862549, "num_tokens": 695141446.0, "step": 18304 }, { "epoch": 2.328584149599288, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8533284664154053, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8787804841995239, "num_tokens": 695179761.0, "step": 18305 }, { "epoch": 2.328711359877878, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.649759292602539, "learning_rate": 1e-05, "loss": 0.3833, "mean_token_accuracy": 0.8982160091400146, "num_tokens": 695217224.0, "step": 18306 }, { "epoch": 2.328838570156469, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7329044342041016, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8769195675849915, "num_tokens": 695253481.0, "step": 18307 }, { "epoch": 2.328965780435059, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5728466510772705, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8794763088226318, "num_tokens": 695287002.0, "step": 18308 }, { "epoch": 2.3290929907136495, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6495776176452637, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8857420682907104, "num_tokens": 695328296.0, "step": 18309 }, { "epoch": 2.32922020099224, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.490616798400879, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8795626163482666, "num_tokens": 695364291.0, "step": 18310 }, { "epoch": 2.3293474112708306, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6330337524414062, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8615829944610596, "num_tokens": 695407263.0, "step": 18311 }, { "epoch": 2.329474621549421, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4175474643707275, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8622452020645142, "num_tokens": 695444243.0, "step": 18312 }, { "epoch": 2.3296018318280116, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.657665252685547, "learning_rate": 1e-05, "loss": 0.3946, "mean_token_accuracy": 0.8884275555610657, "num_tokens": 695475207.0, "step": 18313 }, { "epoch": 2.329729042106602, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3356635570526123, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8859521150588989, "num_tokens": 695515915.0, "step": 18314 }, { "epoch": 2.3298562523851927, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7412970066070557, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.879287600517273, "num_tokens": 695555168.0, "step": 18315 }, { "epoch": 2.329983462663783, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2636935710906982, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8765435218811035, "num_tokens": 695591829.0, "step": 18316 }, { "epoch": 2.3301106729423737, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8800182342529297, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8748146295547485, "num_tokens": 695628518.0, "step": 18317 }, { "epoch": 2.3302378832209643, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.236896276473999, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8811739683151245, "num_tokens": 695661769.0, "step": 18318 }, { "epoch": 2.330365093499555, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.065830707550049, "learning_rate": 1e-05, "loss": 0.5003, "mean_token_accuracy": 0.8606119155883789, "num_tokens": 695699830.0, "step": 18319 }, { "epoch": 2.3304923037781453, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.281233310699463, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8834788203239441, "num_tokens": 695738871.0, "step": 18320 }, { "epoch": 2.330619514056736, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.244246482849121, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.870273768901825, "num_tokens": 695771868.0, "step": 18321 }, { "epoch": 2.3307467243353264, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4912776947021484, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8702316284179688, "num_tokens": 695808145.0, "step": 18322 }, { "epoch": 2.330873934613917, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.236246585845947, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.897006630897522, "num_tokens": 695837611.0, "step": 18323 }, { "epoch": 2.3310011448925074, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6665704250335693, "learning_rate": 1e-05, "loss": 0.4054, "mean_token_accuracy": 0.8915486931800842, "num_tokens": 695874640.0, "step": 18324 }, { "epoch": 2.331128355171098, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.057318210601807, "learning_rate": 1e-05, "loss": 0.4849, "mean_token_accuracy": 0.8674237728118896, "num_tokens": 695912580.0, "step": 18325 }, { "epoch": 2.3312555654496885, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.707141160964966, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.8867023587226868, "num_tokens": 695952061.0, "step": 18326 }, { "epoch": 2.331382775728279, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.858076810836792, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8899199962615967, "num_tokens": 695986389.0, "step": 18327 }, { "epoch": 2.3315099860068695, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6603493690490723, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8853960633277893, "num_tokens": 696026153.0, "step": 18328 }, { "epoch": 2.3316371962854596, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6952316761016846, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8640189170837402, "num_tokens": 696062693.0, "step": 18329 }, { "epoch": 2.3317644065640506, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.609004497528076, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8832079768180847, "num_tokens": 696101933.0, "step": 18330 }, { "epoch": 2.3318916168426407, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.573777437210083, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8738416433334351, "num_tokens": 696137607.0, "step": 18331 }, { "epoch": 2.332018827121231, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.558117151260376, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8788678050041199, "num_tokens": 696181277.0, "step": 18332 }, { "epoch": 2.3321460373998217, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5136122703552246, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8777540922164917, "num_tokens": 696216592.0, "step": 18333 }, { "epoch": 2.3322732476784123, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.522461175918579, "learning_rate": 1e-05, "loss": 0.3812, "mean_token_accuracy": 0.8938660621643066, "num_tokens": 696250762.0, "step": 18334 }, { "epoch": 2.332400457957003, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4680464267730713, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8767792582511902, "num_tokens": 696289819.0, "step": 18335 }, { "epoch": 2.3325276682355933, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4980533123016357, "learning_rate": 1e-05, "loss": 0.5074, "mean_token_accuracy": 0.8526493310928345, "num_tokens": 696335551.0, "step": 18336 }, { "epoch": 2.332654878514184, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.447923183441162, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8823943138122559, "num_tokens": 696369973.0, "step": 18337 }, { "epoch": 2.3327820887927744, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4874842166900635, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8730243444442749, "num_tokens": 696403718.0, "step": 18338 }, { "epoch": 2.332909299071365, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4304728507995605, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8740547895431519, "num_tokens": 696440769.0, "step": 18339 }, { "epoch": 2.3330365093499554, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.472407579421997, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8783918619155884, "num_tokens": 696482570.0, "step": 18340 }, { "epoch": 2.333163719628546, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4034037590026855, "learning_rate": 1e-05, "loss": 0.3886, "mean_token_accuracy": 0.8915302157402039, "num_tokens": 696519199.0, "step": 18341 }, { "epoch": 2.3332909299071365, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4920692443847656, "learning_rate": 1e-05, "loss": 0.4689, "mean_token_accuracy": 0.8634167313575745, "num_tokens": 696558460.0, "step": 18342 }, { "epoch": 2.333418140185727, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.388310432434082, "learning_rate": 1e-05, "loss": 0.5127, "mean_token_accuracy": 0.8514024019241333, "num_tokens": 696600395.0, "step": 18343 }, { "epoch": 2.3335453504643175, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5209782123565674, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8749381303787231, "num_tokens": 696631478.0, "step": 18344 }, { "epoch": 2.333672560742908, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.334986925125122, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8836930394172668, "num_tokens": 696669634.0, "step": 18345 }, { "epoch": 2.3337997710214986, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5758230686187744, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8724141120910645, "num_tokens": 696707642.0, "step": 18346 }, { "epoch": 2.333926981300089, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2557361125946045, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8841173648834229, "num_tokens": 696747426.0, "step": 18347 }, { "epoch": 2.3340541915786797, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6955111026763916, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.866370439529419, "num_tokens": 696784114.0, "step": 18348 }, { "epoch": 2.33418140185727, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.173261880874634, "learning_rate": 1e-05, "loss": 0.3966, "mean_token_accuracy": 0.8861740827560425, "num_tokens": 696822303.0, "step": 18349 }, { "epoch": 2.3343086121358607, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.907036781311035, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8761848211288452, "num_tokens": 696855460.0, "step": 18350 }, { "epoch": 2.3344358224144512, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.106640100479126, "learning_rate": 1e-05, "loss": 0.3554, "mean_token_accuracy": 0.9029549360275269, "num_tokens": 696891070.0, "step": 18351 }, { "epoch": 2.3345630326930418, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.227504730224609, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8807156085968018, "num_tokens": 696924679.0, "step": 18352 }, { "epoch": 2.3346902429716323, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.1960911750793457, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8837891817092896, "num_tokens": 696958894.0, "step": 18353 }, { "epoch": 2.3348174532502224, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 7.390975952148438e-05, "grad_norm": 4.574072360992432, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8721051812171936, "num_tokens": 696995371.0, "step": 18354 }, { "epoch": 2.3349446635288134, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.627758264541626, "learning_rate": 1e-05, "loss": 0.4039, "mean_token_accuracy": 0.8912058472633362, "num_tokens": 697034284.0, "step": 18355 }, { "epoch": 2.3350718738074034, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.576668739318848, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8913089036941528, "num_tokens": 697069805.0, "step": 18356 }, { "epoch": 2.335199084085994, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 3.990994930267334, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8816001415252686, "num_tokens": 697105326.0, "step": 18357 }, { "epoch": 2.3353262943645845, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.27451753616333, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8816014528274536, "num_tokens": 697144157.0, "step": 18358 }, { "epoch": 2.335453504643175, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.955874443054199, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8874307870864868, "num_tokens": 697178261.0, "step": 18359 }, { "epoch": 2.3355807149217656, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.043233871459961, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.882208526134491, "num_tokens": 697224044.0, "step": 18360 }, { "epoch": 2.335707925200356, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.802173137664795, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8827300667762756, "num_tokens": 697262007.0, "step": 18361 }, { "epoch": 2.3358351354789466, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8964405059814453, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8868362903594971, "num_tokens": 697301689.0, "step": 18362 }, { "epoch": 2.335962345757537, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6536142826080322, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8818406462669373, "num_tokens": 697343509.0, "step": 18363 }, { "epoch": 2.3360895560361277, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8187694549560547, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8816474676132202, "num_tokens": 697380620.0, "step": 18364 }, { "epoch": 2.336216766314718, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5281214714050293, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8736485242843628, "num_tokens": 697422811.0, "step": 18365 }, { "epoch": 2.3363439765933087, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.8118326663970947, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8753488063812256, "num_tokens": 697460527.0, "step": 18366 }, { "epoch": 2.3364711868718993, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.421576976776123, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8826574683189392, "num_tokens": 697495417.0, "step": 18367 }, { "epoch": 2.33659839715049, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8706812858581543, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8856250047683716, "num_tokens": 697534745.0, "step": 18368 }, { "epoch": 2.3367256074290803, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.330789089202881, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8800594806671143, "num_tokens": 697572074.0, "step": 18369 }, { "epoch": 2.336852817707671, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 4.020228862762451, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8757457733154297, "num_tokens": 697606943.0, "step": 18370 }, { "epoch": 2.3369800279862614, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3276636600494385, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.878614604473114, "num_tokens": 697643634.0, "step": 18371 }, { "epoch": 2.337107238264852, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.175157070159912, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.876876175403595, "num_tokens": 697682429.0, "step": 18372 }, { "epoch": 2.3372344485434424, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.47183895111084, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8774551153182983, "num_tokens": 697718039.0, "step": 18373 }, { "epoch": 2.337361658822033, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.215310573577881, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8848061561584473, "num_tokens": 697756179.0, "step": 18374 }, { "epoch": 2.3374888691006235, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6476187705993652, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8826469779014587, "num_tokens": 697795327.0, "step": 18375 }, { "epoch": 2.337616079379214, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0749592781066895, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8772748112678528, "num_tokens": 697830423.0, "step": 18376 }, { "epoch": 2.3377432896578045, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.7419979572296143, "learning_rate": 1e-05, "loss": 0.4958, "mean_token_accuracy": 0.8583008050918579, "num_tokens": 697870890.0, "step": 18377 }, { "epoch": 2.337870499936395, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.827986478805542, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8697213530540466, "num_tokens": 697910880.0, "step": 18378 }, { "epoch": 2.337997710214985, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.716625928878784, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8860056400299072, "num_tokens": 697945509.0, "step": 18379 }, { "epoch": 2.338124920493576, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.619098663330078, "learning_rate": 1e-05, "loss": 0.4985, "mean_token_accuracy": 0.8595731854438782, "num_tokens": 697987374.0, "step": 18380 }, { "epoch": 2.338252130772166, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6907334327697754, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8776854276657104, "num_tokens": 698025130.0, "step": 18381 }, { "epoch": 2.3383793410507567, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4671356678009033, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8781177401542664, "num_tokens": 698057785.0, "step": 18382 }, { "epoch": 2.3385065513293473, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.676227569580078, "learning_rate": 1e-05, "loss": 0.4121, "mean_token_accuracy": 0.8871732950210571, "num_tokens": 698098952.0, "step": 18383 }, { "epoch": 2.338633761607938, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.352062940597534, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8869496583938599, "num_tokens": 698142280.0, "step": 18384 }, { "epoch": 2.3387609718865283, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.725541353225708, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8740944862365723, "num_tokens": 698183083.0, "step": 18385 }, { "epoch": 2.338888182165119, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2987561225891113, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8760453462600708, "num_tokens": 698215446.0, "step": 18386 }, { "epoch": 2.3390153924437094, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.82356333732605, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8751641511917114, "num_tokens": 698257353.0, "step": 18387 }, { "epoch": 2.3391426027223, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2401912212371826, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8765287399291992, "num_tokens": 698292753.0, "step": 18388 }, { "epoch": 2.3392698130008904, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9710216522216797, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.881819486618042, "num_tokens": 698327282.0, "step": 18389 }, { "epoch": 2.339397023279481, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2431106567382812, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8773266673088074, "num_tokens": 698368006.0, "step": 18390 }, { "epoch": 2.3395242335580715, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.145537376403809, "learning_rate": 1e-05, "loss": 0.5203, "mean_token_accuracy": 0.8495277762413025, "num_tokens": 698406898.0, "step": 18391 }, { "epoch": 2.339651443836662, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.354127883911133, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8741180300712585, "num_tokens": 698448159.0, "step": 18392 }, { "epoch": 2.3397786541152525, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.244535446166992, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8861593008041382, "num_tokens": 698489075.0, "step": 18393 }, { "epoch": 2.339905864393843, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5544354915618896, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8752070665359497, "num_tokens": 698527619.0, "step": 18394 }, { "epoch": 2.3400330746724336, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.185967922210693, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8703587055206299, "num_tokens": 698565250.0, "step": 18395 }, { "epoch": 2.340160284951024, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6827642917633057, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8806340098381042, "num_tokens": 698603207.0, "step": 18396 }, { "epoch": 2.3402874952296147, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9977951049804688, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8919728994369507, "num_tokens": 698636747.0, "step": 18397 }, { "epoch": 2.340414705508205, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.677419424057007, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8890257477760315, "num_tokens": 698677885.0, "step": 18398 }, { "epoch": 2.3405419157867957, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.813875436782837, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8766542077064514, "num_tokens": 698721924.0, "step": 18399 }, { "epoch": 2.3406691260653862, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.646981954574585, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8664910793304443, "num_tokens": 698754847.0, "step": 18400 }, { "epoch": 2.3407963363439768, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003185272216796875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.711850166320801, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8894054889678955, "num_tokens": 698798567.0, "step": 18401 }, { "epoch": 2.3409235466225673, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.563065528869629, "learning_rate": 1e-05, "loss": 0.4029, "mean_token_accuracy": 0.8848723769187927, "num_tokens": 698837574.0, "step": 18402 }, { "epoch": 2.341050756901158, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6347174644470215, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.8746192455291748, "num_tokens": 698867550.0, "step": 18403 }, { "epoch": 2.341177967179748, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4921634197235107, "learning_rate": 1e-05, "loss": 0.3957, "mean_token_accuracy": 0.8891257643699646, "num_tokens": 698898859.0, "step": 18404 }, { "epoch": 2.341305177458339, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5953943729400635, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8827900886535645, "num_tokens": 698933449.0, "step": 18405 }, { "epoch": 2.341432387736929, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4331490993499756, "learning_rate": 1e-05, "loss": 0.3925, "mean_token_accuracy": 0.8903952836990356, "num_tokens": 698969045.0, "step": 18406 }, { "epoch": 2.3415595980155195, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.60842227935791, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8680557012557983, "num_tokens": 699006201.0, "step": 18407 }, { "epoch": 2.34168680829411, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3549704551696777, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.886489748954773, "num_tokens": 699044091.0, "step": 18408 }, { "epoch": 2.3418140185727006, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6658735275268555, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8766596913337708, "num_tokens": 699075519.0, "step": 18409 }, { "epoch": 2.341941228851291, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2926583290100098, "learning_rate": 1e-05, "loss": 0.4046, "mean_token_accuracy": 0.8880220651626587, "num_tokens": 699113129.0, "step": 18410 }, { "epoch": 2.3420684391298816, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7674033641815186, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8790515065193176, "num_tokens": 699147515.0, "step": 18411 }, { "epoch": 2.342195649408472, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.222999095916748, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8745758533477783, "num_tokens": 699190578.0, "step": 18412 }, { "epoch": 2.3423228596870627, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.935610055923462, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.872852087020874, "num_tokens": 699230422.0, "step": 18413 }, { "epoch": 2.342450069965653, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.240431308746338, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8808144330978394, "num_tokens": 699268662.0, "step": 18414 }, { "epoch": 2.3425772802442437, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.109660625457764, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8728495836257935, "num_tokens": 699312498.0, "step": 18415 }, { "epoch": 2.3427044905228342, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3153090476989746, "learning_rate": 1e-05, "loss": 0.3969, "mean_token_accuracy": 0.888088583946228, "num_tokens": 699354143.0, "step": 18416 }, { "epoch": 2.3428317008014248, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2554931640625, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8849399089813232, "num_tokens": 699392099.0, "step": 18417 }, { "epoch": 2.3429589110800153, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5349442958831787, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.8844963312149048, "num_tokens": 699428449.0, "step": 18418 }, { "epoch": 2.343086121358606, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.189116477966309, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8764756917953491, "num_tokens": 699469359.0, "step": 18419 }, { "epoch": 2.3432133316371964, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6771037578582764, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8867557048797607, "num_tokens": 699507481.0, "step": 18420 }, { "epoch": 2.343340541915787, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.001259803771973, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8702999949455261, "num_tokens": 699548335.0, "step": 18421 }, { "epoch": 2.3434677521943774, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.703195333480835, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8702099323272705, "num_tokens": 699587460.0, "step": 18422 }, { "epoch": 2.343594962472968, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.792011260986328, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8713130950927734, "num_tokens": 699631128.0, "step": 18423 }, { "epoch": 2.3437221727515585, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6557722091674805, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.8685156106948853, "num_tokens": 699672133.0, "step": 18424 }, { "epoch": 2.343849383030149, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6447818279266357, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8743975162506104, "num_tokens": 699705280.0, "step": 18425 }, { "epoch": 2.3439765933087395, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6101438999176025, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8743182420730591, "num_tokens": 699740959.0, "step": 18426 }, { "epoch": 2.3441038035873296, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.546239137649536, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8625787496566772, "num_tokens": 699775546.0, "step": 18427 }, { "epoch": 2.3442310138659206, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5610713958740234, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8788143396377563, "num_tokens": 699809777.0, "step": 18428 }, { "epoch": 2.3443582241445107, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4655911922454834, "learning_rate": 1e-05, "loss": 0.3867, "mean_token_accuracy": 0.8913018703460693, "num_tokens": 699846140.0, "step": 18429 }, { "epoch": 2.344485434423101, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5304222106933594, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8791667222976685, "num_tokens": 699882769.0, "step": 18430 }, { "epoch": 2.3446126447016917, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4276230335235596, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8753572106361389, "num_tokens": 699923658.0, "step": 18431 }, { "epoch": 2.3447398549802823, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.525630235671997, "learning_rate": 1e-05, "loss": 0.3772, "mean_token_accuracy": 0.891158938407898, "num_tokens": 699957457.0, "step": 18432 }, { "epoch": 2.344867065258873, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3912229537963867, "learning_rate": 1e-05, "loss": 0.3937, "mean_token_accuracy": 0.8886920213699341, "num_tokens": 699994852.0, "step": 18433 }, { "epoch": 2.3449942755374633, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5441653728485107, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8850245475769043, "num_tokens": 700030983.0, "step": 18434 }, { "epoch": 2.345121485816054, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.349478006362915, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8836489915847778, "num_tokens": 700067907.0, "step": 18435 }, { "epoch": 2.3452486960946444, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5826289653778076, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8789363503456116, "num_tokens": 700107989.0, "step": 18436 }, { "epoch": 2.345375906373235, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3288512229919434, "learning_rate": 1e-05, "loss": 0.3887, "mean_token_accuracy": 0.891416072845459, "num_tokens": 700145870.0, "step": 18437 }, { "epoch": 2.3455031166518254, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6266298294067383, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8879076242446899, "num_tokens": 700186762.0, "step": 18438 }, { "epoch": 2.345630326930416, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2520387172698975, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8846852779388428, "num_tokens": 700227307.0, "step": 18439 }, { "epoch": 2.3457575372090065, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.749767541885376, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8753522634506226, "num_tokens": 700263660.0, "step": 18440 }, { "epoch": 2.345884747487597, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1890106201171875, "learning_rate": 1e-05, "loss": 0.3808, "mean_token_accuracy": 0.894253134727478, "num_tokens": 700294835.0, "step": 18441 }, { "epoch": 2.3460119577661875, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9319074153900146, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8854703903198242, "num_tokens": 700335977.0, "step": 18442 }, { "epoch": 2.346139168044778, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.144362211227417, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8781002759933472, "num_tokens": 700369388.0, "step": 18443 }, { "epoch": 2.3462663783233686, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.200260639190674, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8778332471847534, "num_tokens": 700409575.0, "step": 18444 }, { "epoch": 2.346393588601959, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.2593514919281006, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.889650821685791, "num_tokens": 700449463.0, "step": 18445 }, { "epoch": 2.3465207988805497, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.440573215484619, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8819261789321899, "num_tokens": 700483765.0, "step": 18446 }, { "epoch": 2.34664800915914, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.5682473182678223, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8842006325721741, "num_tokens": 700521953.0, "step": 18447 }, { "epoch": 2.3467752194377307, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.429668426513672, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.87350994348526, "num_tokens": 700559662.0, "step": 18448 }, { "epoch": 2.3469024297163212, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.8429601192474365, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8806942105293274, "num_tokens": 700598880.0, "step": 18449 }, { "epoch": 2.3470296399949118, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.210608005523682, "learning_rate": 1e-05, "loss": 0.3859, "mean_token_accuracy": 0.8966354131698608, "num_tokens": 700631712.0, "step": 18450 }, { "epoch": 2.3471568502735023, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.85604190826416, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.8680652976036072, "num_tokens": 700665420.0, "step": 18451 }, { "epoch": 2.3472840605520924, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.9832842350006104, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8822256326675415, "num_tokens": 700704603.0, "step": 18452 }, { "epoch": 2.3474112708306833, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.740074872970581, "learning_rate": 1e-05, "loss": 0.4065, "mean_token_accuracy": 0.8850528597831726, "num_tokens": 700738730.0, "step": 18453 }, { "epoch": 2.3475384811092734, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.845144033432007, "learning_rate": 1e-05, "loss": 0.3896, "mean_token_accuracy": 0.8956115245819092, "num_tokens": 700773848.0, "step": 18454 }, { "epoch": 2.347665691387864, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.632338762283325, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8795039057731628, "num_tokens": 700810352.0, "step": 18455 }, { "epoch": 2.3477929016664545, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.752385377883911, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8748502731323242, "num_tokens": 700846432.0, "step": 18456 }, { "epoch": 2.347920111945045, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.526256799697876, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8848133683204651, "num_tokens": 700883674.0, "step": 18457 }, { "epoch": 2.3480473222236355, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.710838556289673, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8605226278305054, "num_tokens": 700925096.0, "step": 18458 }, { "epoch": 2.348174532502226, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4265518188476562, "learning_rate": 1e-05, "loss": 0.385, "mean_token_accuracy": 0.8940243721008301, "num_tokens": 700957301.0, "step": 18459 }, { "epoch": 2.3483017427808166, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7423555850982666, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8882888555526733, "num_tokens": 700990636.0, "step": 18460 }, { "epoch": 2.348428953059407, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3405449390411377, "learning_rate": 1e-05, "loss": 0.3843, "mean_token_accuracy": 0.8909182548522949, "num_tokens": 701026379.0, "step": 18461 }, { "epoch": 2.3485561633379977, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8270580768585205, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8791090846061707, "num_tokens": 701070377.0, "step": 18462 }, { "epoch": 2.348683373616588, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3065106868743896, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8855715990066528, "num_tokens": 701107691.0, "step": 18463 }, { "epoch": 2.3488105838951787, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9508724212646484, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8810331225395203, "num_tokens": 701147775.0, "step": 18464 }, { "epoch": 2.3489377941737692, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.283252716064453, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.890375018119812, "num_tokens": 701187732.0, "step": 18465 }, { "epoch": 2.3490650044523598, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.112305641174316, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8911601305007935, "num_tokens": 701220916.0, "step": 18466 }, { "epoch": 2.3491922147309503, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.387714385986328, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8834342956542969, "num_tokens": 701258831.0, "step": 18467 }, { "epoch": 2.349319425009541, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.180539608001709, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.886539876461029, "num_tokens": 701299310.0, "step": 18468 }, { "epoch": 2.3494466352881314, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5836403369903564, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8720345497131348, "num_tokens": 701331471.0, "step": 18469 }, { "epoch": 2.349573845566722, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.0591254234313965, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.879859447479248, "num_tokens": 701371126.0, "step": 18470 }, { "epoch": 2.3497010558453124, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.648216962814331, "learning_rate": 1e-05, "loss": 0.3657, "mean_token_accuracy": 0.9017306566238403, "num_tokens": 701413612.0, "step": 18471 }, { "epoch": 2.349828266123903, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8760170936584473, "learning_rate": 1e-05, "loss": 0.4993, "mean_token_accuracy": 0.8587349653244019, "num_tokens": 701456085.0, "step": 18472 }, { "epoch": 2.3499554764024935, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.678584098815918, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8788947463035583, "num_tokens": 701494409.0, "step": 18473 }, { "epoch": 2.350082686681084, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6660873889923096, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8782126903533936, "num_tokens": 701533610.0, "step": 18474 }, { "epoch": 2.3502098969596745, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6595165729522705, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8652647733688354, "num_tokens": 701571324.0, "step": 18475 }, { "epoch": 2.350337107238265, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.509777069091797, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8667482733726501, "num_tokens": 701609077.0, "step": 18476 }, { "epoch": 2.350464317516855, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6480534076690674, "learning_rate": 1e-05, "loss": 0.3685, "mean_token_accuracy": 0.8990181684494019, "num_tokens": 701647281.0, "step": 18477 }, { "epoch": 2.350591527795446, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3842523097991943, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8895697593688965, "num_tokens": 701680039.0, "step": 18478 }, { "epoch": 2.350718738074036, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.679443120956421, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8756073713302612, "num_tokens": 701721679.0, "step": 18479 }, { "epoch": 2.3508459483526267, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.297485113143921, "learning_rate": 1e-05, "loss": 0.3715, "mean_token_accuracy": 0.8972774744033813, "num_tokens": 701758692.0, "step": 18480 }, { "epoch": 2.3509731586312173, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7744038105010986, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8650466203689575, "num_tokens": 701793038.0, "step": 18481 }, { "epoch": 2.351100368909808, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.24644136428833, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8790163993835449, "num_tokens": 701831172.0, "step": 18482 }, { "epoch": 2.3512275791883983, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9037981033325195, "learning_rate": 1e-05, "loss": 0.3954, "mean_token_accuracy": 0.8912518620491028, "num_tokens": 701865644.0, "step": 18483 }, { "epoch": 2.351354789466989, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.213637113571167, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8811408877372742, "num_tokens": 701903928.0, "step": 18484 }, { "epoch": 2.3514819997455794, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.09155797958374, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8654590845108032, "num_tokens": 701937440.0, "step": 18485 }, { "epoch": 2.35160921002417, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.289781332015991, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8811436295509338, "num_tokens": 701975936.0, "step": 18486 }, { "epoch": 2.3517364203027604, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.232170581817627, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8769761323928833, "num_tokens": 702009504.0, "step": 18487 }, { "epoch": 2.351863630581351, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4794888496398926, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8748968243598938, "num_tokens": 702046631.0, "step": 18488 }, { "epoch": 2.3519908408599415, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.22573709487915, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8805428743362427, "num_tokens": 702086915.0, "step": 18489 }, { "epoch": 2.352118051138532, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6272826194763184, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8746254444122314, "num_tokens": 702129214.0, "step": 18490 }, { "epoch": 2.3522452614171225, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.109397888183594, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8748700022697449, "num_tokens": 702167450.0, "step": 18491 }, { "epoch": 2.352372471695713, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6857998371124268, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8807336688041687, "num_tokens": 702205967.0, "step": 18492 }, { "epoch": 2.3524996819743036, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9309194087982178, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8779894113540649, "num_tokens": 702239570.0, "step": 18493 }, { "epoch": 2.352626892252894, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.665252208709717, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8750459551811218, "num_tokens": 702279276.0, "step": 18494 }, { "epoch": 2.3527541025314846, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7632784843444824, "learning_rate": 1e-05, "loss": 0.4167, "mean_token_accuracy": 0.8867385387420654, "num_tokens": 702319695.0, "step": 18495 }, { "epoch": 2.352881312810075, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.590816020965576, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8853693008422852, "num_tokens": 702361933.0, "step": 18496 }, { "epoch": 2.3530085230886657, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6733040809631348, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8804884552955627, "num_tokens": 702398532.0, "step": 18497 }, { "epoch": 2.3531357333672562, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5260515213012695, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8813276290893555, "num_tokens": 702437880.0, "step": 18498 }, { "epoch": 2.3532629436458468, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6133129596710205, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8743928670883179, "num_tokens": 702478908.0, "step": 18499 }, { "epoch": 2.353390153924437, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.464576005935669, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8710054159164429, "num_tokens": 702523817.0, "step": 18500 }, { "epoch": 2.353517364203028, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.586916923522949, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.880959153175354, "num_tokens": 702569682.0, "step": 18501 }, { "epoch": 2.353644574481618, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.415864944458008, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.873185396194458, "num_tokens": 702604746.0, "step": 18502 }, { "epoch": 2.353771784760209, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6332714557647705, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8702619671821594, "num_tokens": 702637660.0, "step": 18503 }, { "epoch": 2.353898995038799, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3295278549194336, "learning_rate": 1e-05, "loss": 0.3696, "mean_token_accuracy": 0.8977431654930115, "num_tokens": 702676942.0, "step": 18504 }, { "epoch": 2.3540262053173895, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.716552257537842, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8661683201789856, "num_tokens": 702718629.0, "step": 18505 }, { "epoch": 2.35415341559598, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2643649578094482, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8813631534576416, "num_tokens": 702757698.0, "step": 18506 }, { "epoch": 2.3542806258745705, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.845745801925659, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8878191709518433, "num_tokens": 702797337.0, "step": 18507 }, { "epoch": 2.354407836153161, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.225539445877075, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8695604205131531, "num_tokens": 702832002.0, "step": 18508 }, { "epoch": 2.3545350464317516, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0110979080200195, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.885056734085083, "num_tokens": 702874472.0, "step": 18509 }, { "epoch": 2.354662256710342, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2678277492523193, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8769679665565491, "num_tokens": 702909867.0, "step": 18510 }, { "epoch": 2.3547894669889327, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.173895835876465, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8878889083862305, "num_tokens": 702945817.0, "step": 18511 }, { "epoch": 2.354916677267523, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3991622924804688, "learning_rate": 1e-05, "loss": 0.3916, "mean_token_accuracy": 0.895822286605835, "num_tokens": 702986373.0, "step": 18512 }, { "epoch": 2.3550438875461137, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.218924522399902, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8780820369720459, "num_tokens": 703026077.0, "step": 18513 }, { "epoch": 2.3551710978247042, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6000912189483643, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8799189329147339, "num_tokens": 703067042.0, "step": 18514 }, { "epoch": 2.3552983081032948, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.0868144035339355, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.887120246887207, "num_tokens": 703105290.0, "step": 18515 }, { "epoch": 2.3554255183818853, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.686610698699951, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.876494824886322, "num_tokens": 703144446.0, "step": 18516 }, { "epoch": 2.355552728660476, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9013431072235107, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.874535322189331, "num_tokens": 703179490.0, "step": 18517 }, { "epoch": 2.3556799389390664, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6704275608062744, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8734349012374878, "num_tokens": 703214218.0, "step": 18518 }, { "epoch": 2.355807149217657, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6988492012023926, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8837103843688965, "num_tokens": 703252576.0, "step": 18519 }, { "epoch": 2.3559343594962474, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.618530035018921, "learning_rate": 1e-05, "loss": 0.3629, "mean_token_accuracy": 0.899766743183136, "num_tokens": 703291129.0, "step": 18520 }, { "epoch": 2.356061569774838, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5795645713806152, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8702757954597473, "num_tokens": 703326864.0, "step": 18521 }, { "epoch": 2.3561887800534285, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5682811737060547, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8793361186981201, "num_tokens": 703369719.0, "step": 18522 }, { "epoch": 2.356315990332019, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4919729232788086, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8791257739067078, "num_tokens": 703415304.0, "step": 18523 }, { "epoch": 2.3564432006106095, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5440845489501953, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8735430240631104, "num_tokens": 703456273.0, "step": 18524 }, { "epoch": 2.3565704108891996, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.454660177230835, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8740891814231873, "num_tokens": 703491355.0, "step": 18525 }, { "epoch": 2.3566976211677906, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.520460844039917, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8823456168174744, "num_tokens": 703532951.0, "step": 18526 }, { "epoch": 2.3568248314463807, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4204182624816895, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8801268339157104, "num_tokens": 703575306.0, "step": 18527 }, { "epoch": 2.356952041724971, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5070154666900635, "learning_rate": 1e-05, "loss": 0.3965, "mean_token_accuracy": 0.8876440525054932, "num_tokens": 703615198.0, "step": 18528 }, { "epoch": 2.3570792520035617, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.395519495010376, "learning_rate": 1e-05, "loss": 0.3935, "mean_token_accuracy": 0.8885795474052429, "num_tokens": 703651780.0, "step": 18529 }, { "epoch": 2.3572064622821522, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.538283348083496, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8729805946350098, "num_tokens": 703685273.0, "step": 18530 }, { "epoch": 2.3573336725607428, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.335134983062744, "learning_rate": 1e-05, "loss": 0.388, "mean_token_accuracy": 0.890580415725708, "num_tokens": 703722185.0, "step": 18531 }, { "epoch": 2.3574608828393333, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.592825174331665, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8855428695678711, "num_tokens": 703760229.0, "step": 18532 }, { "epoch": 2.357588093117924, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2929232120513916, "learning_rate": 1e-05, "loss": 0.3941, "mean_token_accuracy": 0.8880192637443542, "num_tokens": 703794286.0, "step": 18533 }, { "epoch": 2.3577153033965144, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.682650089263916, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8848884105682373, "num_tokens": 703827486.0, "step": 18534 }, { "epoch": 2.357842513675105, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2197535037994385, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.864960789680481, "num_tokens": 703861398.0, "step": 18535 }, { "epoch": 2.3579697239536954, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8399147987365723, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8751264810562134, "num_tokens": 703901690.0, "step": 18536 }, { "epoch": 2.358096934232286, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.138617753982544, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8809181451797485, "num_tokens": 703943828.0, "step": 18537 }, { "epoch": 2.3582241445108765, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.087514877319336, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8730700016021729, "num_tokens": 703984893.0, "step": 18538 }, { "epoch": 2.358351354789467, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.173750400543213, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8700117468833923, "num_tokens": 704023633.0, "step": 18539 }, { "epoch": 2.3584785650680575, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.388429164886475, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8777929544448853, "num_tokens": 704058762.0, "step": 18540 }, { "epoch": 2.358605775346648, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.434678077697754, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8765854835510254, "num_tokens": 704099100.0, "step": 18541 }, { "epoch": 2.3587329856252386, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.522438049316406, "learning_rate": 1e-05, "loss": 0.5033, "mean_token_accuracy": 0.8620957136154175, "num_tokens": 704131900.0, "step": 18542 }, { "epoch": 2.358860195903829, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.816322088241577, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8852057456970215, "num_tokens": 704161940.0, "step": 18543 }, { "epoch": 2.3589874061824196, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.306733131408691, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.887343168258667, "num_tokens": 704200698.0, "step": 18544 }, { "epoch": 2.35911461646101, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.8841989040374756, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8777573704719543, "num_tokens": 704243317.0, "step": 18545 }, { "epoch": 2.3592418267396007, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.092462539672852, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8899906277656555, "num_tokens": 704277258.0, "step": 18546 }, { "epoch": 2.3593690370181912, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.783881425857544, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8904991745948792, "num_tokens": 704315932.0, "step": 18547 }, { "epoch": 2.3594962472967818, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.929316759109497, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8650315999984741, "num_tokens": 704357355.0, "step": 18548 }, { "epoch": 2.3596234575753723, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6822047233581543, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8721715807914734, "num_tokens": 704392981.0, "step": 18549 }, { "epoch": 2.3597506678539624, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.822345018386841, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8809322118759155, "num_tokens": 704430530.0, "step": 18550 }, { "epoch": 2.3598778781325533, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5600662231445312, "learning_rate": 1e-05, "loss": 0.3949, "mean_token_accuracy": 0.8896242380142212, "num_tokens": 704467444.0, "step": 18551 }, { "epoch": 2.3600050884111434, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7848494052886963, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8914123177528381, "num_tokens": 704504179.0, "step": 18552 }, { "epoch": 2.360132298689734, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4557840824127197, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8839970827102661, "num_tokens": 704545346.0, "step": 18553 }, { "epoch": 2.3602595089683245, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8262033462524414, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8690468072891235, "num_tokens": 704583453.0, "step": 18554 }, { "epoch": 2.360386719246915, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3692586421966553, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8773574829101562, "num_tokens": 704617210.0, "step": 18555 }, { "epoch": 2.3605139295255055, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.925374984741211, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8850749731063843, "num_tokens": 704651002.0, "step": 18556 }, { "epoch": 2.360641139804096, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3132364749908447, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8778252601623535, "num_tokens": 704694718.0, "step": 18557 }, { "epoch": 2.3607683500826866, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.091943740844727, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8887039422988892, "num_tokens": 704730266.0, "step": 18558 }, { "epoch": 2.360895560361277, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3947980403900146, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8835660815238953, "num_tokens": 704766422.0, "step": 18559 }, { "epoch": 2.3610227706398677, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.173646450042725, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8817992210388184, "num_tokens": 704802133.0, "step": 18560 }, { "epoch": 2.361149980918458, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.5604500770568848, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8880254030227661, "num_tokens": 704832108.0, "step": 18561 }, { "epoch": 2.3612771911970487, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.0952467918396, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8720088601112366, "num_tokens": 704867056.0, "step": 18562 }, { "epoch": 2.3614044014756392, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.661325454711914, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8662125468254089, "num_tokens": 704905565.0, "step": 18563 }, { "epoch": 2.3615316117542298, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.88818359375, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8805804252624512, "num_tokens": 704949690.0, "step": 18564 }, { "epoch": 2.3616588220328203, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6977005004882812, "learning_rate": 1e-05, "loss": 0.5015, "mean_token_accuracy": 0.8629765510559082, "num_tokens": 704985123.0, "step": 18565 }, { "epoch": 2.361786032311411, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6820385456085205, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8832640051841736, "num_tokens": 705020317.0, "step": 18566 }, { "epoch": 2.3619132425900013, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6597602367401123, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.8852049112319946, "num_tokens": 705059025.0, "step": 18567 }, { "epoch": 2.362040452868592, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5119469165802, "learning_rate": 1e-05, "loss": 0.3741, "mean_token_accuracy": 0.8953550457954407, "num_tokens": 705094157.0, "step": 18568 }, { "epoch": 2.3621676631471824, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6448376178741455, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8812432885169983, "num_tokens": 705133642.0, "step": 18569 }, { "epoch": 2.362294873425773, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4299213886260986, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8712763786315918, "num_tokens": 705165253.0, "step": 18570 }, { "epoch": 2.3624220837043635, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.639728546142578, "learning_rate": 1e-05, "loss": 0.3933, "mean_token_accuracy": 0.8897558450698853, "num_tokens": 705203847.0, "step": 18571 }, { "epoch": 2.362549293982954, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.343312978744507, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8745976686477661, "num_tokens": 705241984.0, "step": 18572 }, { "epoch": 2.3626765042615445, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6941559314727783, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8688207864761353, "num_tokens": 705278462.0, "step": 18573 }, { "epoch": 2.362803714540135, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2672219276428223, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8879409432411194, "num_tokens": 705313004.0, "step": 18574 }, { "epoch": 2.362930924818725, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7939505577087402, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8908865451812744, "num_tokens": 705350225.0, "step": 18575 }, { "epoch": 2.363058135097316, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2313880920410156, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.8853477239608765, "num_tokens": 705380213.0, "step": 18576 }, { "epoch": 2.363185345375906, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9352481365203857, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.8713507652282715, "num_tokens": 705419059.0, "step": 18577 }, { "epoch": 2.3633125556544967, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2046735286712646, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8759818077087402, "num_tokens": 705459506.0, "step": 18578 }, { "epoch": 2.3634397659330872, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.144927501678467, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8785489797592163, "num_tokens": 705495719.0, "step": 18579 }, { "epoch": 2.3635669762116778, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3209946155548096, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8656543493270874, "num_tokens": 705536040.0, "step": 18580 }, { "epoch": 2.3636941864902683, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.290246486663818, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8795111775398254, "num_tokens": 705571541.0, "step": 18581 }, { "epoch": 2.363821396768859, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.532606840133667, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8834177851676941, "num_tokens": 705610048.0, "step": 18582 }, { "epoch": 2.3639486070474494, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.258018493652344, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.886449933052063, "num_tokens": 705645098.0, "step": 18583 }, { "epoch": 2.36407581732604, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.708181381225586, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8795226812362671, "num_tokens": 705680962.0, "step": 18584 }, { "epoch": 2.3642030276046304, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.076006889343262, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8758601546287537, "num_tokens": 705716983.0, "step": 18585 }, { "epoch": 2.364330237883221, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.737557888031006, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8810573816299438, "num_tokens": 705749735.0, "step": 18586 }, { "epoch": 2.3644574481618115, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.873229503631592, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8717730045318604, "num_tokens": 705788747.0, "step": 18587 }, { "epoch": 2.364584658440402, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6770341396331787, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8745998740196228, "num_tokens": 705827475.0, "step": 18588 }, { "epoch": 2.3647118687189925, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.710753917694092, "learning_rate": 1e-05, "loss": 0.3951, "mean_token_accuracy": 0.8920252323150635, "num_tokens": 705864940.0, "step": 18589 }, { "epoch": 2.364839078997583, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003204345703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6058201789855957, "learning_rate": 1e-05, "loss": 0.3716, "mean_token_accuracy": 0.8952213525772095, "num_tokens": 705898274.0, "step": 18590 }, { "epoch": 2.3649662892761736, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.610330581665039, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8768405318260193, "num_tokens": 705936619.0, "step": 18591 }, { "epoch": 2.365093499554764, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.531278133392334, "learning_rate": 1e-05, "loss": 0.3562, "mean_token_accuracy": 0.9028502106666565, "num_tokens": 705975208.0, "step": 18592 }, { "epoch": 2.3652207098333546, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5665488243103027, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8711011409759521, "num_tokens": 706019948.0, "step": 18593 }, { "epoch": 2.365347920111945, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.476940393447876, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8820450305938721, "num_tokens": 706056862.0, "step": 18594 }, { "epoch": 2.3654751303905357, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.56129789352417, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8749299049377441, "num_tokens": 706092591.0, "step": 18595 }, { "epoch": 2.3656023406691262, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4223105907440186, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8774707317352295, "num_tokens": 706132572.0, "step": 18596 }, { "epoch": 2.3657295509477168, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.572324752807617, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8820867538452148, "num_tokens": 706168264.0, "step": 18597 }, { "epoch": 2.365856761226307, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.358994960784912, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8789708614349365, "num_tokens": 706206684.0, "step": 18598 }, { "epoch": 2.365983971504898, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6284565925598145, "learning_rate": 1e-05, "loss": 0.3808, "mean_token_accuracy": 0.8943394422531128, "num_tokens": 706247547.0, "step": 18599 }, { "epoch": 2.366111181783488, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2876436710357666, "learning_rate": 1e-05, "loss": 0.3816, "mean_token_accuracy": 0.8933665752410889, "num_tokens": 706286286.0, "step": 18600 }, { "epoch": 2.3662383920620784, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.729790449142456, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8734645843505859, "num_tokens": 706329990.0, "step": 18601 }, { "epoch": 2.366365602340669, "ewc_loss": 0.078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2164711952209473, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8759959936141968, "num_tokens": 706370423.0, "step": 18602 }, { "epoch": 2.3664928126192595, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.9004080295562744, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8826925158500671, "num_tokens": 706414380.0, "step": 18603 }, { "epoch": 2.36662002289785, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1771841049194336, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8796190023422241, "num_tokens": 706455403.0, "step": 18604 }, { "epoch": 2.3667472331764405, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.139806747436523, "learning_rate": 1e-05, "loss": 0.3777, "mean_token_accuracy": 0.8988941311836243, "num_tokens": 706490114.0, "step": 18605 }, { "epoch": 2.366874443455031, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.27390193939209, "learning_rate": 1e-05, "loss": 0.3837, "mean_token_accuracy": 0.8917326927185059, "num_tokens": 706523926.0, "step": 18606 }, { "epoch": 2.3670016537336216, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.3293776512146, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8932410478591919, "num_tokens": 706562626.0, "step": 18607 }, { "epoch": 2.367128864012212, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5169577598571777, "learning_rate": 1e-05, "loss": 0.3934, "mean_token_accuracy": 0.8928183317184448, "num_tokens": 706598735.0, "step": 18608 }, { "epoch": 2.3672560742908026, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.33820915222168, "learning_rate": 1e-05, "loss": 0.4796, "mean_token_accuracy": 0.8677628040313721, "num_tokens": 706638399.0, "step": 18609 }, { "epoch": 2.367383284569393, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7358157634735107, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8792599439620972, "num_tokens": 706680274.0, "step": 18610 }, { "epoch": 2.3675104948479837, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.168320655822754, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8653046488761902, "num_tokens": 706713918.0, "step": 18611 }, { "epoch": 2.3676377051265742, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7856082916259766, "learning_rate": 1e-05, "loss": 0.4418, "mean_token_accuracy": 0.8776856660842896, "num_tokens": 706751758.0, "step": 18612 }, { "epoch": 2.3677649154051648, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.949742555618286, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.889782190322876, "num_tokens": 706788243.0, "step": 18613 }, { "epoch": 2.3678921256837553, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.709296703338623, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8884005546569824, "num_tokens": 706821960.0, "step": 18614 }, { "epoch": 2.368019335962346, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7825193405151367, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.883281409740448, "num_tokens": 706858768.0, "step": 18615 }, { "epoch": 2.3681465462409363, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6163787841796875, "learning_rate": 1e-05, "loss": 0.3845, "mean_token_accuracy": 0.8934522867202759, "num_tokens": 706898889.0, "step": 18616 }, { "epoch": 2.368273756519527, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6821932792663574, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8794112205505371, "num_tokens": 706934245.0, "step": 18617 }, { "epoch": 2.3684009667981174, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.52278995513916, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8752601146697998, "num_tokens": 706982092.0, "step": 18618 }, { "epoch": 2.368528177076708, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.654330253601074, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8861202001571655, "num_tokens": 707014748.0, "step": 18619 }, { "epoch": 2.3686553873552985, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4547011852264404, "learning_rate": 1e-05, "loss": 0.405, "mean_token_accuracy": 0.8839691281318665, "num_tokens": 707048363.0, "step": 18620 }, { "epoch": 2.368782597633889, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6606087684631348, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8729103803634644, "num_tokens": 707084089.0, "step": 18621 }, { "epoch": 2.3689098079124795, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3693459033966064, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8710708618164062, "num_tokens": 707127334.0, "step": 18622 }, { "epoch": 2.3690370181910696, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7146458625793457, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8827029466629028, "num_tokens": 707166485.0, "step": 18623 }, { "epoch": 2.3691642284696606, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.30966854095459, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8823691010475159, "num_tokens": 707202031.0, "step": 18624 }, { "epoch": 2.3692914387482507, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.821286201477051, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8825026750564575, "num_tokens": 707240649.0, "step": 18625 }, { "epoch": 2.369418649026841, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2476541996002197, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8797642588615417, "num_tokens": 707275996.0, "step": 18626 }, { "epoch": 2.3695458593054317, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.993642807006836, "learning_rate": 1e-05, "loss": 0.4922, "mean_token_accuracy": 0.8646538257598877, "num_tokens": 707312776.0, "step": 18627 }, { "epoch": 2.3696730695840222, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2667787075042725, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8738952875137329, "num_tokens": 707349519.0, "step": 18628 }, { "epoch": 2.3698002798626128, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.146601676940918, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8773646354675293, "num_tokens": 707389180.0, "step": 18629 }, { "epoch": 2.3699274901412033, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3880136013031006, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8823608160018921, "num_tokens": 707428603.0, "step": 18630 }, { "epoch": 2.370054700419794, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.2171101570129395, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8684542775154114, "num_tokens": 707470297.0, "step": 18631 }, { "epoch": 2.3701819106983844, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.587374448776245, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8804754614830017, "num_tokens": 707507353.0, "step": 18632 }, { "epoch": 2.370309120976975, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.116944789886475, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8877390623092651, "num_tokens": 707541795.0, "step": 18633 }, { "epoch": 2.3704363312555654, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6883609294891357, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8793802261352539, "num_tokens": 707577333.0, "step": 18634 }, { "epoch": 2.370563541534156, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9194977283477783, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8684677481651306, "num_tokens": 707611306.0, "step": 18635 }, { "epoch": 2.3706907518127465, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6777243614196777, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8739529848098755, "num_tokens": 707647698.0, "step": 18636 }, { "epoch": 2.370817962091337, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7196874618530273, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8840917348861694, "num_tokens": 707683934.0, "step": 18637 }, { "epoch": 2.3709451723699275, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6425435543060303, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8738565444946289, "num_tokens": 707723216.0, "step": 18638 }, { "epoch": 2.371072382648518, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.569718599319458, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8832306861877441, "num_tokens": 707761194.0, "step": 18639 }, { "epoch": 2.3711995929271086, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6006078720092773, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8637177348136902, "num_tokens": 707807327.0, "step": 18640 }, { "epoch": 2.371326803205699, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.479198455810547, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8799458742141724, "num_tokens": 707846394.0, "step": 18641 }, { "epoch": 2.3714540134842896, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5782763957977295, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8756580352783203, "num_tokens": 707889099.0, "step": 18642 }, { "epoch": 2.37158122376288, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.425095796585083, "learning_rate": 1e-05, "loss": 0.3899, "mean_token_accuracy": 0.8869500160217285, "num_tokens": 707924729.0, "step": 18643 }, { "epoch": 2.3717084340414707, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.58235239982605, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8570651412010193, "num_tokens": 707965096.0, "step": 18644 }, { "epoch": 2.371835644320061, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.376004695892334, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.8812158703804016, "num_tokens": 708002277.0, "step": 18645 }, { "epoch": 2.3719628545986517, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.615203380584717, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8736927509307861, "num_tokens": 708038978.0, "step": 18646 }, { "epoch": 2.3720900648772423, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.30540132522583, "learning_rate": 1e-05, "loss": 0.3822, "mean_token_accuracy": 0.893356442451477, "num_tokens": 708076022.0, "step": 18647 }, { "epoch": 2.3722172751558324, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6857402324676514, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8661631941795349, "num_tokens": 708119606.0, "step": 18648 }, { "epoch": 2.3723444854344233, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2612287998199463, "learning_rate": 1e-05, "loss": 0.3864, "mean_token_accuracy": 0.8900594711303711, "num_tokens": 708159570.0, "step": 18649 }, { "epoch": 2.3724716957130134, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8009424209594727, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8641237616539001, "num_tokens": 708201092.0, "step": 18650 }, { "epoch": 2.372598905991604, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.182161569595337, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8827184438705444, "num_tokens": 708235808.0, "step": 18651 }, { "epoch": 2.3727261162701945, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.993535280227661, "learning_rate": 1e-05, "loss": 0.3878, "mean_token_accuracy": 0.8986403346061707, "num_tokens": 708270863.0, "step": 18652 }, { "epoch": 2.372853326548785, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.177537441253662, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8661978244781494, "num_tokens": 708309653.0, "step": 18653 }, { "epoch": 2.3729805368273755, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.351521015167236, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8830546736717224, "num_tokens": 708346242.0, "step": 18654 }, { "epoch": 2.373107747105966, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.394972562789917, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8724424242973328, "num_tokens": 708383391.0, "step": 18655 }, { "epoch": 2.3732349573845566, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.427392482757568, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8874304294586182, "num_tokens": 708417094.0, "step": 18656 }, { "epoch": 2.373362167663147, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.6798901557922363, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8772342205047607, "num_tokens": 708456648.0, "step": 18657 }, { "epoch": 2.3734893779417376, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.366299152374268, "learning_rate": 1e-05, "loss": 0.4702, "mean_token_accuracy": 0.8695997595787048, "num_tokens": 708499125.0, "step": 18658 }, { "epoch": 2.373616588220328, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.836897611618042, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8820693492889404, "num_tokens": 708544844.0, "step": 18659 }, { "epoch": 2.3737437984989187, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1104207038879395, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8767533302307129, "num_tokens": 708586852.0, "step": 18660 }, { "epoch": 2.3738710087775092, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.8048932552337646, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8798934817314148, "num_tokens": 708624108.0, "step": 18661 }, { "epoch": 2.3739982190560998, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.917459011077881, "learning_rate": 1e-05, "loss": 0.3803, "mean_token_accuracy": 0.8977894186973572, "num_tokens": 708658214.0, "step": 18662 }, { "epoch": 2.3741254293346903, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.695119619369507, "learning_rate": 1e-05, "loss": 0.3937, "mean_token_accuracy": 0.8939300775527954, "num_tokens": 708695315.0, "step": 18663 }, { "epoch": 2.374252639613281, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7916738986968994, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8704955577850342, "num_tokens": 708734034.0, "step": 18664 }, { "epoch": 2.3743798498918713, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.599388599395752, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8778473138809204, "num_tokens": 708769758.0, "step": 18665 }, { "epoch": 2.374507060170462, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7249131202697754, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8689243197441101, "num_tokens": 708803639.0, "step": 18666 }, { "epoch": 2.3746342704490524, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5083131790161133, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8752156496047974, "num_tokens": 708837085.0, "step": 18667 }, { "epoch": 2.374761480727643, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.68454647064209, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8811855912208557, "num_tokens": 708876102.0, "step": 18668 }, { "epoch": 2.3748886910062335, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.409184455871582, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8901901245117188, "num_tokens": 708914260.0, "step": 18669 }, { "epoch": 2.375015901284824, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.727454662322998, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8677946329116821, "num_tokens": 708952071.0, "step": 18670 }, { "epoch": 2.3751431115634145, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.333327531814575, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8724834322929382, "num_tokens": 708989394.0, "step": 18671 }, { "epoch": 2.375270321842005, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8150405883789062, "learning_rate": 1e-05, "loss": 0.5163, "mean_token_accuracy": 0.8553614616394043, "num_tokens": 709033558.0, "step": 18672 }, { "epoch": 2.375397532120595, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.277007579803467, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8761649131774902, "num_tokens": 709071420.0, "step": 18673 }, { "epoch": 2.375524742399186, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9614291191101074, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8883095979690552, "num_tokens": 709104810.0, "step": 18674 }, { "epoch": 2.375651952677776, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2797908782958984, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8730194568634033, "num_tokens": 709142789.0, "step": 18675 }, { "epoch": 2.3757791629563667, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1156487464904785, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8882417678833008, "num_tokens": 709177590.0, "step": 18676 }, { "epoch": 2.3759063732349572, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3865246772766113, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8868628740310669, "num_tokens": 709217177.0, "step": 18677 }, { "epoch": 2.3760335835135478, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.199995994567871, "learning_rate": 1e-05, "loss": 0.5387, "mean_token_accuracy": 0.8492856621742249, "num_tokens": 709252566.0, "step": 18678 }, { "epoch": 2.3761607937921383, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5719826221466064, "learning_rate": 1e-05, "loss": 0.3859, "mean_token_accuracy": 0.891811192035675, "num_tokens": 709290386.0, "step": 18679 }, { "epoch": 2.376288004070729, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.088813304901123, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8782234191894531, "num_tokens": 709331110.0, "step": 18680 }, { "epoch": 2.3764152143493193, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.657637596130371, "learning_rate": 1e-05, "loss": 0.3997, "mean_token_accuracy": 0.8922713398933411, "num_tokens": 709371390.0, "step": 18681 }, { "epoch": 2.37654242462791, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.894029140472412, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.877493143081665, "num_tokens": 709413428.0, "step": 18682 }, { "epoch": 2.3766696349065004, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6850879192352295, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8817780017852783, "num_tokens": 709450332.0, "step": 18683 }, { "epoch": 2.376796845185091, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7022738456726074, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8578012585639954, "num_tokens": 709490039.0, "step": 18684 }, { "epoch": 2.3769240554636815, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6545450687408447, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8877654671669006, "num_tokens": 709531839.0, "step": 18685 }, { "epoch": 2.377051265742272, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.538212299346924, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.8830103874206543, "num_tokens": 709568126.0, "step": 18686 }, { "epoch": 2.3771784760208625, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6345105171203613, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8824992775917053, "num_tokens": 709610546.0, "step": 18687 }, { "epoch": 2.377305686299453, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.410999059677124, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8717889785766602, "num_tokens": 709656869.0, "step": 18688 }, { "epoch": 2.3774328965780436, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6575510501861572, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.8873317241668701, "num_tokens": 709695691.0, "step": 18689 }, { "epoch": 2.377560106856634, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.332454204559326, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8777906894683838, "num_tokens": 709736691.0, "step": 18690 }, { "epoch": 2.3776873171352246, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7169225215911865, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8738178014755249, "num_tokens": 709776277.0, "step": 18691 }, { "epoch": 2.377814527413815, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2699966430664062, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8705315589904785, "num_tokens": 709817222.0, "step": 18692 }, { "epoch": 2.3779417376924057, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.816419839859009, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8754105567932129, "num_tokens": 709859060.0, "step": 18693 }, { "epoch": 2.378068947970996, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2233035564422607, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8775364756584167, "num_tokens": 709897396.0, "step": 18694 }, { "epoch": 2.3781961582495867, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.978215456008911, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8819801807403564, "num_tokens": 709939797.0, "step": 18695 }, { "epoch": 2.378323368528177, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2263224124908447, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8758951425552368, "num_tokens": 709977102.0, "step": 18696 }, { "epoch": 2.378450578806768, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.172591209411621, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8759127259254456, "num_tokens": 710020824.0, "step": 18697 }, { "epoch": 2.378577789085358, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.364931344985962, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8806169033050537, "num_tokens": 710052263.0, "step": 18698 }, { "epoch": 2.3787049993639484, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.263488292694092, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8816325068473816, "num_tokens": 710095196.0, "step": 18699 }, { "epoch": 2.378832209642539, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.578458786010742, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8730760812759399, "num_tokens": 710127255.0, "step": 18700 }, { "epoch": 2.3789594199211295, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.201609134674072, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8779497146606445, "num_tokens": 710159129.0, "step": 18701 }, { "epoch": 2.37908663019972, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.687561273574829, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.877160906791687, "num_tokens": 710197244.0, "step": 18702 }, { "epoch": 2.3792138404783105, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.012310028076172, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.8625613451004028, "num_tokens": 710241133.0, "step": 18703 }, { "epoch": 2.379341050756901, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6832685470581055, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8817655444145203, "num_tokens": 710283264.0, "step": 18704 }, { "epoch": 2.3794682610354916, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.838921308517456, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8716074228286743, "num_tokens": 710329203.0, "step": 18705 }, { "epoch": 2.379595471314082, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.628714084625244, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8833743333816528, "num_tokens": 710366772.0, "step": 18706 }, { "epoch": 2.3797226815926726, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7084732055664062, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8765466809272766, "num_tokens": 710407458.0, "step": 18707 }, { "epoch": 2.379849891871263, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.556671142578125, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8814882636070251, "num_tokens": 710449200.0, "step": 18708 }, { "epoch": 2.3799771021498537, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.627490520477295, "learning_rate": 1e-05, "loss": 0.3752, "mean_token_accuracy": 0.8993093371391296, "num_tokens": 710490764.0, "step": 18709 }, { "epoch": 2.3801043124284442, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.495231866836548, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8789247274398804, "num_tokens": 710526282.0, "step": 18710 }, { "epoch": 2.3802315227070348, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.605116367340088, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.886703610420227, "num_tokens": 710566523.0, "step": 18711 }, { "epoch": 2.3803587329856253, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4288134574890137, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8828778862953186, "num_tokens": 710607640.0, "step": 18712 }, { "epoch": 2.380485943264216, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6252307891845703, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8768622279167175, "num_tokens": 710644785.0, "step": 18713 }, { "epoch": 2.3806131535428063, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.370375633239746, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8819754123687744, "num_tokens": 710681857.0, "step": 18714 }, { "epoch": 2.380740363821397, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6839895248413086, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8871784210205078, "num_tokens": 710718119.0, "step": 18715 }, { "epoch": 2.3808675740999874, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.303204298019409, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8667865991592407, "num_tokens": 710758573.0, "step": 18716 }, { "epoch": 2.380994784378578, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.790337324142456, "learning_rate": 1e-05, "loss": 0.3956, "mean_token_accuracy": 0.8919152617454529, "num_tokens": 710791567.0, "step": 18717 }, { "epoch": 2.3811219946571685, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2381255626678467, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8852852582931519, "num_tokens": 710833060.0, "step": 18718 }, { "epoch": 2.381249204935759, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.954037666320801, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.874342679977417, "num_tokens": 710872175.0, "step": 18719 }, { "epoch": 2.3813764152143495, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2375731468200684, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8830102682113647, "num_tokens": 710910442.0, "step": 18720 }, { "epoch": 2.3815036254929396, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.134493827819824, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8801646828651428, "num_tokens": 710948656.0, "step": 18721 }, { "epoch": 2.3816308357715306, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.357576847076416, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8757907152175903, "num_tokens": 710989300.0, "step": 18722 }, { "epoch": 2.3817580460501206, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.236718654632568, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8917132616043091, "num_tokens": 711027661.0, "step": 18723 }, { "epoch": 2.381885256328711, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5546886920928955, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8698357343673706, "num_tokens": 711067895.0, "step": 18724 }, { "epoch": 2.3820124666073017, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1565117835998535, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.875169038772583, "num_tokens": 711105567.0, "step": 18725 }, { "epoch": 2.3821396768858922, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.671027183532715, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.8926951885223389, "num_tokens": 711145263.0, "step": 18726 }, { "epoch": 2.3822668871644828, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9740490913391113, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8840221166610718, "num_tokens": 711187024.0, "step": 18727 }, { "epoch": 2.3823940974430733, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.684335708618164, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8776735663414001, "num_tokens": 711226261.0, "step": 18728 }, { "epoch": 2.382521307721664, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.77670955657959, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.872074544429779, "num_tokens": 711267005.0, "step": 18729 }, { "epoch": 2.3826485180002543, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.644150972366333, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8722714185714722, "num_tokens": 711307145.0, "step": 18730 }, { "epoch": 2.382775728278845, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6254289150238037, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8834733963012695, "num_tokens": 711351128.0, "step": 18731 }, { "epoch": 2.3829029385574354, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6080050468444824, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8859502673149109, "num_tokens": 711385344.0, "step": 18732 }, { "epoch": 2.383030148836026, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5308895111083984, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8748663663864136, "num_tokens": 711421548.0, "step": 18733 }, { "epoch": 2.3831573591146165, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5651967525482178, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8821520805358887, "num_tokens": 711456952.0, "step": 18734 }, { "epoch": 2.383284569393207, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.478785753250122, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8830959796905518, "num_tokens": 711486985.0, "step": 18735 }, { "epoch": 2.3834117796717975, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.541128396987915, "learning_rate": 1e-05, "loss": 0.3918, "mean_token_accuracy": 0.8952307105064392, "num_tokens": 711521047.0, "step": 18736 }, { "epoch": 2.383538989950388, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4296200275421143, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.8874521255493164, "num_tokens": 711561151.0, "step": 18737 }, { "epoch": 2.3836662002289786, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5390024185180664, "learning_rate": 1e-05, "loss": 0.4824, "mean_token_accuracy": 0.8611364960670471, "num_tokens": 711597975.0, "step": 18738 }, { "epoch": 2.383793410507569, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3873305320739746, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8777656555175781, "num_tokens": 711637923.0, "step": 18739 }, { "epoch": 2.3839206207861596, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5641262531280518, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8770287036895752, "num_tokens": 711675408.0, "step": 18740 }, { "epoch": 2.38404783106475, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3368921279907227, "learning_rate": 1e-05, "loss": 0.3991, "mean_token_accuracy": 0.8863343000411987, "num_tokens": 711712781.0, "step": 18741 }, { "epoch": 2.3841750413433407, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.613863945007324, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8801567554473877, "num_tokens": 711752974.0, "step": 18742 }, { "epoch": 2.384302251621931, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2754077911376953, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8809471726417542, "num_tokens": 711793526.0, "step": 18743 }, { "epoch": 2.3844294619005217, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7084908485412598, "learning_rate": 1e-05, "loss": 0.3796, "mean_token_accuracy": 0.8947884440422058, "num_tokens": 711830409.0, "step": 18744 }, { "epoch": 2.3845566721791123, "ewc_loss": 0.07861328125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.649162292480469e-05, "grad_norm": 3.1950626373291016, "learning_rate": 1e-05, "loss": 0.3932, "mean_token_accuracy": 0.8888390064239502, "num_tokens": 711873476.0, "step": 18745 }, { "epoch": 2.3846838824577024, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8916914463043213, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8884021043777466, "num_tokens": 711912507.0, "step": 18746 }, { "epoch": 2.3848110927362933, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.161348819732666, "learning_rate": 1e-05, "loss": 0.3951, "mean_token_accuracy": 0.8888724446296692, "num_tokens": 711946884.0, "step": 18747 }, { "epoch": 2.3849383030148834, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.140595436096191, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8690544962882996, "num_tokens": 711987060.0, "step": 18748 }, { "epoch": 2.385065513293474, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.2079520225524902, "learning_rate": 1e-05, "loss": 0.3783, "mean_token_accuracy": 0.8957020044326782, "num_tokens": 712023197.0, "step": 18749 }, { "epoch": 2.3851927235720645, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.398436546325684, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8909905552864075, "num_tokens": 712060337.0, "step": 18750 }, { "epoch": 2.385319933850655, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.498303174972534, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.8898279666900635, "num_tokens": 712101601.0, "step": 18751 }, { "epoch": 2.3854471441292455, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.458507061004639, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8882482051849365, "num_tokens": 712136311.0, "step": 18752 }, { "epoch": 2.385574354407836, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.8149876594543457, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8720874786376953, "num_tokens": 712174386.0, "step": 18753 }, { "epoch": 2.3857015646864266, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.248219966888428, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8835421800613403, "num_tokens": 712209540.0, "step": 18754 }, { "epoch": 2.385828774965017, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.8633131980895996, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8722365498542786, "num_tokens": 712245405.0, "step": 18755 }, { "epoch": 2.3859559852436076, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.025064945220947, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8597780466079712, "num_tokens": 712285013.0, "step": 18756 }, { "epoch": 2.386083195522198, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.766718864440918, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8624985814094543, "num_tokens": 712324850.0, "step": 18757 }, { "epoch": 2.3862104058007887, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8458526134490967, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8758789300918579, "num_tokens": 712364336.0, "step": 18758 }, { "epoch": 2.386337616079379, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6643130779266357, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8648090362548828, "num_tokens": 712403773.0, "step": 18759 }, { "epoch": 2.3864648263579697, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7445168495178223, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8664937615394592, "num_tokens": 712440307.0, "step": 18760 }, { "epoch": 2.3865920366365603, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5451998710632324, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.882654070854187, "num_tokens": 712472008.0, "step": 18761 }, { "epoch": 2.386719246915151, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.688870429992676, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.886039137840271, "num_tokens": 712512935.0, "step": 18762 }, { "epoch": 2.3868464571937413, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.441244602203369, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.888075590133667, "num_tokens": 712552783.0, "step": 18763 }, { "epoch": 2.386973667472332, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.720689058303833, "learning_rate": 1e-05, "loss": 0.3649, "mean_token_accuracy": 0.8989939093589783, "num_tokens": 712583752.0, "step": 18764 }, { "epoch": 2.3871008777509224, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.365201473236084, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8678101301193237, "num_tokens": 712623125.0, "step": 18765 }, { "epoch": 2.387228088029513, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.797504425048828, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8719353079795837, "num_tokens": 712664965.0, "step": 18766 }, { "epoch": 2.3873552983081034, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2845818996429443, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8757076859474182, "num_tokens": 712703571.0, "step": 18767 }, { "epoch": 2.387482508586694, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9352643489837646, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8815691471099854, "num_tokens": 712739830.0, "step": 18768 }, { "epoch": 2.3876097188652845, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2561404705047607, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8733499050140381, "num_tokens": 712782367.0, "step": 18769 }, { "epoch": 2.387736929143875, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.138164043426514, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8827937245368958, "num_tokens": 712816937.0, "step": 18770 }, { "epoch": 2.387864139422465, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.372192621231079, "learning_rate": 1e-05, "loss": 0.397, "mean_token_accuracy": 0.8898307085037231, "num_tokens": 712850198.0, "step": 18771 }, { "epoch": 2.387991349701056, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.245366096496582, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8814053535461426, "num_tokens": 712886201.0, "step": 18772 }, { "epoch": 2.388118559979646, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5799527168273926, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8736042380332947, "num_tokens": 712924309.0, "step": 18773 }, { "epoch": 2.3882457702582367, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.14496374130249, "learning_rate": 1e-05, "loss": 0.3981, "mean_token_accuracy": 0.8946269750595093, "num_tokens": 712964508.0, "step": 18774 }, { "epoch": 2.3883729805368272, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.697953701019287, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8723735809326172, "num_tokens": 713002737.0, "step": 18775 }, { "epoch": 2.3885001908154178, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.945922374725342, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8801590800285339, "num_tokens": 713040859.0, "step": 18776 }, { "epoch": 2.3886274010940083, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.7098429203033447, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8830491900444031, "num_tokens": 713073504.0, "step": 18777 }, { "epoch": 2.388754611372599, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7403597831726074, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8822099566459656, "num_tokens": 713110285.0, "step": 18778 }, { "epoch": 2.3888818216511893, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6762685775756836, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8731162548065186, "num_tokens": 713146227.0, "step": 18779 }, { "epoch": 2.38900903192978, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5711474418640137, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8760169744491577, "num_tokens": 713187632.0, "step": 18780 }, { "epoch": 2.3891362422083704, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.657876491546631, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8768444061279297, "num_tokens": 713223023.0, "step": 18781 }, { "epoch": 2.389263452486961, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4507381916046143, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8707289099693298, "num_tokens": 713264390.0, "step": 18782 }, { "epoch": 2.3893906627655515, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6484906673431396, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.8871064186096191, "num_tokens": 713306504.0, "step": 18783 }, { "epoch": 2.389517873044142, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3681676387786865, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8807003498077393, "num_tokens": 713347556.0, "step": 18784 }, { "epoch": 2.3896450833227325, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6993603706359863, "learning_rate": 1e-05, "loss": 0.5021, "mean_token_accuracy": 0.8538111448287964, "num_tokens": 713386645.0, "step": 18785 }, { "epoch": 2.389772293601323, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2949612140655518, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8844367265701294, "num_tokens": 713419754.0, "step": 18786 }, { "epoch": 2.3898995038799136, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.792607307434082, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.879199743270874, "num_tokens": 713456374.0, "step": 18787 }, { "epoch": 2.390026714158504, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2389774322509766, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.882314145565033, "num_tokens": 713487961.0, "step": 18788 }, { "epoch": 2.3901539244370946, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9346256256103516, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8782062530517578, "num_tokens": 713525697.0, "step": 18789 }, { "epoch": 2.390281134715685, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2143335342407227, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8784021139144897, "num_tokens": 713563902.0, "step": 18790 }, { "epoch": 2.3904083449942757, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.13020658493042, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8694510459899902, "num_tokens": 713602996.0, "step": 18791 }, { "epoch": 2.390535555272866, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.307934284210205, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8792116641998291, "num_tokens": 713641725.0, "step": 18792 }, { "epoch": 2.3906627655514567, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2898736000061035, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.857742428779602, "num_tokens": 713682973.0, "step": 18793 }, { "epoch": 2.390789975830047, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.526553153991699, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8854002952575684, "num_tokens": 713718919.0, "step": 18794 }, { "epoch": 2.390917186108638, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.263474941253662, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8771092295646667, "num_tokens": 713757416.0, "step": 18795 }, { "epoch": 2.391044396387228, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7073729038238525, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.863483726978302, "num_tokens": 713795648.0, "step": 18796 }, { "epoch": 2.3911716066658184, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.071750640869141, "learning_rate": 1e-05, "loss": 0.3892, "mean_token_accuracy": 0.8969756960868835, "num_tokens": 713830095.0, "step": 18797 }, { "epoch": 2.391298816944409, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7254111766815186, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8714373707771301, "num_tokens": 713868958.0, "step": 18798 }, { "epoch": 2.3914260272229995, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.8821053504943848, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8682234287261963, "num_tokens": 713911533.0, "step": 18799 }, { "epoch": 2.39155323750159, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.677013635635376, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8719991445541382, "num_tokens": 713952585.0, "step": 18800 }, { "epoch": 2.3916804477801805, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7176010608673096, "learning_rate": 1e-05, "loss": 0.3869, "mean_token_accuracy": 0.8951111435890198, "num_tokens": 713990929.0, "step": 18801 }, { "epoch": 2.391807658058771, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6044886112213135, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8692569136619568, "num_tokens": 714027098.0, "step": 18802 }, { "epoch": 2.3919348683373616, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6296935081481934, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8709752559661865, "num_tokens": 714062444.0, "step": 18803 }, { "epoch": 2.392062078615952, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5278468132019043, "learning_rate": 1e-05, "loss": 0.3752, "mean_token_accuracy": 0.8961607217788696, "num_tokens": 714099091.0, "step": 18804 }, { "epoch": 2.3921892888945426, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.580845594406128, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8728147745132446, "num_tokens": 714135694.0, "step": 18805 }, { "epoch": 2.392316499173133, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4710731506347656, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.8914918303489685, "num_tokens": 714170324.0, "step": 18806 }, { "epoch": 2.3924437094517237, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5753509998321533, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8811830878257751, "num_tokens": 714205080.0, "step": 18807 }, { "epoch": 2.392570919730314, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.397073984146118, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8770934343338013, "num_tokens": 714245297.0, "step": 18808 }, { "epoch": 2.3926981300089047, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6017470359802246, "learning_rate": 1e-05, "loss": 0.3932, "mean_token_accuracy": 0.8913171291351318, "num_tokens": 714284155.0, "step": 18809 }, { "epoch": 2.3928253402874953, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.342777729034424, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8845624923706055, "num_tokens": 714318688.0, "step": 18810 }, { "epoch": 2.392952550566086, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6603808403015137, "learning_rate": 1e-05, "loss": 0.3763, "mean_token_accuracy": 0.8949910402297974, "num_tokens": 714353730.0, "step": 18811 }, { "epoch": 2.3930797608446763, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2795236110687256, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.877950131893158, "num_tokens": 714390156.0, "step": 18812 }, { "epoch": 2.393206971123267, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7810616493225098, "learning_rate": 1e-05, "loss": 0.468, "mean_token_accuracy": 0.8643279075622559, "num_tokens": 714429435.0, "step": 18813 }, { "epoch": 2.3933341814018574, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2090389728546143, "learning_rate": 1e-05, "loss": 0.3692, "mean_token_accuracy": 0.8977693319320679, "num_tokens": 714470034.0, "step": 18814 }, { "epoch": 2.393461391680448, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9665095806121826, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8769198656082153, "num_tokens": 714505989.0, "step": 18815 }, { "epoch": 2.3935886019590384, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2078123092651367, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8730838298797607, "num_tokens": 714541727.0, "step": 18816 }, { "epoch": 2.393715812237629, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.190530300140381, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8814092874526978, "num_tokens": 714581429.0, "step": 18817 }, { "epoch": 2.3938430225162195, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3328166007995605, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.878645658493042, "num_tokens": 714621839.0, "step": 18818 }, { "epoch": 2.3939702327948096, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.324146747589111, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8769539594650269, "num_tokens": 714665459.0, "step": 18819 }, { "epoch": 2.3940974430734006, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5998942852020264, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.8855569362640381, "num_tokens": 714699259.0, "step": 18820 }, { "epoch": 2.3942246533519906, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.237995147705078, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.8708767890930176, "num_tokens": 714740330.0, "step": 18821 }, { "epoch": 2.394351863630581, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7423439025878906, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.8829060196876526, "num_tokens": 714773330.0, "step": 18822 }, { "epoch": 2.3944790739091717, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0261759757995605, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.870119571685791, "num_tokens": 714812707.0, "step": 18823 }, { "epoch": 2.3946062841877622, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.721449375152588, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8829107284545898, "num_tokens": 714853496.0, "step": 18824 }, { "epoch": 2.3947334944663528, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8353495597839355, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8838824033737183, "num_tokens": 714895766.0, "step": 18825 }, { "epoch": 2.3948607047449433, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6466686725616455, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8812835812568665, "num_tokens": 714936089.0, "step": 18826 }, { "epoch": 2.394987915023534, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7034683227539062, "learning_rate": 1e-05, "loss": 0.3859, "mean_token_accuracy": 0.8939365148544312, "num_tokens": 714980759.0, "step": 18827 }, { "epoch": 2.3951151253021243, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.571023941040039, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8847122192382812, "num_tokens": 715019671.0, "step": 18828 }, { "epoch": 2.395242335580715, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.637897491455078, "learning_rate": 1e-05, "loss": 0.4883, "mean_token_accuracy": 0.8610250949859619, "num_tokens": 715059639.0, "step": 18829 }, { "epoch": 2.3953695458593054, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4901418685913086, "learning_rate": 1e-05, "loss": 0.395, "mean_token_accuracy": 0.891097366809845, "num_tokens": 715096315.0, "step": 18830 }, { "epoch": 2.395496756137896, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6188502311706543, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8796391487121582, "num_tokens": 715130800.0, "step": 18831 }, { "epoch": 2.3956239664164865, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4100708961486816, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8739544153213501, "num_tokens": 715175369.0, "step": 18832 }, { "epoch": 2.395751176695077, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6418628692626953, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8751676082611084, "num_tokens": 715216480.0, "step": 18833 }, { "epoch": 2.3958783869736675, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3357341289520264, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.886249303817749, "num_tokens": 715257182.0, "step": 18834 }, { "epoch": 2.396005597252258, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7525434494018555, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.874474287033081, "num_tokens": 715288793.0, "step": 18835 }, { "epoch": 2.3961328075308486, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2722997665405273, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8776051998138428, "num_tokens": 715323720.0, "step": 18836 }, { "epoch": 2.396260017809439, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 4.768461227416992, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8724962472915649, "num_tokens": 715356104.0, "step": 18837 }, { "epoch": 2.3963872280880296, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5134024620056152, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8881131410598755, "num_tokens": 715390727.0, "step": 18838 }, { "epoch": 2.39651443836662, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.250861167907715, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8830527663230896, "num_tokens": 715426340.0, "step": 18839 }, { "epoch": 2.3966416486452107, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.54740834236145, "learning_rate": 1e-05, "loss": 0.3517, "mean_token_accuracy": 0.905077338218689, "num_tokens": 715458195.0, "step": 18840 }, { "epoch": 2.396768858923801, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.325839996337891, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8909124135971069, "num_tokens": 715496896.0, "step": 18841 }, { "epoch": 2.3968960692023917, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7164807319641113, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8665590882301331, "num_tokens": 715537985.0, "step": 18842 }, { "epoch": 2.3970232794809823, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.178779125213623, "learning_rate": 1e-05, "loss": 0.3837, "mean_token_accuracy": 0.8939894437789917, "num_tokens": 715574767.0, "step": 18843 }, { "epoch": 2.3971504897595723, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003223419189453125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.761887311935425, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8778052926063538, "num_tokens": 715608649.0, "step": 18844 }, { "epoch": 2.3972777000381633, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.018815517425537, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8793355226516724, "num_tokens": 715650278.0, "step": 18845 }, { "epoch": 2.3974049103167534, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.693488836288452, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8750934600830078, "num_tokens": 715692678.0, "step": 18846 }, { "epoch": 2.397532120595344, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.889822244644165, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8701843023300171, "num_tokens": 715733128.0, "step": 18847 }, { "epoch": 2.3976593308739345, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6299257278442383, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8768244385719299, "num_tokens": 715767272.0, "step": 18848 }, { "epoch": 2.397786541152525, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.791029691696167, "learning_rate": 1e-05, "loss": 0.3688, "mean_token_accuracy": 0.8982623815536499, "num_tokens": 715803497.0, "step": 18849 }, { "epoch": 2.3979137514311155, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.513227701187134, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8833688497543335, "num_tokens": 715847069.0, "step": 18850 }, { "epoch": 2.398040961709706, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7781918048858643, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8846648931503296, "num_tokens": 715887639.0, "step": 18851 }, { "epoch": 2.3981681719882966, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4391582012176514, "learning_rate": 1e-05, "loss": 0.3607, "mean_token_accuracy": 0.8991621732711792, "num_tokens": 715919427.0, "step": 18852 }, { "epoch": 2.398295382266887, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8197154998779297, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.865269124507904, "num_tokens": 715961287.0, "step": 18853 }, { "epoch": 2.3984225925454776, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.373497724533081, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8760325908660889, "num_tokens": 715995021.0, "step": 18854 }, { "epoch": 2.398549802824068, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8848204612731934, "learning_rate": 1e-05, "loss": 0.3948, "mean_token_accuracy": 0.8938796520233154, "num_tokens": 716039517.0, "step": 18855 }, { "epoch": 2.3986770131026587, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3384711742401123, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8643009066581726, "num_tokens": 716077116.0, "step": 18856 }, { "epoch": 2.398804223381249, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.012680530548096, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8728383183479309, "num_tokens": 716112800.0, "step": 18857 }, { "epoch": 2.3989314336598397, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.366999387741089, "learning_rate": 1e-05, "loss": 0.3971, "mean_token_accuracy": 0.891102135181427, "num_tokens": 716148165.0, "step": 18858 }, { "epoch": 2.3990586439384303, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.10539436340332, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8737447261810303, "num_tokens": 716185745.0, "step": 18859 }, { "epoch": 2.399185854217021, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4738528728485107, "learning_rate": 1e-05, "loss": 0.3848, "mean_token_accuracy": 0.893114447593689, "num_tokens": 716225177.0, "step": 18860 }, { "epoch": 2.3993130644956113, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.064630508422852, "learning_rate": 1e-05, "loss": 0.3881, "mean_token_accuracy": 0.8976564407348633, "num_tokens": 716265969.0, "step": 18861 }, { "epoch": 2.399440274774202, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5991456508636475, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8795055150985718, "num_tokens": 716302056.0, "step": 18862 }, { "epoch": 2.3995674850527924, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.922499656677246, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.868874728679657, "num_tokens": 716339452.0, "step": 18863 }, { "epoch": 2.399694695331383, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6384146213531494, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8810044527053833, "num_tokens": 716379930.0, "step": 18864 }, { "epoch": 2.3998219056099734, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.723940849304199, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8762309551239014, "num_tokens": 716416367.0, "step": 18865 }, { "epoch": 2.399949115888564, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6413238048553467, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8645293116569519, "num_tokens": 716460112.0, "step": 18866 }, { "epoch": 2.4000763261671545, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5497779846191406, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.874599814414978, "num_tokens": 716497085.0, "step": 18867 }, { "epoch": 2.400203536445745, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.628291368484497, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8808900117874146, "num_tokens": 716542892.0, "step": 18868 }, { "epoch": 2.400330746724335, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.42757248878479, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8673586845397949, "num_tokens": 716578765.0, "step": 18869 }, { "epoch": 2.400457957002926, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6504173278808594, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8832899332046509, "num_tokens": 716618881.0, "step": 18870 }, { "epoch": 2.400585167281516, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.344158887863159, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8792425394058228, "num_tokens": 716651583.0, "step": 18871 }, { "epoch": 2.4007123775601067, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7134602069854736, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8683298826217651, "num_tokens": 716687118.0, "step": 18872 }, { "epoch": 2.400839587838697, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.274663209915161, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8699421882629395, "num_tokens": 716722543.0, "step": 18873 }, { "epoch": 2.4009667981172877, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8324358463287354, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8818627595901489, "num_tokens": 716757447.0, "step": 18874 }, { "epoch": 2.4010940083958783, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2189321517944336, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8780829906463623, "num_tokens": 716799736.0, "step": 18875 }, { "epoch": 2.401221218674469, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9769949913024902, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8759855628013611, "num_tokens": 716844850.0, "step": 18876 }, { "epoch": 2.4013484289530593, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.230090379714966, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8777707815170288, "num_tokens": 716878801.0, "step": 18877 }, { "epoch": 2.40147563923165, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.178918361663818, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.879115879535675, "num_tokens": 716914271.0, "step": 18878 }, { "epoch": 2.4016028495102404, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3485565185546875, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8772000074386597, "num_tokens": 716954094.0, "step": 18879 }, { "epoch": 2.401730059788831, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.269952297210693, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8830785751342773, "num_tokens": 716995061.0, "step": 18880 }, { "epoch": 2.4018572700674214, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.565671443939209, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8773806095123291, "num_tokens": 717032271.0, "step": 18881 }, { "epoch": 2.401984480346012, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.195943832397461, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8833783864974976, "num_tokens": 717068551.0, "step": 18882 }, { "epoch": 2.4021116906246025, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6983532905578613, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.882483184337616, "num_tokens": 717107460.0, "step": 18883 }, { "epoch": 2.402238900903193, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.021946430206299, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8751317262649536, "num_tokens": 717144774.0, "step": 18884 }, { "epoch": 2.4023661111817836, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7034363746643066, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8768243789672852, "num_tokens": 717180808.0, "step": 18885 }, { "epoch": 2.402493321460374, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8333404064178467, "learning_rate": 1e-05, "loss": 0.4121, "mean_token_accuracy": 0.8840199708938599, "num_tokens": 717220788.0, "step": 18886 }, { "epoch": 2.4026205317389646, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6453568935394287, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.884803295135498, "num_tokens": 717260257.0, "step": 18887 }, { "epoch": 2.402747742017555, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7111945152282715, "learning_rate": 1e-05, "loss": 0.4065, "mean_token_accuracy": 0.8856476545333862, "num_tokens": 717291911.0, "step": 18888 }, { "epoch": 2.4028749522961457, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5998048782348633, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8770281076431274, "num_tokens": 717322356.0, "step": 18889 }, { "epoch": 2.403002162574736, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6151349544525146, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8771520853042603, "num_tokens": 717363439.0, "step": 18890 }, { "epoch": 2.4031293728533267, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.524536609649658, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8761062622070312, "num_tokens": 717401050.0, "step": 18891 }, { "epoch": 2.403256583131917, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.568788528442383, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8766746520996094, "num_tokens": 717442346.0, "step": 18892 }, { "epoch": 2.403383793410508, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.467785596847534, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.8824299573898315, "num_tokens": 717483387.0, "step": 18893 }, { "epoch": 2.403511003689098, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5588622093200684, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8800961971282959, "num_tokens": 717522208.0, "step": 18894 }, { "epoch": 2.4036382139676884, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.420147180557251, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8899154663085938, "num_tokens": 717563338.0, "step": 18895 }, { "epoch": 2.403765424246279, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.575181007385254, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.87890625, "num_tokens": 717603351.0, "step": 18896 }, { "epoch": 2.4038926345248695, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3667516708374023, "learning_rate": 1e-05, "loss": 0.3823, "mean_token_accuracy": 0.8914439678192139, "num_tokens": 717639886.0, "step": 18897 }, { "epoch": 2.40401984480346, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.654801368713379, "learning_rate": 1e-05, "loss": 0.5042, "mean_token_accuracy": 0.8571922779083252, "num_tokens": 717672228.0, "step": 18898 }, { "epoch": 2.4041470550820505, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.292834758758545, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8706831932067871, "num_tokens": 717709703.0, "step": 18899 }, { "epoch": 2.404274265360641, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7576754093170166, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.875342607498169, "num_tokens": 717746785.0, "step": 18900 }, { "epoch": 2.4044014756392316, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.218834638595581, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8735222816467285, "num_tokens": 717785786.0, "step": 18901 }, { "epoch": 2.404528685917822, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9206478595733643, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8761827945709229, "num_tokens": 717821833.0, "step": 18902 }, { "epoch": 2.4046558961964126, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1990554332733154, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8876491785049438, "num_tokens": 717858107.0, "step": 18903 }, { "epoch": 2.404783106475003, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000324249267578125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.128319263458252, "learning_rate": 1e-05, "loss": 0.3956, "mean_token_accuracy": 0.8937098979949951, "num_tokens": 717893090.0, "step": 18904 }, { "epoch": 2.4049103167535937, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.310302495956421, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8895955681800842, "num_tokens": 717919703.0, "step": 18905 }, { "epoch": 2.405037527032184, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2768354415893555, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8909370303153992, "num_tokens": 717956877.0, "step": 18906 }, { "epoch": 2.4051647373107747, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.491442918777466, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8688790202140808, "num_tokens": 718002812.0, "step": 18907 }, { "epoch": 2.4052919475893653, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.28637170791626, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8811131119728088, "num_tokens": 718041170.0, "step": 18908 }, { "epoch": 2.405419157867956, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.680056571960449, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8759824633598328, "num_tokens": 718084168.0, "step": 18909 }, { "epoch": 2.4055463681465463, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.137198448181152, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8899157047271729, "num_tokens": 718118663.0, "step": 18910 }, { "epoch": 2.405673578425137, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.734285593032837, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8802422881126404, "num_tokens": 718155490.0, "step": 18911 }, { "epoch": 2.4058007887037274, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9452567100524902, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8765544295310974, "num_tokens": 718192301.0, "step": 18912 }, { "epoch": 2.405927998982318, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.677748203277588, "learning_rate": 1e-05, "loss": 0.407, "mean_token_accuracy": 0.8891290426254272, "num_tokens": 718231390.0, "step": 18913 }, { "epoch": 2.4060552092609084, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.7801876068115234, "learning_rate": 1e-05, "loss": 0.3987, "mean_token_accuracy": 0.8908427953720093, "num_tokens": 718269248.0, "step": 18914 }, { "epoch": 2.406182419539499, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.610173463821411, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8831689953804016, "num_tokens": 718305961.0, "step": 18915 }, { "epoch": 2.4063096298180895, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6818888187408447, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8745239973068237, "num_tokens": 718344659.0, "step": 18916 }, { "epoch": 2.4064368400966796, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5590121746063232, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8668476343154907, "num_tokens": 718386624.0, "step": 18917 }, { "epoch": 2.4065640503752705, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.608306884765625, "learning_rate": 1e-05, "loss": 0.3849, "mean_token_accuracy": 0.8925307393074036, "num_tokens": 718422154.0, "step": 18918 }, { "epoch": 2.4066912606538606, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.486229419708252, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8649477958679199, "num_tokens": 718460253.0, "step": 18919 }, { "epoch": 2.406818470932451, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5904154777526855, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8815697431564331, "num_tokens": 718495328.0, "step": 18920 }, { "epoch": 2.4069456812110417, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.431292772293091, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8827155828475952, "num_tokens": 718531244.0, "step": 18921 }, { "epoch": 2.407072891489632, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.598660945892334, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8821401596069336, "num_tokens": 718569766.0, "step": 18922 }, { "epoch": 2.4072001017682227, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.352398157119751, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8791168928146362, "num_tokens": 718612496.0, "step": 18923 }, { "epoch": 2.4073273120468133, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6615400314331055, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8800262808799744, "num_tokens": 718651251.0, "step": 18924 }, { "epoch": 2.407454522325404, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.3061704635620117, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8787038922309875, "num_tokens": 718683204.0, "step": 18925 }, { "epoch": 2.4075817326039943, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7617504596710205, "learning_rate": 1e-05, "loss": 0.3959, "mean_token_accuracy": 0.8885771036148071, "num_tokens": 718717114.0, "step": 18926 }, { "epoch": 2.407708942882585, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.237067937850952, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8715085983276367, "num_tokens": 718754756.0, "step": 18927 }, { "epoch": 2.4078361531611754, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9299421310424805, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8714011907577515, "num_tokens": 718792312.0, "step": 18928 }, { "epoch": 2.407963363439766, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.1961417198181152, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8761832118034363, "num_tokens": 718834771.0, "step": 18929 }, { "epoch": 2.4080905737183564, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.165024280548096, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8747755289077759, "num_tokens": 718871124.0, "step": 18930 }, { "epoch": 2.408217783996947, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.3077821731567383, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8765383362770081, "num_tokens": 718912037.0, "step": 18931 }, { "epoch": 2.4083449942755375, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.318528652191162, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.871987521648407, "num_tokens": 718953048.0, "step": 18932 }, { "epoch": 2.408472204554128, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.559967041015625, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8722620010375977, "num_tokens": 718991825.0, "step": 18933 }, { "epoch": 2.4085994148327186, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.271683216094971, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.876349925994873, "num_tokens": 719026847.0, "step": 18934 }, { "epoch": 2.408726625111309, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.739812135696411, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8834437131881714, "num_tokens": 719061415.0, "step": 18935 }, { "epoch": 2.4088538353898996, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.087611198425293, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8854790329933167, "num_tokens": 719097432.0, "step": 18936 }, { "epoch": 2.40898104566849, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7466673851013184, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8810005187988281, "num_tokens": 719129301.0, "step": 18937 }, { "epoch": 2.4091082559470807, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.876221179962158, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8775089383125305, "num_tokens": 719171772.0, "step": 18938 }, { "epoch": 2.409235466225671, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.682328224182129, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8735524415969849, "num_tokens": 719211751.0, "step": 18939 }, { "epoch": 2.4093626765042617, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7154502868652344, "learning_rate": 1e-05, "loss": 0.3838, "mean_token_accuracy": 0.896080493927002, "num_tokens": 719246812.0, "step": 18940 }, { "epoch": 2.4094898867828523, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.612807273864746, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8610517978668213, "num_tokens": 719290962.0, "step": 18941 }, { "epoch": 2.4096170970614423, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6328322887420654, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.878959059715271, "num_tokens": 719323506.0, "step": 18942 }, { "epoch": 2.4097443073400333, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5394246578216553, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8703804016113281, "num_tokens": 719361435.0, "step": 18943 }, { "epoch": 2.4098715176186234, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5691752433776855, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8897764682769775, "num_tokens": 719405617.0, "step": 18944 }, { "epoch": 2.409998727897214, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4650838375091553, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.87811279296875, "num_tokens": 719450588.0, "step": 18945 }, { "epoch": 2.4101259381758044, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5726311206817627, "learning_rate": 1e-05, "loss": 0.3937, "mean_token_accuracy": 0.8876179456710815, "num_tokens": 719489028.0, "step": 18946 }, { "epoch": 2.410253148454395, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.41282320022583, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8740968704223633, "num_tokens": 719527921.0, "step": 18947 }, { "epoch": 2.4103803587329855, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.588526964187622, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8813114166259766, "num_tokens": 719572461.0, "step": 18948 }, { "epoch": 2.410507569011576, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3444793224334717, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8762094974517822, "num_tokens": 719613478.0, "step": 18949 }, { "epoch": 2.4106347792901666, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6707558631896973, "learning_rate": 1e-05, "loss": 0.3916, "mean_token_accuracy": 0.8911827802658081, "num_tokens": 719643665.0, "step": 18950 }, { "epoch": 2.410761989568757, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2885987758636475, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8697585463523865, "num_tokens": 719684168.0, "step": 18951 }, { "epoch": 2.4108891998473476, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7817323207855225, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8785338997840881, "num_tokens": 719722484.0, "step": 18952 }, { "epoch": 2.411016410125938, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.216184616088867, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8810065984725952, "num_tokens": 719756659.0, "step": 18953 }, { "epoch": 2.4111436204045287, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9751315116882324, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8804997205734253, "num_tokens": 719787302.0, "step": 18954 }, { "epoch": 2.411270830683119, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.194974422454834, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8821475505828857, "num_tokens": 719823359.0, "step": 18955 }, { "epoch": 2.4113980409617097, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.213099479675293, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8807809352874756, "num_tokens": 719861263.0, "step": 18956 }, { "epoch": 2.4115252512403003, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3312087059020996, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8820155262947083, "num_tokens": 719897987.0, "step": 18957 }, { "epoch": 2.411652461518891, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.387185096740723, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8874462842941284, "num_tokens": 719931067.0, "step": 18958 }, { "epoch": 2.4117796717974813, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.618901252746582, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8890368938446045, "num_tokens": 719971174.0, "step": 18959 }, { "epoch": 2.411906882076072, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.304909706115723, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8861327171325684, "num_tokens": 720005781.0, "step": 18960 }, { "epoch": 2.4120340923546624, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.788726568222046, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8877053260803223, "num_tokens": 720040917.0, "step": 18961 }, { "epoch": 2.412161302633253, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.097137928009033, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8685067296028137, "num_tokens": 720080490.0, "step": 18962 }, { "epoch": 2.4122885129118434, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7754297256469727, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8803327083587646, "num_tokens": 720116461.0, "step": 18963 }, { "epoch": 2.412415723190434, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8925600051879883, "learning_rate": 1e-05, "loss": 0.3654, "mean_token_accuracy": 0.9004961252212524, "num_tokens": 720156756.0, "step": 18964 }, { "epoch": 2.4125429334690245, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6893198490142822, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8794220685958862, "num_tokens": 720195271.0, "step": 18965 }, { "epoch": 2.412670143747615, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.747070789337158, "learning_rate": 1e-05, "loss": 0.4023, "mean_token_accuracy": 0.8889304399490356, "num_tokens": 720232159.0, "step": 18966 }, { "epoch": 2.412797354026205, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6121935844421387, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.885440468788147, "num_tokens": 720264591.0, "step": 18967 }, { "epoch": 2.412924564304796, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6556341648101807, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8817278742790222, "num_tokens": 720302957.0, "step": 18968 }, { "epoch": 2.413051774583386, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5470833778381348, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8726254105567932, "num_tokens": 720335679.0, "step": 18969 }, { "epoch": 2.4131789848619767, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.593445062637329, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8769715428352356, "num_tokens": 720377261.0, "step": 18970 }, { "epoch": 2.413306195140567, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4685566425323486, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8799923062324524, "num_tokens": 720415344.0, "step": 18971 }, { "epoch": 2.4134334054191577, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5973117351531982, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8751424551010132, "num_tokens": 720449133.0, "step": 18972 }, { "epoch": 2.4135606156977483, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4123427867889404, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8707015514373779, "num_tokens": 720484250.0, "step": 18973 }, { "epoch": 2.413687825976339, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.629747152328491, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.874160647392273, "num_tokens": 720520466.0, "step": 18974 }, { "epoch": 2.4138150362549293, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3263967037200928, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8835620880126953, "num_tokens": 720557013.0, "step": 18975 }, { "epoch": 2.41394224653352, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.711210250854492, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.882729709148407, "num_tokens": 720594957.0, "step": 18976 }, { "epoch": 2.4140694568121104, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2607967853546143, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8767143487930298, "num_tokens": 720630531.0, "step": 18977 }, { "epoch": 2.414196667090701, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8462092876434326, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8748723864555359, "num_tokens": 720673502.0, "step": 18978 }, { "epoch": 2.4143238773692914, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1973893642425537, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8797934651374817, "num_tokens": 720716765.0, "step": 18979 }, { "epoch": 2.414451087647882, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.06395959854126, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.8913682699203491, "num_tokens": 720754711.0, "step": 18980 }, { "epoch": 2.4145782979264725, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2533934116363525, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8746803998947144, "num_tokens": 720789067.0, "step": 18981 }, { "epoch": 2.414705508205063, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.278843402862549, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8717097640037537, "num_tokens": 720821725.0, "step": 18982 }, { "epoch": 2.4148327184836536, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.448415517807007, "learning_rate": 1e-05, "loss": 0.4219, "mean_token_accuracy": 0.8837986588478088, "num_tokens": 720857587.0, "step": 18983 }, { "epoch": 2.414959928762244, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.321002960205078, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8828825950622559, "num_tokens": 720900481.0, "step": 18984 }, { "epoch": 2.4150871390408346, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.699845790863037, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8794045448303223, "num_tokens": 720933131.0, "step": 18985 }, { "epoch": 2.415214349319425, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.152284145355225, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8779468536376953, "num_tokens": 720972142.0, "step": 18986 }, { "epoch": 2.4153415595980157, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.739132881164551, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8890652656555176, "num_tokens": 721014251.0, "step": 18987 }, { "epoch": 2.415468769876606, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.951282501220703, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8698718547821045, "num_tokens": 721053327.0, "step": 18988 }, { "epoch": 2.4155959801551967, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6989569664001465, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8834443688392639, "num_tokens": 721092047.0, "step": 18989 }, { "epoch": 2.415723190433787, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.7819125652313232, "learning_rate": 1e-05, "loss": 0.3739, "mean_token_accuracy": 0.8982367515563965, "num_tokens": 721125318.0, "step": 18990 }, { "epoch": 2.4158504007123778, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.63142728805542, "learning_rate": 1e-05, "loss": 0.3965, "mean_token_accuracy": 0.8933068513870239, "num_tokens": 721162058.0, "step": 18991 }, { "epoch": 2.415977610990968, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6702957153320312, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8822357654571533, "num_tokens": 721199219.0, "step": 18992 }, { "epoch": 2.4161048212695584, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5538291931152344, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8758689761161804, "num_tokens": 721239696.0, "step": 18993 }, { "epoch": 2.416232031548149, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6133158206939697, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8718454837799072, "num_tokens": 721277660.0, "step": 18994 }, { "epoch": 2.4163592418267394, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.495692729949951, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8792112469673157, "num_tokens": 721311805.0, "step": 18995 }, { "epoch": 2.41648645210533, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5769097805023193, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8598809242248535, "num_tokens": 721355968.0, "step": 18996 }, { "epoch": 2.4166136623839205, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4307239055633545, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8733018636703491, "num_tokens": 721397930.0, "step": 18997 }, { "epoch": 2.416740872662511, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5882201194763184, "learning_rate": 1e-05, "loss": 0.3928, "mean_token_accuracy": 0.8900173902511597, "num_tokens": 721432408.0, "step": 18998 }, { "epoch": 2.4168680829411016, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.362074136734009, "learning_rate": 1e-05, "loss": 0.3672, "mean_token_accuracy": 0.8985925316810608, "num_tokens": 721468903.0, "step": 18999 }, { "epoch": 2.416995293219692, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.640514850616455, "learning_rate": 1e-05, "loss": 0.383, "mean_token_accuracy": 0.896183967590332, "num_tokens": 721509687.0, "step": 19000 }, { "epoch": 2.4171225034982826, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.307863473892212, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8722545504570007, "num_tokens": 721548130.0, "step": 19001 }, { "epoch": 2.417249713776873, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.73702335357666, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8850895166397095, "num_tokens": 721588504.0, "step": 19002 }, { "epoch": 2.4173769240554637, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.229142189025879, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8766920566558838, "num_tokens": 721632621.0, "step": 19003 }, { "epoch": 2.417504134334054, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.896913528442383, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8782973289489746, "num_tokens": 721670898.0, "step": 19004 }, { "epoch": 2.4176313446126447, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.191305160522461, "learning_rate": 1e-05, "loss": 0.3967, "mean_token_accuracy": 0.8882346153259277, "num_tokens": 721709012.0, "step": 19005 }, { "epoch": 2.4177585548912353, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.117392539978027, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8799867033958435, "num_tokens": 721747135.0, "step": 19006 }, { "epoch": 2.417885765169826, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.277170181274414, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8796289563179016, "num_tokens": 721784447.0, "step": 19007 }, { "epoch": 2.4180129754484163, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.324129104614258, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.869307816028595, "num_tokens": 721817853.0, "step": 19008 }, { "epoch": 2.418140185727007, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5220041275024414, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8760623931884766, "num_tokens": 721852537.0, "step": 19009 }, { "epoch": 2.4182673960055974, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.298777103424072, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8821743726730347, "num_tokens": 721892156.0, "step": 19010 }, { "epoch": 2.418394606284188, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.710897445678711, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8813272714614868, "num_tokens": 721929425.0, "step": 19011 }, { "epoch": 2.4185218165627784, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.136035919189453, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8710367679595947, "num_tokens": 721967492.0, "step": 19012 }, { "epoch": 2.418649026841369, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7805416584014893, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8754668235778809, "num_tokens": 722007754.0, "step": 19013 }, { "epoch": 2.4187762371199595, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9223108291625977, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8802402019500732, "num_tokens": 722043380.0, "step": 19014 }, { "epoch": 2.4189034473985496, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.689786195755005, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8755433559417725, "num_tokens": 722086200.0, "step": 19015 }, { "epoch": 2.4190306576771405, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.752817392349243, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8834058046340942, "num_tokens": 722122907.0, "step": 19016 }, { "epoch": 2.4191578679557306, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.620300769805908, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8823654651641846, "num_tokens": 722155951.0, "step": 19017 }, { "epoch": 2.419285078234321, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.643625497817993, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8753643035888672, "num_tokens": 722194479.0, "step": 19018 }, { "epoch": 2.4194122885129117, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.549044370651245, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8831652402877808, "num_tokens": 722228470.0, "step": 19019 }, { "epoch": 2.419539498791502, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.597705841064453, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8708029985427856, "num_tokens": 722265960.0, "step": 19020 }, { "epoch": 2.4196667090700927, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.473738670349121, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8750870227813721, "num_tokens": 722301377.0, "step": 19021 }, { "epoch": 2.4197939193486833, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.580336570739746, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8826730847358704, "num_tokens": 722338534.0, "step": 19022 }, { "epoch": 2.419921129627274, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.408770799636841, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8769394159317017, "num_tokens": 722377460.0, "step": 19023 }, { "epoch": 2.4200483399058643, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.613276243209839, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.869229257106781, "num_tokens": 722415167.0, "step": 19024 }, { "epoch": 2.420175550184455, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.345392942428589, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.867347240447998, "num_tokens": 722456114.0, "step": 19025 }, { "epoch": 2.4203027604630454, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.679760217666626, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8847121596336365, "num_tokens": 722496378.0, "step": 19026 }, { "epoch": 2.420429970741636, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2791051864624023, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.871173083782196, "num_tokens": 722534903.0, "step": 19027 }, { "epoch": 2.4205571810202264, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8607170581817627, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8743689060211182, "num_tokens": 722576649.0, "step": 19028 }, { "epoch": 2.420684391298817, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2137835025787354, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8710938692092896, "num_tokens": 722618555.0, "step": 19029 }, { "epoch": 2.4208116015774075, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.034794330596924, "learning_rate": 1e-05, "loss": 0.3819, "mean_token_accuracy": 0.8976295590400696, "num_tokens": 722652045.0, "step": 19030 }, { "epoch": 2.420938811855998, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2358009815216064, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8766272068023682, "num_tokens": 722686987.0, "step": 19031 }, { "epoch": 2.4210660221345885, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.254090309143066, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8736525774002075, "num_tokens": 722724161.0, "step": 19032 }, { "epoch": 2.421193232413179, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.4103221893310547, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8863760232925415, "num_tokens": 722766473.0, "step": 19033 }, { "epoch": 2.4213204426917696, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.335250377655029, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.879987359046936, "num_tokens": 722809499.0, "step": 19034 }, { "epoch": 2.42144765297036, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6688649654388428, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8838247060775757, "num_tokens": 722847295.0, "step": 19035 }, { "epoch": 2.4215748632489507, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.192945957183838, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8801243305206299, "num_tokens": 722884313.0, "step": 19036 }, { "epoch": 2.421702073527541, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.760566234588623, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8823933601379395, "num_tokens": 722924530.0, "step": 19037 }, { "epoch": 2.4218292838061317, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.9884698390960693, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8682846426963806, "num_tokens": 722962968.0, "step": 19038 }, { "epoch": 2.4219564940847222, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7254743576049805, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8880756497383118, "num_tokens": 722997801.0, "step": 19039 }, { "epoch": 2.4220837043633123, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8018031120300293, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8678563833236694, "num_tokens": 723037724.0, "step": 19040 }, { "epoch": 2.4222109146419033, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.654867649078369, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.870051383972168, "num_tokens": 723078199.0, "step": 19041 }, { "epoch": 2.4223381249204934, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6792516708374023, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8729868531227112, "num_tokens": 723110894.0, "step": 19042 }, { "epoch": 2.422465335199084, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5819191932678223, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8683452010154724, "num_tokens": 723150688.0, "step": 19043 }, { "epoch": 2.4225925454776744, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5908689498901367, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8778183460235596, "num_tokens": 723189863.0, "step": 19044 }, { "epoch": 2.422719755756265, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5092506408691406, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8687961101531982, "num_tokens": 723232333.0, "step": 19045 }, { "epoch": 2.4228469660348555, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5918428897857666, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8653856515884399, "num_tokens": 723267469.0, "step": 19046 }, { "epoch": 2.422974176313446, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4292283058166504, "learning_rate": 1e-05, "loss": 0.3855, "mean_token_accuracy": 0.8914923667907715, "num_tokens": 723304367.0, "step": 19047 }, { "epoch": 2.4231013865920366, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.590416431427002, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8769542574882507, "num_tokens": 723343247.0, "step": 19048 }, { "epoch": 2.423228596870627, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3663766384124756, "learning_rate": 1e-05, "loss": 0.3946, "mean_token_accuracy": 0.8898732662200928, "num_tokens": 723378022.0, "step": 19049 }, { "epoch": 2.4233558071492176, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6601767539978027, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.8809774518013, "num_tokens": 723409827.0, "step": 19050 }, { "epoch": 2.423483017427808, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2951765060424805, "learning_rate": 1e-05, "loss": 0.3982, "mean_token_accuracy": 0.886601984500885, "num_tokens": 723447192.0, "step": 19051 }, { "epoch": 2.4236102277063987, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7591679096221924, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8839346170425415, "num_tokens": 723492123.0, "step": 19052 }, { "epoch": 2.423737437984989, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2215991020202637, "learning_rate": 1e-05, "loss": 0.3748, "mean_token_accuracy": 0.8951293230056763, "num_tokens": 723525680.0, "step": 19053 }, { "epoch": 2.4238646482635797, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.949000597000122, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8773781061172485, "num_tokens": 723565667.0, "step": 19054 }, { "epoch": 2.4239918585421703, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.207366704940796, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8742771744728088, "num_tokens": 723599486.0, "step": 19055 }, { "epoch": 2.424119068820761, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.167287826538086, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.884442925453186, "num_tokens": 723631992.0, "step": 19056 }, { "epoch": 2.4242462790993513, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3122189044952393, "learning_rate": 1e-05, "loss": 0.394, "mean_token_accuracy": 0.8877179622650146, "num_tokens": 723667654.0, "step": 19057 }, { "epoch": 2.424373489377942, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003261566162109375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.320970058441162, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8845454454421997, "num_tokens": 723705368.0, "step": 19058 }, { "epoch": 2.4245006996565324, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5628600120544434, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8817277550697327, "num_tokens": 723744973.0, "step": 19059 }, { "epoch": 2.424627909935123, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.291449069976807, "learning_rate": 1e-05, "loss": 0.4757, "mean_token_accuracy": 0.8678823709487915, "num_tokens": 723778868.0, "step": 19060 }, { "epoch": 2.4247551202137134, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7560815811157227, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8827803134918213, "num_tokens": 723822413.0, "step": 19061 }, { "epoch": 2.424882330492304, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.076171875, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8712867498397827, "num_tokens": 723865826.0, "step": 19062 }, { "epoch": 2.4250095407708945, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7598373889923096, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8855248093605042, "num_tokens": 723901839.0, "step": 19063 }, { "epoch": 2.425136751049485, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8566887378692627, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8861269950866699, "num_tokens": 723944883.0, "step": 19064 }, { "epoch": 2.425263961328075, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6907083988189697, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8851895928382874, "num_tokens": 723985084.0, "step": 19065 }, { "epoch": 2.425391171606666, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6965348720550537, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8771950602531433, "num_tokens": 724027919.0, "step": 19066 }, { "epoch": 2.425518381885256, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.61924147605896, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8849829435348511, "num_tokens": 724064593.0, "step": 19067 }, { "epoch": 2.4256455921638467, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.611140251159668, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.8900255560874939, "num_tokens": 724095967.0, "step": 19068 }, { "epoch": 2.425772802442437, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.549564838409424, "learning_rate": 1e-05, "loss": 0.3851, "mean_token_accuracy": 0.8946139812469482, "num_tokens": 724128420.0, "step": 19069 }, { "epoch": 2.4259000127210277, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.564981460571289, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8703463077545166, "num_tokens": 724165073.0, "step": 19070 }, { "epoch": 2.4260272229996183, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.483837604522705, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8780845999717712, "num_tokens": 724201377.0, "step": 19071 }, { "epoch": 2.426154433278209, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.547626256942749, "learning_rate": 1e-05, "loss": 0.385, "mean_token_accuracy": 0.8947796821594238, "num_tokens": 724241783.0, "step": 19072 }, { "epoch": 2.4262816435567993, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4296700954437256, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8792970180511475, "num_tokens": 724277903.0, "step": 19073 }, { "epoch": 2.42640885383539, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.552136182785034, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8876473903656006, "num_tokens": 724318834.0, "step": 19074 }, { "epoch": 2.4265360641139804, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3832826614379883, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8843239545822144, "num_tokens": 724356406.0, "step": 19075 }, { "epoch": 2.426663274392571, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5875115394592285, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8893504738807678, "num_tokens": 724394774.0, "step": 19076 }, { "epoch": 2.4267904846711614, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.33100962638855, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8721292614936829, "num_tokens": 724431825.0, "step": 19077 }, { "epoch": 2.426917694949752, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6894948482513428, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8791518211364746, "num_tokens": 724466402.0, "step": 19078 }, { "epoch": 2.4270449052283425, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.2570297718048096, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8702956438064575, "num_tokens": 724503353.0, "step": 19079 }, { "epoch": 2.427172115506933, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8069779872894287, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8854339718818665, "num_tokens": 724540348.0, "step": 19080 }, { "epoch": 2.4272993257855235, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.186389923095703, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8766626119613647, "num_tokens": 724577999.0, "step": 19081 }, { "epoch": 2.427426536064114, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.009300708770752, "learning_rate": 1e-05, "loss": 0.3821, "mean_token_accuracy": 0.8938761353492737, "num_tokens": 724614754.0, "step": 19082 }, { "epoch": 2.4275537463427046, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.1882834434509277, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.869560718536377, "num_tokens": 724653232.0, "step": 19083 }, { "epoch": 2.427680956621295, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.270847797393799, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8711336851119995, "num_tokens": 724686461.0, "step": 19084 }, { "epoch": 2.4278081668998857, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.3593997955322266, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.88982093334198, "num_tokens": 724727330.0, "step": 19085 }, { "epoch": 2.427935377178476, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.419840335845947, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8707090616226196, "num_tokens": 724766509.0, "step": 19086 }, { "epoch": 2.4280625874570667, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.686641216278076, "learning_rate": 1e-05, "loss": 0.3938, "mean_token_accuracy": 0.8964310884475708, "num_tokens": 724795128.0, "step": 19087 }, { "epoch": 2.428189797735657, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.287237644195557, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8744215965270996, "num_tokens": 724833941.0, "step": 19088 }, { "epoch": 2.4283170080142478, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.8094258308410645, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8851913213729858, "num_tokens": 724870429.0, "step": 19089 }, { "epoch": 2.428444218292838, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.094407081604004, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8827353715896606, "num_tokens": 724905574.0, "step": 19090 }, { "epoch": 2.4285714285714284, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7671236991882324, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.885668933391571, "num_tokens": 724949845.0, "step": 19091 }, { "epoch": 2.428698638850019, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.8990602493286133, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8720499277114868, "num_tokens": 724987586.0, "step": 19092 }, { "epoch": 2.4288258491286094, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6714582443237305, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8864938020706177, "num_tokens": 725026432.0, "step": 19093 }, { "epoch": 2.4289530594072, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7771339416503906, "learning_rate": 1e-05, "loss": 0.4972, "mean_token_accuracy": 0.8606375455856323, "num_tokens": 725070627.0, "step": 19094 }, { "epoch": 2.4290802696857905, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5831499099731445, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8798542618751526, "num_tokens": 725106005.0, "step": 19095 }, { "epoch": 2.429207479964381, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.713221311569214, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8849353790283203, "num_tokens": 725149667.0, "step": 19096 }, { "epoch": 2.4293346902429716, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4821574687957764, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8833906054496765, "num_tokens": 725185526.0, "step": 19097 }, { "epoch": 2.429461900521562, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.710287570953369, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8839306831359863, "num_tokens": 725221789.0, "step": 19098 }, { "epoch": 2.4295891108001526, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4024956226348877, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.87235426902771, "num_tokens": 725258275.0, "step": 19099 }, { "epoch": 2.429716321078743, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7556569576263428, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8658692836761475, "num_tokens": 725297753.0, "step": 19100 }, { "epoch": 2.4298435313573337, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.327728748321533, "learning_rate": 1e-05, "loss": 0.3972, "mean_token_accuracy": 0.8860973119735718, "num_tokens": 725331978.0, "step": 19101 }, { "epoch": 2.429970741635924, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8537285327911377, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8832148313522339, "num_tokens": 725373876.0, "step": 19102 }, { "epoch": 2.4300979519145147, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2802796363830566, "learning_rate": 1e-05, "loss": 0.3917, "mean_token_accuracy": 0.8870707154273987, "num_tokens": 725411280.0, "step": 19103 }, { "epoch": 2.4302251621931052, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.006048202514648, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8662946224212646, "num_tokens": 725452050.0, "step": 19104 }, { "epoch": 2.4303523724716958, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.302215814590454, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.886805534362793, "num_tokens": 725489922.0, "step": 19105 }, { "epoch": 2.4304795827502863, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.152813911437988, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8753635883331299, "num_tokens": 725527871.0, "step": 19106 }, { "epoch": 2.430606793028877, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.477801561355591, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8907899260520935, "num_tokens": 725565045.0, "step": 19107 }, { "epoch": 2.4307340033074674, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.130019187927246, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8907801508903503, "num_tokens": 725596604.0, "step": 19108 }, { "epoch": 2.430861213586058, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5880799293518066, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8801926970481873, "num_tokens": 725632299.0, "step": 19109 }, { "epoch": 2.4309884238646484, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9881956577301025, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8783252239227295, "num_tokens": 725673778.0, "step": 19110 }, { "epoch": 2.431115634143239, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.644723653793335, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8679170608520508, "num_tokens": 725715087.0, "step": 19111 }, { "epoch": 2.4312428444218295, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8026928901672363, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8825231790542603, "num_tokens": 725754047.0, "step": 19112 }, { "epoch": 2.4313700547004196, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.640104293823242, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8684003353118896, "num_tokens": 725793168.0, "step": 19113 }, { "epoch": 2.4314972649790105, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6195127964019775, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8835617303848267, "num_tokens": 725834713.0, "step": 19114 }, { "epoch": 2.4316244752576006, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.63149094581604, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8807744979858398, "num_tokens": 725872759.0, "step": 19115 }, { "epoch": 2.431751685536191, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4759013652801514, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.88399738073349, "num_tokens": 725916192.0, "step": 19116 }, { "epoch": 2.4318788958147817, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.634600877761841, "learning_rate": 1e-05, "loss": 0.3739, "mean_token_accuracy": 0.8956094980239868, "num_tokens": 725949838.0, "step": 19117 }, { "epoch": 2.432006106093372, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.400599956512451, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8741720914840698, "num_tokens": 725986538.0, "step": 19118 }, { "epoch": 2.4321333163719627, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6488335132598877, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8840976357460022, "num_tokens": 726029007.0, "step": 19119 }, { "epoch": 2.4322605266505533, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.32661509513855, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8815414905548096, "num_tokens": 726066745.0, "step": 19120 }, { "epoch": 2.432387736929144, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7074344158172607, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8737679123878479, "num_tokens": 726106768.0, "step": 19121 }, { "epoch": 2.4325149472077343, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.272394895553589, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8810244798660278, "num_tokens": 726144879.0, "step": 19122 }, { "epoch": 2.432642157486325, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8201563358306885, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.880955696105957, "num_tokens": 726181278.0, "step": 19123 }, { "epoch": 2.4327693677649154, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2080886363983154, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8815908432006836, "num_tokens": 726226090.0, "step": 19124 }, { "epoch": 2.432896578043506, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.001902103424072, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8781005144119263, "num_tokens": 726261567.0, "step": 19125 }, { "epoch": 2.4330237883220964, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2222163677215576, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8800620436668396, "num_tokens": 726300527.0, "step": 19126 }, { "epoch": 2.433150998600687, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.194005012512207, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8858091831207275, "num_tokens": 726341240.0, "step": 19127 }, { "epoch": 2.4332782088792775, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.36961030960083, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8720259070396423, "num_tokens": 726375286.0, "step": 19128 }, { "epoch": 2.433405419157868, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.302908897399902, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.89417564868927, "num_tokens": 726414810.0, "step": 19129 }, { "epoch": 2.4335326294364585, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5778276920318604, "learning_rate": 1e-05, "loss": 0.3897, "mean_token_accuracy": 0.8962689638137817, "num_tokens": 726451204.0, "step": 19130 }, { "epoch": 2.433659839715049, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.230783462524414, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8795778751373291, "num_tokens": 726491107.0, "step": 19131 }, { "epoch": 2.4337870499936396, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7158255577087402, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8776094913482666, "num_tokens": 726535297.0, "step": 19132 }, { "epoch": 2.43391426027223, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.045344829559326, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8711364269256592, "num_tokens": 726574841.0, "step": 19133 }, { "epoch": 2.4340414705508207, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7235071659088135, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8746622800827026, "num_tokens": 726611101.0, "step": 19134 }, { "epoch": 2.434168680829411, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8516650199890137, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8750964403152466, "num_tokens": 726654323.0, "step": 19135 }, { "epoch": 2.4342958911080017, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6653459072113037, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8827115893363953, "num_tokens": 726691504.0, "step": 19136 }, { "epoch": 2.4344231013865922, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6987709999084473, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8747701048851013, "num_tokens": 726737257.0, "step": 19137 }, { "epoch": 2.4345503116651823, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.589209794998169, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.889336884021759, "num_tokens": 726779917.0, "step": 19138 }, { "epoch": 2.4346775219437733, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6051723957061768, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8832709193229675, "num_tokens": 726823130.0, "step": 19139 }, { "epoch": 2.4348047322223634, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5263636112213135, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8827661275863647, "num_tokens": 726861787.0, "step": 19140 }, { "epoch": 2.434931942500954, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5723140239715576, "learning_rate": 1e-05, "loss": 0.471, "mean_token_accuracy": 0.8654317855834961, "num_tokens": 726904413.0, "step": 19141 }, { "epoch": 2.4350591527795444, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4620094299316406, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8911104202270508, "num_tokens": 726946984.0, "step": 19142 }, { "epoch": 2.435186363058135, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5773606300354004, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8799002766609192, "num_tokens": 726982587.0, "step": 19143 }, { "epoch": 2.4353135733367255, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.402301549911499, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8835026025772095, "num_tokens": 727017255.0, "step": 19144 }, { "epoch": 2.435440783615316, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.622406482696533, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8868828415870667, "num_tokens": 727046147.0, "step": 19145 }, { "epoch": 2.4355679938939065, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3327887058258057, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.878230631351471, "num_tokens": 727085982.0, "step": 19146 }, { "epoch": 2.435695204172497, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.681919574737549, "learning_rate": 1e-05, "loss": 0.393, "mean_token_accuracy": 0.8922491073608398, "num_tokens": 727123991.0, "step": 19147 }, { "epoch": 2.4358224144510876, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.276332139968872, "learning_rate": 1e-05, "loss": 0.469, "mean_token_accuracy": 0.8646088242530823, "num_tokens": 727162155.0, "step": 19148 }, { "epoch": 2.435949624729678, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8091914653778076, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8704794645309448, "num_tokens": 727201234.0, "step": 19149 }, { "epoch": 2.4360768350082687, "ewc_loss": 0.0791015625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.210665702819824, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8784589767456055, "num_tokens": 727243429.0, "step": 19150 }, { "epoch": 2.436204045286859, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.008186340332031, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8806084394454956, "num_tokens": 727279705.0, "step": 19151 }, { "epoch": 2.4363312555654497, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.216961622238159, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8853947520256042, "num_tokens": 727315503.0, "step": 19152 }, { "epoch": 2.4364584658440402, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.220003128051758, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8764740228652954, "num_tokens": 727355272.0, "step": 19153 }, { "epoch": 2.4365856761226308, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3908638954162598, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8662657737731934, "num_tokens": 727391058.0, "step": 19154 }, { "epoch": 2.4367128864012213, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.315145015716553, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8740348815917969, "num_tokens": 727433011.0, "step": 19155 }, { "epoch": 2.436840096679812, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.637223243713379, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8859447836875916, "num_tokens": 727462282.0, "step": 19156 }, { "epoch": 2.4369673069584024, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1956257820129395, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8773941397666931, "num_tokens": 727501636.0, "step": 19157 }, { "epoch": 2.437094517236993, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7276315689086914, "learning_rate": 1e-05, "loss": 0.3753, "mean_token_accuracy": 0.8971768617630005, "num_tokens": 727536828.0, "step": 19158 }, { "epoch": 2.4372217275155834, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.998715877532959, "learning_rate": 1e-05, "loss": 0.5108, "mean_token_accuracy": 0.852845311164856, "num_tokens": 727574858.0, "step": 19159 }, { "epoch": 2.437348937794174, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7099528312683105, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8945496082305908, "num_tokens": 727618168.0, "step": 19160 }, { "epoch": 2.4374761480727645, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8129987716674805, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8685007095336914, "num_tokens": 727660205.0, "step": 19161 }, { "epoch": 2.437603358351355, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.644166946411133, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8866612911224365, "num_tokens": 727701980.0, "step": 19162 }, { "epoch": 2.437730568629945, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6934688091278076, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8678563833236694, "num_tokens": 727743766.0, "step": 19163 }, { "epoch": 2.437857778908536, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5721356868743896, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8815574645996094, "num_tokens": 727785504.0, "step": 19164 }, { "epoch": 2.437984989187126, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6000843048095703, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8865891098976135, "num_tokens": 727831296.0, "step": 19165 }, { "epoch": 2.4381121994657167, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5190305709838867, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8868405818939209, "num_tokens": 727867343.0, "step": 19166 }, { "epoch": 2.438239409744307, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.572695255279541, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8803573846817017, "num_tokens": 727903856.0, "step": 19167 }, { "epoch": 2.4383666200228977, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4702248573303223, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8785312175750732, "num_tokens": 727938385.0, "step": 19168 }, { "epoch": 2.4384938303014883, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5758135318756104, "learning_rate": 1e-05, "loss": 0.3939, "mean_token_accuracy": 0.8874319791793823, "num_tokens": 727965566.0, "step": 19169 }, { "epoch": 2.438621040580079, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.398577928543091, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8930459022521973, "num_tokens": 728002203.0, "step": 19170 }, { "epoch": 2.4387482508586693, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.603703260421753, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8868800401687622, "num_tokens": 728037055.0, "step": 19171 }, { "epoch": 2.43887546113726, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3381903171539307, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8726950287818909, "num_tokens": 728077226.0, "step": 19172 }, { "epoch": 2.4390026714158504, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6854848861694336, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8849767446517944, "num_tokens": 728111396.0, "step": 19173 }, { "epoch": 2.439129881694441, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2756869792938232, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8758266568183899, "num_tokens": 728149609.0, "step": 19174 }, { "epoch": 2.4392570919730314, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8049516677856445, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8796789050102234, "num_tokens": 728186668.0, "step": 19175 }, { "epoch": 2.439384302251622, "ewc_loss": 0.07958984375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.215782880783081, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8643234968185425, "num_tokens": 728226504.0, "step": 19176 }, { "epoch": 2.4395115125302125, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.983734369277954, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8778421878814697, "num_tokens": 728264648.0, "step": 19177 }, { "epoch": 2.439638722808803, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2158918380737305, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8817815780639648, "num_tokens": 728300292.0, "step": 19178 }, { "epoch": 2.4397659330873935, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.197122573852539, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8782345056533813, "num_tokens": 728336169.0, "step": 19179 }, { "epoch": 2.439893143365984, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.335174560546875, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8848304152488708, "num_tokens": 728380160.0, "step": 19180 }, { "epoch": 2.4400203536445746, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.330208778381348, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8819724321365356, "num_tokens": 728417978.0, "step": 19181 }, { "epoch": 2.440147563923165, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.593895196914673, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8835688829421997, "num_tokens": 728455264.0, "step": 19182 }, { "epoch": 2.4402747742017556, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.250142574310303, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8681041598320007, "num_tokens": 728498880.0, "step": 19183 }, { "epoch": 2.440401984480346, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.758960723876953, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8758516311645508, "num_tokens": 728531925.0, "step": 19184 }, { "epoch": 2.4405291947589367, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.039703369140625, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.876250684261322, "num_tokens": 728570573.0, "step": 19185 }, { "epoch": 2.440656405037527, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.73268723487854, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8816182017326355, "num_tokens": 728610877.0, "step": 19186 }, { "epoch": 2.4407836153161178, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.843655586242676, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8784977197647095, "num_tokens": 728650963.0, "step": 19187 }, { "epoch": 2.440910825594708, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.658590316772461, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8894591927528381, "num_tokens": 728688699.0, "step": 19188 }, { "epoch": 2.4410380358732984, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7225232124328613, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8687528967857361, "num_tokens": 728730142.0, "step": 19189 }, { "epoch": 2.441165246151889, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.586601495742798, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8713686466217041, "num_tokens": 728767808.0, "step": 19190 }, { "epoch": 2.4412924564304794, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.655322790145874, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8667101263999939, "num_tokens": 728816171.0, "step": 19191 }, { "epoch": 2.44141966670907, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5184173583984375, "learning_rate": 1e-05, "loss": 0.4837, "mean_token_accuracy": 0.8645815253257751, "num_tokens": 728858350.0, "step": 19192 }, { "epoch": 2.4415468769876605, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.605199098587036, "learning_rate": 1e-05, "loss": 0.3882, "mean_token_accuracy": 0.8906877040863037, "num_tokens": 728894408.0, "step": 19193 }, { "epoch": 2.441674087266251, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4380099773406982, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.8595688343048096, "num_tokens": 728936563.0, "step": 19194 }, { "epoch": 2.4418012975448415, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.617703676223755, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8717071413993835, "num_tokens": 728978796.0, "step": 19195 }, { "epoch": 2.441928507823432, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3558027744293213, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8867089152336121, "num_tokens": 729018524.0, "step": 19196 }, { "epoch": 2.4420557181020226, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6983237266540527, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8778951168060303, "num_tokens": 729055336.0, "step": 19197 }, { "epoch": 2.442182928380613, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.00032806396484375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2894015312194824, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8707172870635986, "num_tokens": 729095505.0, "step": 19198 }, { "epoch": 2.4423101386592037, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8044779300689697, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8776520490646362, "num_tokens": 729138384.0, "step": 19199 }, { "epoch": 2.442437348937794, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2290613651275635, "learning_rate": 1e-05, "loss": 0.4825, "mean_token_accuracy": 0.8586210012435913, "num_tokens": 729179942.0, "step": 19200 }, { "epoch": 2.4425645592163847, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9841041564941406, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.8615221977233887, "num_tokens": 729223051.0, "step": 19201 }, { "epoch": 2.4426917694949752, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.23163104057312, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.8824412226676941, "num_tokens": 729259368.0, "step": 19202 }, { "epoch": 2.4428189797735658, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.194967746734619, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8742376565933228, "num_tokens": 729295148.0, "step": 19203 }, { "epoch": 2.4429461900521563, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3847389221191406, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8717029094696045, "num_tokens": 729330649.0, "step": 19204 }, { "epoch": 2.443073400330747, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.278520107269287, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8854974508285522, "num_tokens": 729369260.0, "step": 19205 }, { "epoch": 2.4432006106093374, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.601559638977051, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8803132772445679, "num_tokens": 729406029.0, "step": 19206 }, { "epoch": 2.443327820887928, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.175998210906982, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8749188184738159, "num_tokens": 729444102.0, "step": 19207 }, { "epoch": 2.4434550311665184, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7046515941619873, "learning_rate": 1e-05, "loss": 0.3671, "mean_token_accuracy": 0.9008744359016418, "num_tokens": 729477121.0, "step": 19208 }, { "epoch": 2.443582241445109, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.006131649017334, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8690651655197144, "num_tokens": 729511682.0, "step": 19209 }, { "epoch": 2.4437094517236995, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003299713134765625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.706786870956421, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8854540586471558, "num_tokens": 729545675.0, "step": 19210 }, { "epoch": 2.4438366620022896, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.807997226715088, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8785732984542847, "num_tokens": 729585323.0, "step": 19211 }, { "epoch": 2.4439638722808805, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6441385746002197, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8727270364761353, "num_tokens": 729623881.0, "step": 19212 }, { "epoch": 2.4440910825594706, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6722538471221924, "learning_rate": 1e-05, "loss": 0.3923, "mean_token_accuracy": 0.893883466720581, "num_tokens": 729655486.0, "step": 19213 }, { "epoch": 2.444218292838061, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5818562507629395, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8796948194503784, "num_tokens": 729693606.0, "step": 19214 }, { "epoch": 2.4443455031166517, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.586132287979126, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.870477557182312, "num_tokens": 729729467.0, "step": 19215 }, { "epoch": 2.444472713395242, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.521456003189087, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8885051012039185, "num_tokens": 729767264.0, "step": 19216 }, { "epoch": 2.4445999236738327, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.540107488632202, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8826398849487305, "num_tokens": 729807352.0, "step": 19217 }, { "epoch": 2.4447271339524232, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4812960624694824, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8605796694755554, "num_tokens": 729847489.0, "step": 19218 }, { "epoch": 2.4448543442310138, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5122854709625244, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8854523301124573, "num_tokens": 729889675.0, "step": 19219 }, { "epoch": 2.4449815545096043, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4501380920410156, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8765418529510498, "num_tokens": 729926562.0, "step": 19220 }, { "epoch": 2.445108764788195, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5226523876190186, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8642293810844421, "num_tokens": 729964533.0, "step": 19221 }, { "epoch": 2.4452359750667854, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3972182273864746, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8898025751113892, "num_tokens": 730004206.0, "step": 19222 }, { "epoch": 2.445363185345376, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5436975955963135, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8730354905128479, "num_tokens": 730046779.0, "step": 19223 }, { "epoch": 2.4454903956239664, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3525822162628174, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8750690221786499, "num_tokens": 730085882.0, "step": 19224 }, { "epoch": 2.445617605902557, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5991156101226807, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8765566349029541, "num_tokens": 730126072.0, "step": 19225 }, { "epoch": 2.4457448161811475, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2853164672851562, "learning_rate": 1e-05, "loss": 0.3772, "mean_token_accuracy": 0.8959270715713501, "num_tokens": 730160038.0, "step": 19226 }, { "epoch": 2.445872026459738, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 4.022450923919678, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8808988928794861, "num_tokens": 730195701.0, "step": 19227 }, { "epoch": 2.4459992367383285, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.158435821533203, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.8848680257797241, "num_tokens": 730233938.0, "step": 19228 }, { "epoch": 2.446126447016919, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.159160614013672, "learning_rate": 1e-05, "loss": 0.3931, "mean_token_accuracy": 0.8940992951393127, "num_tokens": 730272333.0, "step": 19229 }, { "epoch": 2.4462536572955096, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.1923460960388184, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8689148426055908, "num_tokens": 730312967.0, "step": 19230 }, { "epoch": 2.4463808675741, "ewc_loss": 0.10595703125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.504088401794434, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8812375664710999, "num_tokens": 730345525.0, "step": 19231 }, { "epoch": 2.4465080778526906, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.5709264278411865, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.882867693901062, "num_tokens": 730384329.0, "step": 19232 }, { "epoch": 2.446635288131281, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.5197858810424805, "learning_rate": 1e-05, "loss": 0.4008, "mean_token_accuracy": 0.8948578834533691, "num_tokens": 730423139.0, "step": 19233 }, { "epoch": 2.4467624984098717, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.901329517364502, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8791913986206055, "num_tokens": 730463989.0, "step": 19234 }, { "epoch": 2.4468897086884622, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.2878313064575195, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.872275710105896, "num_tokens": 730500854.0, "step": 19235 }, { "epoch": 2.4470169189670523, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.9178824424743652, "learning_rate": 1e-05, "loss": 0.3955, "mean_token_accuracy": 0.895538866519928, "num_tokens": 730537742.0, "step": 19236 }, { "epoch": 2.4471441292456433, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.066356658935547, "learning_rate": 1e-05, "loss": 0.3843, "mean_token_accuracy": 0.8976960778236389, "num_tokens": 730573090.0, "step": 19237 }, { "epoch": 2.4472713395242334, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7954468727111816, "learning_rate": 1e-05, "loss": 0.3905, "mean_token_accuracy": 0.8961688280105591, "num_tokens": 730611865.0, "step": 19238 }, { "epoch": 2.447398549802824, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.929572582244873, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8610525727272034, "num_tokens": 730646831.0, "step": 19239 }, { "epoch": 2.4475257600814144, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.661968946456909, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8825910091400146, "num_tokens": 730689789.0, "step": 19240 }, { "epoch": 2.447652970360005, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8611104488372803, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.881072461605072, "num_tokens": 730720289.0, "step": 19241 }, { "epoch": 2.4477801806385955, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.531344175338745, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8868378400802612, "num_tokens": 730757499.0, "step": 19242 }, { "epoch": 2.447907390917186, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8416168689727783, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8842799663543701, "num_tokens": 730799897.0, "step": 19243 }, { "epoch": 2.4480346011957765, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.425914764404297, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.885985255241394, "num_tokens": 730837709.0, "step": 19244 }, { "epoch": 2.448161811474367, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.91606068611145, "learning_rate": 1e-05, "loss": 0.379, "mean_token_accuracy": 0.8974651098251343, "num_tokens": 730874128.0, "step": 19245 }, { "epoch": 2.4482890217529576, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.363271951675415, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8742579221725464, "num_tokens": 730917027.0, "step": 19246 }, { "epoch": 2.448416232031548, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.058371067047119, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8672983646392822, "num_tokens": 730955107.0, "step": 19247 }, { "epoch": 2.4485434423101387, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3950202465057373, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8686634302139282, "num_tokens": 730995807.0, "step": 19248 }, { "epoch": 2.448670652588729, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.149117946624756, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.8924987316131592, "num_tokens": 731034935.0, "step": 19249 }, { "epoch": 2.4487978628673197, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5205535888671875, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8775925636291504, "num_tokens": 731070934.0, "step": 19250 }, { "epoch": 2.4489250731459102, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.11048698425293, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8876475691795349, "num_tokens": 731112625.0, "step": 19251 }, { "epoch": 2.4490522834245008, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.646702289581299, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8748223781585693, "num_tokens": 731151011.0, "step": 19252 }, { "epoch": 2.4491794937030913, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.948044776916504, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8819156885147095, "num_tokens": 731190885.0, "step": 19253 }, { "epoch": 2.449306703981682, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6778299808502197, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.881635844707489, "num_tokens": 731229056.0, "step": 19254 }, { "epoch": 2.4494339142602723, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7472524642944336, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8795382976531982, "num_tokens": 731268098.0, "step": 19255 }, { "epoch": 2.449561124538863, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.669448137283325, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8656725287437439, "num_tokens": 731307809.0, "step": 19256 }, { "epoch": 2.4496883348174534, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.576871633529663, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8723676204681396, "num_tokens": 731343904.0, "step": 19257 }, { "epoch": 2.449815545096044, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.641308546066284, "learning_rate": 1e-05, "loss": 0.3857, "mean_token_accuracy": 0.8946466445922852, "num_tokens": 731383120.0, "step": 19258 }, { "epoch": 2.4499427553746345, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.449090003967285, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.880357563495636, "num_tokens": 731423020.0, "step": 19259 }, { "epoch": 2.450069965653225, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.643035650253296, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8767063617706299, "num_tokens": 731463522.0, "step": 19260 }, { "epoch": 2.450197175931815, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.387016534805298, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8685131669044495, "num_tokens": 731502327.0, "step": 19261 }, { "epoch": 2.450324386210406, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6566174030303955, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8904433250427246, "num_tokens": 731545253.0, "step": 19262 }, { "epoch": 2.450451596488996, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3155229091644287, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.8870140910148621, "num_tokens": 731575388.0, "step": 19263 }, { "epoch": 2.4505788067675867, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7196080684661865, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8774022459983826, "num_tokens": 731618195.0, "step": 19264 }, { "epoch": 2.450706017046177, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.257880449295044, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8716250658035278, "num_tokens": 731658551.0, "step": 19265 }, { "epoch": 2.4508332273247677, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.824369192123413, "learning_rate": 1e-05, "loss": 0.3933, "mean_token_accuracy": 0.8922656178474426, "num_tokens": 731700477.0, "step": 19266 }, { "epoch": 2.4509604376033582, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2185604572296143, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8823932409286499, "num_tokens": 731735841.0, "step": 19267 }, { "epoch": 2.4510876478819488, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.008356094360352, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8745483160018921, "num_tokens": 731774974.0, "step": 19268 }, { "epoch": 2.4512148581605393, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2369611263275146, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8849325180053711, "num_tokens": 731806010.0, "step": 19269 }, { "epoch": 2.45134206843913, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.203363418579102, "learning_rate": 1e-05, "loss": 0.3913, "mean_token_accuracy": 0.89590984582901, "num_tokens": 731841773.0, "step": 19270 }, { "epoch": 2.4514692787177204, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.373920202255249, "learning_rate": 1e-05, "loss": 0.4071, "mean_token_accuracy": 0.8876648545265198, "num_tokens": 731877974.0, "step": 19271 }, { "epoch": 2.451596488996311, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.303354263305664, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8805188536643982, "num_tokens": 731919375.0, "step": 19272 }, { "epoch": 2.4517236992749014, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6048343181610107, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8813087344169617, "num_tokens": 731957518.0, "step": 19273 }, { "epoch": 2.451850909553492, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.219889163970947, "learning_rate": 1e-05, "loss": 0.4121, "mean_token_accuracy": 0.8886032104492188, "num_tokens": 731996597.0, "step": 19274 }, { "epoch": 2.4519781198320825, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7203402519226074, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8794293403625488, "num_tokens": 732035888.0, "step": 19275 }, { "epoch": 2.452105330110673, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.046003818511963, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8779685497283936, "num_tokens": 732073026.0, "step": 19276 }, { "epoch": 2.4522325403892635, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7410202026367188, "learning_rate": 1e-05, "loss": 0.5094, "mean_token_accuracy": 0.8564368486404419, "num_tokens": 732110794.0, "step": 19277 }, { "epoch": 2.452359750667854, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.836322546005249, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8872563242912292, "num_tokens": 732141051.0, "step": 19278 }, { "epoch": 2.4524869609464446, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.66255784034729, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8851544857025146, "num_tokens": 732187023.0, "step": 19279 }, { "epoch": 2.452614171225035, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6970179080963135, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.8676482439041138, "num_tokens": 732220960.0, "step": 19280 }, { "epoch": 2.4527413815036256, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.602720022201538, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8796489834785461, "num_tokens": 732260656.0, "step": 19281 }, { "epoch": 2.452868591782216, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.613652229309082, "learning_rate": 1e-05, "loss": 0.5082, "mean_token_accuracy": 0.8533056974411011, "num_tokens": 732301771.0, "step": 19282 }, { "epoch": 2.4529958020608067, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5364112854003906, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8732424378395081, "num_tokens": 732341618.0, "step": 19283 }, { "epoch": 2.453123012339397, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5559511184692383, "learning_rate": 1e-05, "loss": 0.3416, "mean_token_accuracy": 0.9093781113624573, "num_tokens": 732375797.0, "step": 19284 }, { "epoch": 2.4532502226179878, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4759366512298584, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8885432481765747, "num_tokens": 732414234.0, "step": 19285 }, { "epoch": 2.453377432896578, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5407493114471436, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8750317096710205, "num_tokens": 732461063.0, "step": 19286 }, { "epoch": 2.4535046431751684, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.425069570541382, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8816123008728027, "num_tokens": 732501622.0, "step": 19287 }, { "epoch": 2.453631853453759, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5639684200286865, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8731051683425903, "num_tokens": 732546938.0, "step": 19288 }, { "epoch": 2.4537590637323494, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.372424364089966, "learning_rate": 1e-05, "loss": 0.378, "mean_token_accuracy": 0.8957071304321289, "num_tokens": 732581250.0, "step": 19289 }, { "epoch": 2.45388627401094, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6147656440734863, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8900620937347412, "num_tokens": 732621081.0, "step": 19290 }, { "epoch": 2.4540134842895305, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.301328659057617, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.8872361183166504, "num_tokens": 732660625.0, "step": 19291 }, { "epoch": 2.454140694568121, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.713191032409668, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8829998970031738, "num_tokens": 732700100.0, "step": 19292 }, { "epoch": 2.4542679048467115, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.23226261138916, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8815470933914185, "num_tokens": 732735252.0, "step": 19293 }, { "epoch": 2.454395115125302, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.896406888961792, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8735599517822266, "num_tokens": 732767912.0, "step": 19294 }, { "epoch": 2.4545223254038926, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1751322746276855, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8729315996170044, "num_tokens": 732808499.0, "step": 19295 }, { "epoch": 2.454649535682483, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.14461088180542, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8781987428665161, "num_tokens": 732840754.0, "step": 19296 }, { "epoch": 2.4547767459610736, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.2538652420043945, "learning_rate": 1e-05, "loss": 0.4912, "mean_token_accuracy": 0.8594523668289185, "num_tokens": 732882432.0, "step": 19297 }, { "epoch": 2.454903956239664, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.375105381011963, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.889246940612793, "num_tokens": 732912418.0, "step": 19298 }, { "epoch": 2.4550311665182547, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5342633724212646, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8817278146743774, "num_tokens": 732946050.0, "step": 19299 }, { "epoch": 2.4551583767968452, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.376354217529297, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8867605924606323, "num_tokens": 732980667.0, "step": 19300 }, { "epoch": 2.4552855870754358, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.770871877670288, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8813985586166382, "num_tokens": 733016169.0, "step": 19301 }, { "epoch": 2.4554127973540263, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.187981605529785, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8752627372741699, "num_tokens": 733055120.0, "step": 19302 }, { "epoch": 2.455540007632617, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.802690267562866, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.8687504529953003, "num_tokens": 733094550.0, "step": 19303 }, { "epoch": 2.4556672179112073, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9716949462890625, "learning_rate": 1e-05, "loss": 0.3886, "mean_token_accuracy": 0.8929083943367004, "num_tokens": 733132338.0, "step": 19304 }, { "epoch": 2.455794428189798, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.734354019165039, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8764352798461914, "num_tokens": 733167053.0, "step": 19305 }, { "epoch": 2.4559216384683884, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.815471887588501, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8693022727966309, "num_tokens": 733208175.0, "step": 19306 }, { "epoch": 2.456048848746979, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6286838054656982, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8686051368713379, "num_tokens": 733248336.0, "step": 19307 }, { "epoch": 2.4561760590255695, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7361721992492676, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.878283679485321, "num_tokens": 733283314.0, "step": 19308 }, { "epoch": 2.4563032693041595, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5315701961517334, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8777250051498413, "num_tokens": 733317541.0, "step": 19309 }, { "epoch": 2.4564304795827505, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6935901641845703, "learning_rate": 1e-05, "loss": 0.3699, "mean_token_accuracy": 0.898917019367218, "num_tokens": 733349975.0, "step": 19310 }, { "epoch": 2.4565576898613406, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4384701251983643, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8911231160163879, "num_tokens": 733388085.0, "step": 19311 }, { "epoch": 2.456684900139931, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.715101957321167, "learning_rate": 1e-05, "loss": 0.3997, "mean_token_accuracy": 0.8887107372283936, "num_tokens": 733428420.0, "step": 19312 }, { "epoch": 2.4568121104185217, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3664987087249756, "learning_rate": 1e-05, "loss": 0.4038, "mean_token_accuracy": 0.8840475082397461, "num_tokens": 733460271.0, "step": 19313 }, { "epoch": 2.456939320697112, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.783839464187622, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8792247772216797, "num_tokens": 733498610.0, "step": 19314 }, { "epoch": 2.4570665309757027, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3187386989593506, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8730186223983765, "num_tokens": 733530501.0, "step": 19315 }, { "epoch": 2.4571937412542932, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.890591859817505, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.885706901550293, "num_tokens": 733563208.0, "step": 19316 }, { "epoch": 2.4573209515328838, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.289168119430542, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.887174129486084, "num_tokens": 733598729.0, "step": 19317 }, { "epoch": 2.4574481618114743, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0432257652282715, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8770197629928589, "num_tokens": 733632033.0, "step": 19318 }, { "epoch": 2.457575372090065, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3380086421966553, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8730456829071045, "num_tokens": 733673088.0, "step": 19319 }, { "epoch": 2.4577025823686554, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.14830207824707, "learning_rate": 1e-05, "loss": 0.4756, "mean_token_accuracy": 0.870324969291687, "num_tokens": 733714096.0, "step": 19320 }, { "epoch": 2.457829792647246, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.468268394470215, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8722452521324158, "num_tokens": 733754495.0, "step": 19321 }, { "epoch": 2.4579570029258364, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.129392147064209, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8807246685028076, "num_tokens": 733788904.0, "step": 19322 }, { "epoch": 2.458084213204427, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5946567058563232, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8833396434783936, "num_tokens": 733831041.0, "step": 19323 }, { "epoch": 2.4582114234830175, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.236361026763916, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8692898750305176, "num_tokens": 733872209.0, "step": 19324 }, { "epoch": 2.458338633761608, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7227673530578613, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8722054958343506, "num_tokens": 733906884.0, "step": 19325 }, { "epoch": 2.4584658440401985, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.831540107727051, "learning_rate": 1e-05, "loss": 0.3912, "mean_token_accuracy": 0.894760012626648, "num_tokens": 733949262.0, "step": 19326 }, { "epoch": 2.458593054318789, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.670945167541504, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8864943981170654, "num_tokens": 733991111.0, "step": 19327 }, { "epoch": 2.4587202645973796, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6709330081939697, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8766297698020935, "num_tokens": 734030968.0, "step": 19328 }, { "epoch": 2.45884747487597, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.655893564224243, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8714537024497986, "num_tokens": 734070612.0, "step": 19329 }, { "epoch": 2.4589746851545606, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5145654678344727, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8827581405639648, "num_tokens": 734108682.0, "step": 19330 }, { "epoch": 2.459101895433151, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6463255882263184, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8845690488815308, "num_tokens": 734141565.0, "step": 19331 }, { "epoch": 2.4592291057117417, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.408583164215088, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8837833404541016, "num_tokens": 734183048.0, "step": 19332 }, { "epoch": 2.459356315990332, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.677335500717163, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8725113868713379, "num_tokens": 734216141.0, "step": 19333 }, { "epoch": 2.4594835262689223, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.331854820251465, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8799794316291809, "num_tokens": 734256365.0, "step": 19334 }, { "epoch": 2.4596107365475133, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7203776836395264, "learning_rate": 1e-05, "loss": 0.3981, "mean_token_accuracy": 0.8899784088134766, "num_tokens": 734297679.0, "step": 19335 }, { "epoch": 2.4597379468261034, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.286504030227661, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.865484356880188, "num_tokens": 734333245.0, "step": 19336 }, { "epoch": 2.459865157104694, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.827589273452759, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8702691793441772, "num_tokens": 734374408.0, "step": 19337 }, { "epoch": 2.4599923673832844, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2425527572631836, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8840748071670532, "num_tokens": 734408105.0, "step": 19338 }, { "epoch": 2.460119577661875, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.974867105484009, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8818472027778625, "num_tokens": 734451700.0, "step": 19339 }, { "epoch": 2.4602467879404655, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.237926959991455, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8715618252754211, "num_tokens": 734492343.0, "step": 19340 }, { "epoch": 2.460373998219056, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.161867618560791, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.880225658416748, "num_tokens": 734531720.0, "step": 19341 }, { "epoch": 2.4605012084976465, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3494858741760254, "learning_rate": 1e-05, "loss": 0.4144, "mean_token_accuracy": 0.883612871170044, "num_tokens": 734567809.0, "step": 19342 }, { "epoch": 2.460628418776237, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.2697834968566895, "learning_rate": 1e-05, "loss": 0.3777, "mean_token_accuracy": 0.8976526856422424, "num_tokens": 734605491.0, "step": 19343 }, { "epoch": 2.4607556290548276, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.562652111053467, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.8867075443267822, "num_tokens": 734637169.0, "step": 19344 }, { "epoch": 2.460882839333418, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2150468826293945, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8842322826385498, "num_tokens": 734668167.0, "step": 19345 }, { "epoch": 2.4610100496120086, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6795897483825684, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8939021825790405, "num_tokens": 734705641.0, "step": 19346 }, { "epoch": 2.461137259890599, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.048750400543213, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8760421872138977, "num_tokens": 734742412.0, "step": 19347 }, { "epoch": 2.4612644701691897, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6968894004821777, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8914309740066528, "num_tokens": 734782930.0, "step": 19348 }, { "epoch": 2.4613916804477802, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.866504669189453, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8885295391082764, "num_tokens": 734825051.0, "step": 19349 }, { "epoch": 2.4615188907263708, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.660640239715576, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8707606792449951, "num_tokens": 734863524.0, "step": 19350 }, { "epoch": 2.4616461010049613, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.716031312942505, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.873639702796936, "num_tokens": 734899692.0, "step": 19351 }, { "epoch": 2.461773311283552, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.611769437789917, "learning_rate": 1e-05, "loss": 0.4907, "mean_token_accuracy": 0.8632516860961914, "num_tokens": 734933114.0, "step": 19352 }, { "epoch": 2.4619005215621423, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.620687246322632, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8606958389282227, "num_tokens": 734975896.0, "step": 19353 }, { "epoch": 2.462027731840733, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.534065008163452, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8698455095291138, "num_tokens": 735017969.0, "step": 19354 }, { "epoch": 2.4621549421193234, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5778164863586426, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.875834584236145, "num_tokens": 735057052.0, "step": 19355 }, { "epoch": 2.462282152397914, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.470348834991455, "learning_rate": 1e-05, "loss": 0.3811, "mean_token_accuracy": 0.8950027227401733, "num_tokens": 735094211.0, "step": 19356 }, { "epoch": 2.4624093626765045, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5672779083251953, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8612527251243591, "num_tokens": 735134870.0, "step": 19357 }, { "epoch": 2.462536572955095, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.441863536834717, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8677780628204346, "num_tokens": 735177218.0, "step": 19358 }, { "epoch": 2.462663783233685, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.621145486831665, "learning_rate": 1e-05, "loss": 0.5141, "mean_token_accuracy": 0.8530902862548828, "num_tokens": 735208310.0, "step": 19359 }, { "epoch": 2.462790993512276, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3909404277801514, "learning_rate": 1e-05, "loss": 0.3922, "mean_token_accuracy": 0.887946367263794, "num_tokens": 735243514.0, "step": 19360 }, { "epoch": 2.462918203790866, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.633580446243286, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8734749555587769, "num_tokens": 735278765.0, "step": 19361 }, { "epoch": 2.4630454140694567, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3045568466186523, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8644958734512329, "num_tokens": 735317966.0, "step": 19362 }, { "epoch": 2.463172624348047, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.715487003326416, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8834668397903442, "num_tokens": 735355552.0, "step": 19363 }, { "epoch": 2.4632998346266377, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2340097427368164, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8722862005233765, "num_tokens": 735398011.0, "step": 19364 }, { "epoch": 2.4634270449052282, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.875718593597412, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8921362161636353, "num_tokens": 735436743.0, "step": 19365 }, { "epoch": 2.4635542551838188, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2077369689941406, "learning_rate": 1e-05, "loss": 0.3849, "mean_token_accuracy": 0.8931657075881958, "num_tokens": 735467932.0, "step": 19366 }, { "epoch": 2.4636814654624093, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.066353797912598, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8893628120422363, "num_tokens": 735510057.0, "step": 19367 }, { "epoch": 2.463808675741, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.243171215057373, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8755224943161011, "num_tokens": 735553700.0, "step": 19368 }, { "epoch": 2.4639358860195903, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.280985355377197, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8620548248291016, "num_tokens": 735598490.0, "step": 19369 }, { "epoch": 2.464063096298181, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.4592721462249756, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8747254610061646, "num_tokens": 735637987.0, "step": 19370 }, { "epoch": 2.4641903065767714, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.32183837890625, "learning_rate": 1e-05, "loss": 0.3838, "mean_token_accuracy": 0.9005386829376221, "num_tokens": 735676327.0, "step": 19371 }, { "epoch": 2.464317516855362, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.6926984786987305, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8765954971313477, "num_tokens": 735713015.0, "step": 19372 }, { "epoch": 2.4644447271339525, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.17174768447876, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8808245658874512, "num_tokens": 735750561.0, "step": 19373 }, { "epoch": 2.464571937412543, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7655322551727295, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8706227540969849, "num_tokens": 735788264.0, "step": 19374 }, { "epoch": 2.4646991476911335, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9574644565582275, "learning_rate": 1e-05, "loss": 0.4546, "mean_token_accuracy": 0.8794369697570801, "num_tokens": 735829384.0, "step": 19375 }, { "epoch": 2.464826357969724, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.700173854827881, "learning_rate": 1e-05, "loss": 0.3911, "mean_token_accuracy": 0.8942000865936279, "num_tokens": 735868323.0, "step": 19376 }, { "epoch": 2.4649535682483146, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.7907350063323975, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8896658420562744, "num_tokens": 735902107.0, "step": 19377 }, { "epoch": 2.465080778526905, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6317529678344727, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8820155262947083, "num_tokens": 735941501.0, "step": 19378 }, { "epoch": 2.4652079888054956, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.674046754837036, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8741244077682495, "num_tokens": 735987721.0, "step": 19379 }, { "epoch": 2.465335199084086, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.569295883178711, "learning_rate": 1e-05, "loss": 0.3926, "mean_token_accuracy": 0.8933773636817932, "num_tokens": 736022126.0, "step": 19380 }, { "epoch": 2.4654624093626767, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6097018718719482, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8857876062393188, "num_tokens": 736059782.0, "step": 19381 }, { "epoch": 2.4655896196412668, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4959235191345215, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8664042949676514, "num_tokens": 736101514.0, "step": 19382 }, { "epoch": 2.4657168299198577, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5871808528900146, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8866590261459351, "num_tokens": 736137861.0, "step": 19383 }, { "epoch": 2.465844040198448, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.428600549697876, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.875556468963623, "num_tokens": 736177567.0, "step": 19384 }, { "epoch": 2.4659712504770384, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.612722396850586, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8802109360694885, "num_tokens": 736214893.0, "step": 19385 }, { "epoch": 2.466098460755629, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3710031509399414, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8681520223617554, "num_tokens": 736250539.0, "step": 19386 }, { "epoch": 2.4662256710342194, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6566264629364014, "learning_rate": 1e-05, "loss": 0.3919, "mean_token_accuracy": 0.8924369215965271, "num_tokens": 736280924.0, "step": 19387 }, { "epoch": 2.46635288131281, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.292135238647461, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8796373605728149, "num_tokens": 736322446.0, "step": 19388 }, { "epoch": 2.4664800915914005, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7583281993865967, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.883959174156189, "num_tokens": 736361370.0, "step": 19389 }, { "epoch": 2.466607301869991, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2269914150238037, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8814184665679932, "num_tokens": 736401931.0, "step": 19390 }, { "epoch": 2.4667345121485815, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9337265491485596, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8818278312683105, "num_tokens": 736445125.0, "step": 19391 }, { "epoch": 2.466861722427172, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.1919844150543213, "learning_rate": 1e-05, "loss": 0.3724, "mean_token_accuracy": 0.8999394178390503, "num_tokens": 736488409.0, "step": 19392 }, { "epoch": 2.4669889327057626, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.163196563720703, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8765710592269897, "num_tokens": 736526979.0, "step": 19393 }, { "epoch": 2.467116142984353, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.2987325191497803, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8884297609329224, "num_tokens": 736571933.0, "step": 19394 }, { "epoch": 2.4672433532629436, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.346904754638672, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8638156652450562, "num_tokens": 736610423.0, "step": 19395 }, { "epoch": 2.467370563541534, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.587956428527832, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8716076612472534, "num_tokens": 736648432.0, "step": 19396 }, { "epoch": 2.4674977738201247, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.272757053375244, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.860573410987854, "num_tokens": 736688032.0, "step": 19397 }, { "epoch": 2.4676249840987152, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7569918632507324, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8786864280700684, "num_tokens": 736725785.0, "step": 19398 }, { "epoch": 2.4677521943773058, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.066715240478516, "learning_rate": 1e-05, "loss": 0.5007, "mean_token_accuracy": 0.8591244220733643, "num_tokens": 736760229.0, "step": 19399 }, { "epoch": 2.4678794046558963, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7558717727661133, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8813490271568298, "num_tokens": 736794981.0, "step": 19400 }, { "epoch": 2.468006614934487, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8521840572357178, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8791438937187195, "num_tokens": 736835387.0, "step": 19401 }, { "epoch": 2.4681338252130773, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6683242321014404, "learning_rate": 1e-05, "loss": 0.3798, "mean_token_accuracy": 0.898883044719696, "num_tokens": 736877047.0, "step": 19402 }, { "epoch": 2.468261035491668, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7093985080718994, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8727695345878601, "num_tokens": 736916637.0, "step": 19403 }, { "epoch": 2.4683882457702584, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5977184772491455, "learning_rate": 1e-05, "loss": 0.392, "mean_token_accuracy": 0.8879047632217407, "num_tokens": 736954905.0, "step": 19404 }, { "epoch": 2.468515456048849, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6206953525543213, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.887911856174469, "num_tokens": 736992264.0, "step": 19405 }, { "epoch": 2.4686426663274394, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.528521776199341, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8629417419433594, "num_tokens": 737032896.0, "step": 19406 }, { "epoch": 2.4687698766060295, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.576817512512207, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8863730430603027, "num_tokens": 737073734.0, "step": 19407 }, { "epoch": 2.4688970868846205, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4700679779052734, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8839769959449768, "num_tokens": 737112246.0, "step": 19408 }, { "epoch": 2.4690242971632106, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.582277536392212, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8726925253868103, "num_tokens": 737148378.0, "step": 19409 }, { "epoch": 2.469151507441801, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.420085906982422, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8668767809867859, "num_tokens": 737182815.0, "step": 19410 }, { "epoch": 2.4692787177203916, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5867764949798584, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8846197128295898, "num_tokens": 737223301.0, "step": 19411 }, { "epoch": 2.469405927998982, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.374676465988159, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8732871413230896, "num_tokens": 737254184.0, "step": 19412 }, { "epoch": 2.4695331382775727, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6245853900909424, "learning_rate": 1e-05, "loss": 0.3684, "mean_token_accuracy": 0.8983607292175293, "num_tokens": 737292540.0, "step": 19413 }, { "epoch": 2.4696603485561632, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.304213285446167, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8642842769622803, "num_tokens": 737329822.0, "step": 19414 }, { "epoch": 2.4697875588347538, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7255666255950928, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8872241973876953, "num_tokens": 737365416.0, "step": 19415 }, { "epoch": 2.4699147691133443, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2183568477630615, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8823572397232056, "num_tokens": 737411498.0, "step": 19416 }, { "epoch": 2.470041979391935, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8959293365478516, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8872947692871094, "num_tokens": 737453667.0, "step": 19417 }, { "epoch": 2.4701691896705253, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.184786796569824, "learning_rate": 1e-05, "loss": 0.3951, "mean_token_accuracy": 0.8860670924186707, "num_tokens": 737487095.0, "step": 19418 }, { "epoch": 2.470296399949116, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.136838436126709, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8890352845191956, "num_tokens": 737521513.0, "step": 19419 }, { "epoch": 2.4704236102277064, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.2539219856262207, "learning_rate": 1e-05, "loss": 0.419, "mean_token_accuracy": 0.8873867988586426, "num_tokens": 737559218.0, "step": 19420 }, { "epoch": 2.470550820506297, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.357967376708984, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8695424795150757, "num_tokens": 737600486.0, "step": 19421 }, { "epoch": 2.4706780307848875, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.516728639602661, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8670141696929932, "num_tokens": 737639035.0, "step": 19422 }, { "epoch": 2.470805241063478, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.363523006439209, "learning_rate": 1e-05, "loss": 0.3843, "mean_token_accuracy": 0.8979295492172241, "num_tokens": 737676582.0, "step": 19423 }, { "epoch": 2.4709324513420685, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7534403800964355, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8827486038208008, "num_tokens": 737714809.0, "step": 19424 }, { "epoch": 2.471059661620659, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.183228492736816, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8754177093505859, "num_tokens": 737757803.0, "step": 19425 }, { "epoch": 2.4711868718992496, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.795624256134033, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8781386613845825, "num_tokens": 737796563.0, "step": 19426 }, { "epoch": 2.47131408217784, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9833879470825195, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8720459938049316, "num_tokens": 737829711.0, "step": 19427 }, { "epoch": 2.4714412924564306, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.7420098781585693, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.883305549621582, "num_tokens": 737862632.0, "step": 19428 }, { "epoch": 2.471568502735021, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.807471513748169, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8790661692619324, "num_tokens": 737900188.0, "step": 19429 }, { "epoch": 2.4716957130136117, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6502268314361572, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8859134912490845, "num_tokens": 737934667.0, "step": 19430 }, { "epoch": 2.471822923292202, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.690537929534912, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8858911991119385, "num_tokens": 737968386.0, "step": 19431 }, { "epoch": 2.4719501335707923, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.550956964492798, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8772232532501221, "num_tokens": 738011761.0, "step": 19432 }, { "epoch": 2.4720773438493833, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6421875953674316, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.885462760925293, "num_tokens": 738049319.0, "step": 19433 }, { "epoch": 2.4722045541279734, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.476393222808838, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8752685785293579, "num_tokens": 738083725.0, "step": 19434 }, { "epoch": 2.472331764406564, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.633333444595337, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8720014095306396, "num_tokens": 738126330.0, "step": 19435 }, { "epoch": 2.4724589746851544, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4029455184936523, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8766753673553467, "num_tokens": 738163564.0, "step": 19436 }, { "epoch": 2.472586184963745, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.682039737701416, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8727148771286011, "num_tokens": 738197014.0, "step": 19437 }, { "epoch": 2.4727133952423355, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3140721321105957, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8807446956634521, "num_tokens": 738232263.0, "step": 19438 }, { "epoch": 2.472840605520926, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7611734867095947, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8895114660263062, "num_tokens": 738273839.0, "step": 19439 }, { "epoch": 2.4729678157995165, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2417643070220947, "learning_rate": 1e-05, "loss": 0.3783, "mean_token_accuracy": 0.8931394815444946, "num_tokens": 738314062.0, "step": 19440 }, { "epoch": 2.473095026078107, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.933983087539673, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8734854459762573, "num_tokens": 738352330.0, "step": 19441 }, { "epoch": 2.4732222363566976, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2352709770202637, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.860689103603363, "num_tokens": 738388527.0, "step": 19442 }, { "epoch": 2.473349446635288, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.136092185974121, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8737578392028809, "num_tokens": 738422913.0, "step": 19443 }, { "epoch": 2.4734766569138786, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3347854614257812, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8729724884033203, "num_tokens": 738463739.0, "step": 19444 }, { "epoch": 2.473603867192469, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.259738922119141, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.868869423866272, "num_tokens": 738499160.0, "step": 19445 }, { "epoch": 2.4737310774710597, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5283241271972656, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.888664960861206, "num_tokens": 738546634.0, "step": 19446 }, { "epoch": 2.47385828774965, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.223698616027832, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8859739899635315, "num_tokens": 738581021.0, "step": 19447 }, { "epoch": 2.4739854980282407, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7029199600219727, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8803083896636963, "num_tokens": 738622606.0, "step": 19448 }, { "epoch": 2.4741127083068313, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.031083583831787, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8775275945663452, "num_tokens": 738656244.0, "step": 19449 }, { "epoch": 2.474239918585422, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7283194065093994, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8699508905410767, "num_tokens": 738694743.0, "step": 19450 }, { "epoch": 2.4743671288640123, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8237242698669434, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8739135265350342, "num_tokens": 738731447.0, "step": 19451 }, { "epoch": 2.474494339142603, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6824581623077393, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8892601132392883, "num_tokens": 738764631.0, "step": 19452 }, { "epoch": 2.4746215494211934, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.641648292541504, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8882520198822021, "num_tokens": 738807685.0, "step": 19453 }, { "epoch": 2.474748759699784, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6352553367614746, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8761134147644043, "num_tokens": 738851717.0, "step": 19454 }, { "epoch": 2.4748759699783744, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5420479774475098, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8642277717590332, "num_tokens": 738889432.0, "step": 19455 }, { "epoch": 2.475003180256965, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.609450340270996, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8726378679275513, "num_tokens": 738923197.0, "step": 19456 }, { "epoch": 2.475130390535555, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4605777263641357, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8835904598236084, "num_tokens": 738959699.0, "step": 19457 }, { "epoch": 2.475257600814146, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.589956521987915, "learning_rate": 1e-05, "loss": 0.3893, "mean_token_accuracy": 0.8943518400192261, "num_tokens": 738990444.0, "step": 19458 }, { "epoch": 2.475384811092736, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.395209550857544, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.889929473400116, "num_tokens": 739030864.0, "step": 19459 }, { "epoch": 2.4755120213713266, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6176724433898926, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8733729124069214, "num_tokens": 739068407.0, "step": 19460 }, { "epoch": 2.475639231649917, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3402035236358643, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8827296495437622, "num_tokens": 739105560.0, "step": 19461 }, { "epoch": 2.4757664419285077, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6729135513305664, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8682059049606323, "num_tokens": 739142056.0, "step": 19462 }, { "epoch": 2.4758936522070982, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2778189182281494, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8594474792480469, "num_tokens": 739185407.0, "step": 19463 }, { "epoch": 2.4760208624856888, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7800004482269287, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8743892908096313, "num_tokens": 739222594.0, "step": 19464 }, { "epoch": 2.4761480727642793, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2159950733184814, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8828035593032837, "num_tokens": 739261021.0, "step": 19465 }, { "epoch": 2.47627528304287, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.954224109649658, "learning_rate": 1e-05, "loss": 0.4603, "mean_token_accuracy": 0.8753265142440796, "num_tokens": 739300828.0, "step": 19466 }, { "epoch": 2.4764024933214603, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.184915065765381, "learning_rate": 1e-05, "loss": 0.3977, "mean_token_accuracy": 0.8880733847618103, "num_tokens": 739338724.0, "step": 19467 }, { "epoch": 2.476529703600051, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.191633701324463, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8878580331802368, "num_tokens": 739377821.0, "step": 19468 }, { "epoch": 2.4766569138786414, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3000898361206055, "learning_rate": 1e-05, "loss": 0.3907, "mean_token_accuracy": 0.8917180895805359, "num_tokens": 739411438.0, "step": 19469 }, { "epoch": 2.476784124157232, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.3879923820495605, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8699776530265808, "num_tokens": 739457836.0, "step": 19470 }, { "epoch": 2.4769113344358225, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.5877697467803955, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8701494932174683, "num_tokens": 739496336.0, "step": 19471 }, { "epoch": 2.477038544714413, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.3397932052612305, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8808872699737549, "num_tokens": 739539777.0, "step": 19472 }, { "epoch": 2.4771657549930035, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.783473014831543, "learning_rate": 1e-05, "loss": 0.3968, "mean_token_accuracy": 0.8947478532791138, "num_tokens": 739572560.0, "step": 19473 }, { "epoch": 2.477292965271594, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.151119232177734, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8859495520591736, "num_tokens": 739604260.0, "step": 19474 }, { "epoch": 2.4774201755501846, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7877259254455566, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8791952729225159, "num_tokens": 739645241.0, "step": 19475 }, { "epoch": 2.477547385828775, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.954744338989258, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8821821808815002, "num_tokens": 739683516.0, "step": 19476 }, { "epoch": 2.4776745961073656, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.7137887477874756, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8773925304412842, "num_tokens": 739720422.0, "step": 19477 }, { "epoch": 2.477801806385956, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.801715612411499, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8932750225067139, "num_tokens": 739758071.0, "step": 19478 }, { "epoch": 2.4779290166645467, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6208598613739014, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8830867409706116, "num_tokens": 739796143.0, "step": 19479 }, { "epoch": 2.4780562269431368, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.713273525238037, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8751876354217529, "num_tokens": 739832038.0, "step": 19480 }, { "epoch": 2.4781834372217277, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5178275108337402, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8832215666770935, "num_tokens": 739873232.0, "step": 19481 }, { "epoch": 2.478310647500318, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.683755397796631, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8861681818962097, "num_tokens": 739912213.0, "step": 19482 }, { "epoch": 2.4784378577789083, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.43365478515625, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8763381838798523, "num_tokens": 739949363.0, "step": 19483 }, { "epoch": 2.478565068057499, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.723360538482666, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8821463584899902, "num_tokens": 739979360.0, "step": 19484 }, { "epoch": 2.4786922783360894, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3533921241760254, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.864203929901123, "num_tokens": 740019691.0, "step": 19485 }, { "epoch": 2.47881948861468, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.800135612487793, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8861726522445679, "num_tokens": 740053604.0, "step": 19486 }, { "epoch": 2.4789466988932705, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2941319942474365, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8756749629974365, "num_tokens": 740090220.0, "step": 19487 }, { "epoch": 2.479073909171861, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9276480674743652, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8916483521461487, "num_tokens": 740126361.0, "step": 19488 }, { "epoch": 2.4792011194504515, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.274385929107666, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8786990642547607, "num_tokens": 740164752.0, "step": 19489 }, { "epoch": 2.479328329729042, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.090823650360107, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8832421898841858, "num_tokens": 740199622.0, "step": 19490 }, { "epoch": 2.4794555400076326, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3610188961029053, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8794606924057007, "num_tokens": 740236457.0, "step": 19491 }, { "epoch": 2.479582750286223, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.197474479675293, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8713313341140747, "num_tokens": 740275094.0, "step": 19492 }, { "epoch": 2.4797099605648136, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5390658378601074, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.87693190574646, "num_tokens": 740308564.0, "step": 19493 }, { "epoch": 2.479837170843404, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.120596885681152, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8799367547035217, "num_tokens": 740354994.0, "step": 19494 }, { "epoch": 2.4799643811219947, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.646047353744507, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8895683884620667, "num_tokens": 740394924.0, "step": 19495 }, { "epoch": 2.480091591400585, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.955331325531006, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8901709318161011, "num_tokens": 740428362.0, "step": 19496 }, { "epoch": 2.4802188016791757, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6752545833587646, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.870375394821167, "num_tokens": 740473987.0, "step": 19497 }, { "epoch": 2.4803460119577663, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7549502849578857, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.864136815071106, "num_tokens": 740512458.0, "step": 19498 }, { "epoch": 2.480473222236357, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.652844190597534, "learning_rate": 1e-05, "loss": 0.4943, "mean_token_accuracy": 0.861523449420929, "num_tokens": 740558744.0, "step": 19499 }, { "epoch": 2.4806004325149473, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.583167314529419, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8685677647590637, "num_tokens": 740599889.0, "step": 19500 }, { "epoch": 2.480727642793538, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.637998580932617, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8802410960197449, "num_tokens": 740636162.0, "step": 19501 }, { "epoch": 2.4808548530721284, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4615633487701416, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8624900579452515, "num_tokens": 740679595.0, "step": 19502 }, { "epoch": 2.480982063350719, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6596124172210693, "learning_rate": 1e-05, "loss": 0.4063, "mean_token_accuracy": 0.8866994380950928, "num_tokens": 740714830.0, "step": 19503 }, { "epoch": 2.4811092736293094, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.378148317337036, "learning_rate": 1e-05, "loss": 0.4032, "mean_token_accuracy": 0.8842194676399231, "num_tokens": 740749748.0, "step": 19504 }, { "epoch": 2.4812364839078995, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6924734115600586, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8662281036376953, "num_tokens": 740785110.0, "step": 19505 }, { "epoch": 2.4813636941864905, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3043525218963623, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8829766511917114, "num_tokens": 740824432.0, "step": 19506 }, { "epoch": 2.4814909044650806, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.788268804550171, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8770202398300171, "num_tokens": 740861403.0, "step": 19507 }, { "epoch": 2.481618114743671, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2417192459106445, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8764718770980835, "num_tokens": 740899299.0, "step": 19508 }, { "epoch": 2.4817453250222616, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9356689453125, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8846739530563354, "num_tokens": 740934178.0, "step": 19509 }, { "epoch": 2.481872535300852, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.230689764022827, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.874748170375824, "num_tokens": 740967010.0, "step": 19510 }, { "epoch": 2.4819997455794427, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.121187210083008, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8758807182312012, "num_tokens": 741004655.0, "step": 19511 }, { "epoch": 2.4821269558580332, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.303436040878296, "learning_rate": 1e-05, "loss": 0.397, "mean_token_accuracy": 0.890313446521759, "num_tokens": 741037365.0, "step": 19512 }, { "epoch": 2.4822541661366238, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.279482364654541, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8750600814819336, "num_tokens": 741072304.0, "step": 19513 }, { "epoch": 2.4823813764152143, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.4964442253112793, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8831396698951721, "num_tokens": 741107187.0, "step": 19514 }, { "epoch": 2.482508586693805, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.278923511505127, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8765113353729248, "num_tokens": 741142152.0, "step": 19515 }, { "epoch": 2.4826357969723953, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6901440620422363, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8859272003173828, "num_tokens": 741175704.0, "step": 19516 }, { "epoch": 2.482763007250986, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1157050132751465, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.890562117099762, "num_tokens": 741212187.0, "step": 19517 }, { "epoch": 2.4828902175295764, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.735044002532959, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8798319101333618, "num_tokens": 741246830.0, "step": 19518 }, { "epoch": 2.483017427808167, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.918833017349243, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8760656714439392, "num_tokens": 741279672.0, "step": 19519 }, { "epoch": 2.4831446380867574, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.683104991912842, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8894165754318237, "num_tokens": 741318205.0, "step": 19520 }, { "epoch": 2.483271848365348, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7461960315704346, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8847429752349854, "num_tokens": 741356312.0, "step": 19521 }, { "epoch": 2.4833990586439385, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6271111965179443, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.870306670665741, "num_tokens": 741393370.0, "step": 19522 }, { "epoch": 2.483526268922529, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6490230560302734, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8680602312088013, "num_tokens": 741428533.0, "step": 19523 }, { "epoch": 2.4836534792011196, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.558893918991089, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8719096183776855, "num_tokens": 741464614.0, "step": 19524 }, { "epoch": 2.48378068947971, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5852622985839844, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8767080903053284, "num_tokens": 741499126.0, "step": 19525 }, { "epoch": 2.4839078997583006, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.483534336090088, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8842979669570923, "num_tokens": 741544381.0, "step": 19526 }, { "epoch": 2.484035110036891, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5611350536346436, "learning_rate": 1e-05, "loss": 0.3884, "mean_token_accuracy": 0.8949981927871704, "num_tokens": 741588627.0, "step": 19527 }, { "epoch": 2.4841623203154817, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4358599185943604, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8838686943054199, "num_tokens": 741631165.0, "step": 19528 }, { "epoch": 2.484289530594072, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5804591178894043, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.881084680557251, "num_tokens": 741668667.0, "step": 19529 }, { "epoch": 2.4844167408726623, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3863608837127686, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8808834552764893, "num_tokens": 741703625.0, "step": 19530 }, { "epoch": 2.4845439511512533, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.609351396560669, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8761422634124756, "num_tokens": 741744371.0, "step": 19531 }, { "epoch": 2.4846711614298433, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3222150802612305, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8714473247528076, "num_tokens": 741786733.0, "step": 19532 }, { "epoch": 2.484798371708434, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6923396587371826, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8795713186264038, "num_tokens": 741825693.0, "step": 19533 }, { "epoch": 2.4849255819870244, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.261903762817383, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8620162010192871, "num_tokens": 741866176.0, "step": 19534 }, { "epoch": 2.485052792265615, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8275177478790283, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8698679804801941, "num_tokens": 741902523.0, "step": 19535 }, { "epoch": 2.4851800025442055, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.207486152648926, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8703840970993042, "num_tokens": 741939005.0, "step": 19536 }, { "epoch": 2.485307212822796, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.105189323425293, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8634825348854065, "num_tokens": 741975012.0, "step": 19537 }, { "epoch": 2.4854344231013865, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2330424785614014, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8791823983192444, "num_tokens": 742008351.0, "step": 19538 }, { "epoch": 2.485561633379977, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.300274848937988, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8821356296539307, "num_tokens": 742045477.0, "step": 19539 }, { "epoch": 2.4856888436585676, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4431099891662598, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.866521954536438, "num_tokens": 742080070.0, "step": 19540 }, { "epoch": 2.485816053937158, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.3703227043151855, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8829502463340759, "num_tokens": 742116735.0, "step": 19541 }, { "epoch": 2.4859432642157486, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.701521396636963, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8775942325592041, "num_tokens": 742151492.0, "step": 19542 }, { "epoch": 2.486070474494339, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.219420433044434, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8912248611450195, "num_tokens": 742196913.0, "step": 19543 }, { "epoch": 2.4861976847729297, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.764230728149414, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8936901092529297, "num_tokens": 742241894.0, "step": 19544 }, { "epoch": 2.48632489505152, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.045097351074219, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8846980333328247, "num_tokens": 742276256.0, "step": 19545 }, { "epoch": 2.4864521053301107, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.726888656616211, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.8947296142578125, "num_tokens": 742316834.0, "step": 19546 }, { "epoch": 2.4865793156087013, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.862461805343628, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8813809156417847, "num_tokens": 742359132.0, "step": 19547 }, { "epoch": 2.486706525887292, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6568925380706787, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8633931279182434, "num_tokens": 742399746.0, "step": 19548 }, { "epoch": 2.4868337361658823, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7506766319274902, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8674812316894531, "num_tokens": 742435400.0, "step": 19549 }, { "epoch": 2.486960946444473, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5731332302093506, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.886298418045044, "num_tokens": 742472935.0, "step": 19550 }, { "epoch": 2.4870881567230634, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6724693775177, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8830289840698242, "num_tokens": 742516841.0, "step": 19551 }, { "epoch": 2.487215367001654, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.500715970993042, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8740640878677368, "num_tokens": 742551057.0, "step": 19552 }, { "epoch": 2.4873425772802444, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6696419715881348, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8759500980377197, "num_tokens": 742584891.0, "step": 19553 }, { "epoch": 2.487469787558835, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.410012722015381, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8853228092193604, "num_tokens": 742622574.0, "step": 19554 }, { "epoch": 2.487596997837425, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.691441059112549, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8788427710533142, "num_tokens": 742664458.0, "step": 19555 }, { "epoch": 2.487724208116016, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3348073959350586, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.88264000415802, "num_tokens": 742704006.0, "step": 19556 }, { "epoch": 2.487851418394606, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7960541248321533, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8683401346206665, "num_tokens": 742736885.0, "step": 19557 }, { "epoch": 2.4879786286731966, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.283315420150757, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.875578761100769, "num_tokens": 742773496.0, "step": 19558 }, { "epoch": 2.488105838951787, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9269354343414307, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8842507600784302, "num_tokens": 742812444.0, "step": 19559 }, { "epoch": 2.4882330492303777, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.26112961769104, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8781826496124268, "num_tokens": 742849910.0, "step": 19560 }, { "epoch": 2.488360259508968, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.090354919433594, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8898011445999146, "num_tokens": 742891412.0, "step": 19561 }, { "epoch": 2.4884874697875587, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3427793979644775, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8741716742515564, "num_tokens": 742929510.0, "step": 19562 }, { "epoch": 2.4886146800661493, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.20115852355957, "learning_rate": 1e-05, "loss": 0.4788, "mean_token_accuracy": 0.8691264986991882, "num_tokens": 742969767.0, "step": 19563 }, { "epoch": 2.48874189034474, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.526933431625366, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8877506256103516, "num_tokens": 743001416.0, "step": 19564 }, { "epoch": 2.4888691006233303, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.142538070678711, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8757346272468567, "num_tokens": 743037273.0, "step": 19565 }, { "epoch": 2.488996310901921, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6520400047302246, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8763300776481628, "num_tokens": 743072774.0, "step": 19566 }, { "epoch": 2.4891235211805114, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9651997089385986, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8675500154495239, "num_tokens": 743115931.0, "step": 19567 }, { "epoch": 2.489250731459102, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6682567596435547, "learning_rate": 1e-05, "loss": 0.4096, "mean_token_accuracy": 0.8903173208236694, "num_tokens": 743158500.0, "step": 19568 }, { "epoch": 2.4893779417376924, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7694575786590576, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8822635412216187, "num_tokens": 743195311.0, "step": 19569 }, { "epoch": 2.489505152016283, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6592836380004883, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8836795687675476, "num_tokens": 743227978.0, "step": 19570 }, { "epoch": 2.4896323622948735, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.604038715362549, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8846875429153442, "num_tokens": 743264297.0, "step": 19571 }, { "epoch": 2.489759572573464, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.628195285797119, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8741492033004761, "num_tokens": 743305351.0, "step": 19572 }, { "epoch": 2.4898867828520546, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.49674129486084, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8743817806243896, "num_tokens": 743345105.0, "step": 19573 }, { "epoch": 2.490013993130645, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.60964298248291, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8891432285308838, "num_tokens": 743383889.0, "step": 19574 }, { "epoch": 2.4901412034092356, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.422609329223633, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8796036243438721, "num_tokens": 743420045.0, "step": 19575 }, { "epoch": 2.490268413687826, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6163361072540283, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8781607151031494, "num_tokens": 743455358.0, "step": 19576 }, { "epoch": 2.4903956239664167, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3691492080688477, "learning_rate": 1e-05, "loss": 0.4787, "mean_token_accuracy": 0.8603667616844177, "num_tokens": 743492337.0, "step": 19577 }, { "epoch": 2.4905228342450068, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.658828020095825, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8868426084518433, "num_tokens": 743526023.0, "step": 19578 }, { "epoch": 2.4906500445235977, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3050575256347656, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8809552192687988, "num_tokens": 743560224.0, "step": 19579 }, { "epoch": 2.490777254802188, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7427804470062256, "learning_rate": 1e-05, "loss": 0.5103, "mean_token_accuracy": 0.8526403307914734, "num_tokens": 743601403.0, "step": 19580 }, { "epoch": 2.4909044650807783, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.241581678390503, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.882207989692688, "num_tokens": 743641677.0, "step": 19581 }, { "epoch": 2.491031675359369, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.878782272338867, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8784491419792175, "num_tokens": 743684615.0, "step": 19582 }, { "epoch": 2.4911588856379594, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2046329975128174, "learning_rate": 1e-05, "loss": 0.3768, "mean_token_accuracy": 0.8937441110610962, "num_tokens": 743718960.0, "step": 19583 }, { "epoch": 2.49128609591655, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.079375267028809, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8892394304275513, "num_tokens": 743758336.0, "step": 19584 }, { "epoch": 2.4914133061951405, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2329485416412354, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8700750470161438, "num_tokens": 743803243.0, "step": 19585 }, { "epoch": 2.491540516473731, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.307776927947998, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8850398659706116, "num_tokens": 743840945.0, "step": 19586 }, { "epoch": 2.4916677267523215, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4391913414001465, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8823490738868713, "num_tokens": 743884338.0, "step": 19587 }, { "epoch": 2.491794937030912, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.391774654388428, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8796648383140564, "num_tokens": 743925499.0, "step": 19588 }, { "epoch": 2.4919221473095026, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7273268699645996, "learning_rate": 1e-05, "loss": 0.4871, "mean_token_accuracy": 0.8648751974105835, "num_tokens": 743965016.0, "step": 19589 }, { "epoch": 2.492049357588093, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.237171649932861, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8730252981185913, "num_tokens": 744000523.0, "step": 19590 }, { "epoch": 2.4921765678666836, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.798297643661499, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8755988478660583, "num_tokens": 744043862.0, "step": 19591 }, { "epoch": 2.492303778145274, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.025251388549805, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8620887994766235, "num_tokens": 744085161.0, "step": 19592 }, { "epoch": 2.4924309884238647, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7569470405578613, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8819940686225891, "num_tokens": 744117575.0, "step": 19593 }, { "epoch": 2.492558198702455, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8479514122009277, "learning_rate": 1e-05, "loss": 0.5197, "mean_token_accuracy": 0.8537082076072693, "num_tokens": 744156312.0, "step": 19594 }, { "epoch": 2.4926854089810457, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6657049655914307, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8811079859733582, "num_tokens": 744191177.0, "step": 19595 }, { "epoch": 2.4928126192596363, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7181894779205322, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8738195896148682, "num_tokens": 744230407.0, "step": 19596 }, { "epoch": 2.492939829538227, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.577329397201538, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8753489255905151, "num_tokens": 744270265.0, "step": 19597 }, { "epoch": 2.4930670398168173, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6515817642211914, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8796879053115845, "num_tokens": 744310913.0, "step": 19598 }, { "epoch": 2.493194250095408, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5079662799835205, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8735463619232178, "num_tokens": 744342918.0, "step": 19599 }, { "epoch": 2.4933214603739984, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6379592418670654, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8759617805480957, "num_tokens": 744379000.0, "step": 19600 }, { "epoch": 2.493448670652589, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4258198738098145, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.8781524896621704, "num_tokens": 744414641.0, "step": 19601 }, { "epoch": 2.4935758809311794, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6847875118255615, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8725699186325073, "num_tokens": 744444860.0, "step": 19602 }, { "epoch": 2.4937030912097695, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3317198753356934, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8886412382125854, "num_tokens": 744479657.0, "step": 19603 }, { "epoch": 2.4938303014883605, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.760909080505371, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8794155716896057, "num_tokens": 744520048.0, "step": 19604 }, { "epoch": 2.4939575117669506, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2719357013702393, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8723554611206055, "num_tokens": 744555265.0, "step": 19605 }, { "epoch": 2.494084722045541, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9061615467071533, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8746048212051392, "num_tokens": 744591011.0, "step": 19606 }, { "epoch": 2.4942119323241316, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.238774538040161, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8736745715141296, "num_tokens": 744632302.0, "step": 19607 }, { "epoch": 2.494339142602722, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.100606918334961, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8885493278503418, "num_tokens": 744669827.0, "step": 19608 }, { "epoch": 2.4944663528813127, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3193624019622803, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.884009599685669, "num_tokens": 744708438.0, "step": 19609 }, { "epoch": 2.494593563159903, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2470598220825195, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8799145221710205, "num_tokens": 744744218.0, "step": 19610 }, { "epoch": 2.4947207734384937, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5093674659729004, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8840984106063843, "num_tokens": 744782805.0, "step": 19611 }, { "epoch": 2.4948479837170843, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.239476203918457, "learning_rate": 1e-05, "loss": 0.3814, "mean_token_accuracy": 0.8979370594024658, "num_tokens": 744814011.0, "step": 19612 }, { "epoch": 2.494975193995675, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.683288812637329, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8803209662437439, "num_tokens": 744857385.0, "step": 19613 }, { "epoch": 2.4951024042742653, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.062242031097412, "learning_rate": 1e-05, "loss": 0.5034, "mean_token_accuracy": 0.8604598045349121, "num_tokens": 744898212.0, "step": 19614 }, { "epoch": 2.495229614552856, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.737070083618164, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8764050006866455, "num_tokens": 744930665.0, "step": 19615 }, { "epoch": 2.4953568248314464, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8452188968658447, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8684940338134766, "num_tokens": 744968725.0, "step": 19616 }, { "epoch": 2.495484035110037, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6908252239227295, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8682252168655396, "num_tokens": 745008386.0, "step": 19617 }, { "epoch": 2.4956112453886274, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.665189743041992, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8742082118988037, "num_tokens": 745047059.0, "step": 19618 }, { "epoch": 2.495738455667218, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.648345708847046, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8705317974090576, "num_tokens": 745089768.0, "step": 19619 }, { "epoch": 2.4958656659458085, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.535881280899048, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.875593900680542, "num_tokens": 745130848.0, "step": 19620 }, { "epoch": 2.495992876224399, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6213433742523193, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8761925101280212, "num_tokens": 745166715.0, "step": 19621 }, { "epoch": 2.4961200865029896, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4761385917663574, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8706746101379395, "num_tokens": 745200718.0, "step": 19622 }, { "epoch": 2.49624729678158, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5857646465301514, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8708460330963135, "num_tokens": 745240736.0, "step": 19623 }, { "epoch": 2.4963745070601706, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4125711917877197, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8783387541770935, "num_tokens": 745279040.0, "step": 19624 }, { "epoch": 2.496501717338761, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5957257747650146, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8737353086471558, "num_tokens": 745316936.0, "step": 19625 }, { "epoch": 2.4966289276173517, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.367172956466675, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.8573225736618042, "num_tokens": 745355438.0, "step": 19626 }, { "epoch": 2.496756137895942, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.627661943435669, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8827979564666748, "num_tokens": 745394509.0, "step": 19627 }, { "epoch": 2.4968833481745323, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3142926692962646, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.876467764377594, "num_tokens": 745426347.0, "step": 19628 }, { "epoch": 2.4970105584531233, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6986324787139893, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.886651873588562, "num_tokens": 745457523.0, "step": 19629 }, { "epoch": 2.4971377687317133, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2543962001800537, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.86307692527771, "num_tokens": 745495155.0, "step": 19630 }, { "epoch": 2.497264979010304, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.805889129638672, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.873971164226532, "num_tokens": 745535466.0, "step": 19631 }, { "epoch": 2.4973921892888944, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1992862224578857, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8755919337272644, "num_tokens": 745569184.0, "step": 19632 }, { "epoch": 2.497519399567485, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.002303123474121, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8724006414413452, "num_tokens": 745610735.0, "step": 19633 }, { "epoch": 2.4976466098460754, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.1903645992279053, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8674442172050476, "num_tokens": 745653513.0, "step": 19634 }, { "epoch": 2.497773820124666, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2505574226379395, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8750482201576233, "num_tokens": 745690007.0, "step": 19635 }, { "epoch": 2.4979010304032565, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3556571006774902, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8686438202857971, "num_tokens": 745724886.0, "step": 19636 }, { "epoch": 2.498028240681847, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.403080463409424, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8849320411682129, "num_tokens": 745760782.0, "step": 19637 }, { "epoch": 2.4981554509604376, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6302151679992676, "learning_rate": 1e-05, "loss": 0.3971, "mean_token_accuracy": 0.8921998143196106, "num_tokens": 745794535.0, "step": 19638 }, { "epoch": 2.498282661239028, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.327976703643799, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8790304064750671, "num_tokens": 745831010.0, "step": 19639 }, { "epoch": 2.4984098715176186, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.801452875137329, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8720473051071167, "num_tokens": 745866761.0, "step": 19640 }, { "epoch": 2.498537081796209, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.108887195587158, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8810996413230896, "num_tokens": 745909081.0, "step": 19641 }, { "epoch": 2.4986642920747997, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7691311836242676, "learning_rate": 1e-05, "loss": 0.3986, "mean_token_accuracy": 0.8938276171684265, "num_tokens": 745942907.0, "step": 19642 }, { "epoch": 2.49879150235339, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9214236736297607, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8759564161300659, "num_tokens": 745981558.0, "step": 19643 }, { "epoch": 2.4989187126319807, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.701434373855591, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8792309165000916, "num_tokens": 746015568.0, "step": 19644 }, { "epoch": 2.4990459229105713, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7902002334594727, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.884969174861908, "num_tokens": 746049503.0, "step": 19645 }, { "epoch": 2.499173133189162, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.59484601020813, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8793315291404724, "num_tokens": 746086979.0, "step": 19646 }, { "epoch": 2.4993003434677523, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7169201374053955, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8809458017349243, "num_tokens": 746125073.0, "step": 19647 }, { "epoch": 2.499427553746343, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5038228034973145, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8814713954925537, "num_tokens": 746161621.0, "step": 19648 }, { "epoch": 2.4995547640249334, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6898560523986816, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8744829893112183, "num_tokens": 746203032.0, "step": 19649 }, { "epoch": 2.499681974303524, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4361021518707275, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8773412704467773, "num_tokens": 746233939.0, "step": 19650 }, { "epoch": 2.499809184582114, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7193655967712402, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.888621985912323, "num_tokens": 746269175.0, "step": 19651 }, { "epoch": 2.499936394860705, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.343409538269043, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8743240833282471, "num_tokens": 746310634.0, "step": 19652 }, { "epoch": 2.500063605139295, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.79927396774292, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8834291696548462, "num_tokens": 746353683.0, "step": 19653 }, { "epoch": 2.500190815417886, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.274960517883301, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8765147924423218, "num_tokens": 746392299.0, "step": 19654 }, { "epoch": 2.500318025696476, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.948059320449829, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.882085382938385, "num_tokens": 746435276.0, "step": 19655 }, { "epoch": 2.5004452359750666, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.266603708267212, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8717977404594421, "num_tokens": 746476175.0, "step": 19656 }, { "epoch": 2.500572446253657, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.125103950500488, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8887289762496948, "num_tokens": 746517084.0, "step": 19657 }, { "epoch": 2.5006996565322477, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.376781463623047, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8638942837715149, "num_tokens": 746556073.0, "step": 19658 }, { "epoch": 2.500826866810838, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.222474575042725, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8906203508377075, "num_tokens": 746590188.0, "step": 19659 }, { "epoch": 2.5009540770894287, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5706372261047363, "learning_rate": 1e-05, "loss": 0.5373, "mean_token_accuracy": 0.8511502742767334, "num_tokens": 746634667.0, "step": 19660 }, { "epoch": 2.5010812873680193, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.116804599761963, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8854237794876099, "num_tokens": 746677398.0, "step": 19661 }, { "epoch": 2.50120849764661, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6804325580596924, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8847649693489075, "num_tokens": 746711607.0, "step": 19662 }, { "epoch": 2.5013357079252003, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9229578971862793, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8863215446472168, "num_tokens": 746751853.0, "step": 19663 }, { "epoch": 2.501462918203791, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.6914820671081543, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8651740550994873, "num_tokens": 746792134.0, "step": 19664 }, { "epoch": 2.5015901284823814, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.730588674545288, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8700898289680481, "num_tokens": 746830087.0, "step": 19665 }, { "epoch": 2.501717338760972, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6644937992095947, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8843823671340942, "num_tokens": 746864738.0, "step": 19666 }, { "epoch": 2.5018445490395624, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.556260585784912, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8622924089431763, "num_tokens": 746904021.0, "step": 19667 }, { "epoch": 2.501971759318153, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.635382890701294, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8839138150215149, "num_tokens": 746942710.0, "step": 19668 }, { "epoch": 2.5020989695967435, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.442394733428955, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8799038529396057, "num_tokens": 746986694.0, "step": 19669 }, { "epoch": 2.502226179875334, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6401784420013428, "learning_rate": 1e-05, "loss": 0.3964, "mean_token_accuracy": 0.8906930088996887, "num_tokens": 747028440.0, "step": 19670 }, { "epoch": 2.5023533901539246, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3809685707092285, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.8871050477027893, "num_tokens": 747063862.0, "step": 19671 }, { "epoch": 2.502480600432515, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6703248023986816, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8761334419250488, "num_tokens": 747100469.0, "step": 19672 }, { "epoch": 2.5026078107111056, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3196218013763428, "learning_rate": 1e-05, "loss": 0.3821, "mean_token_accuracy": 0.8907356262207031, "num_tokens": 747133074.0, "step": 19673 }, { "epoch": 2.5027350209896957, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.746558904647827, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8666913509368896, "num_tokens": 747167010.0, "step": 19674 }, { "epoch": 2.5028622312682867, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.262953042984009, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8740347027778625, "num_tokens": 747201397.0, "step": 19675 }, { "epoch": 2.5029894415468767, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8594093322753906, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8865841031074524, "num_tokens": 747238430.0, "step": 19676 }, { "epoch": 2.5031166518254677, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2239623069763184, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8786400556564331, "num_tokens": 747277760.0, "step": 19677 }, { "epoch": 2.503243862104058, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.028008937835693, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8783086538314819, "num_tokens": 747317763.0, "step": 19678 }, { "epoch": 2.5033710723826488, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2442004680633545, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8691443800926208, "num_tokens": 747362733.0, "step": 19679 }, { "epoch": 2.503498282661239, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.20907735824585, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8875042200088501, "num_tokens": 747400735.0, "step": 19680 }, { "epoch": 2.5036254929398294, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.3930015563964844, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8752255439758301, "num_tokens": 747440263.0, "step": 19681 }, { "epoch": 2.50375270321842, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2972331047058105, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8809581995010376, "num_tokens": 747477918.0, "step": 19682 }, { "epoch": 2.5038799134970104, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.600863218307495, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8777816891670227, "num_tokens": 747513235.0, "step": 19683 }, { "epoch": 2.504007123775601, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2038044929504395, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8962414264678955, "num_tokens": 747553280.0, "step": 19684 }, { "epoch": 2.5041343340541915, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7520642280578613, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8734726905822754, "num_tokens": 747594389.0, "step": 19685 }, { "epoch": 2.504261544332782, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.987736225128174, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8828490376472473, "num_tokens": 747634805.0, "step": 19686 }, { "epoch": 2.5043887546113726, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.7102203369140625, "learning_rate": 1e-05, "loss": 0.479, "mean_token_accuracy": 0.869746744632721, "num_tokens": 747673963.0, "step": 19687 }, { "epoch": 2.504515964889963, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000331878662109375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.812298059463501, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8710488080978394, "num_tokens": 747715067.0, "step": 19688 }, { "epoch": 2.5046431751685536, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6377828121185303, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8846548199653625, "num_tokens": 747752799.0, "step": 19689 }, { "epoch": 2.504770385447144, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6927785873413086, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8700927495956421, "num_tokens": 747789325.0, "step": 19690 }, { "epoch": 2.5048975957257347, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5730903148651123, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.877817690372467, "num_tokens": 747823438.0, "step": 19691 }, { "epoch": 2.505024806004325, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.603113889694214, "learning_rate": 1e-05, "loss": 0.4063, "mean_token_accuracy": 0.8867531418800354, "num_tokens": 747862310.0, "step": 19692 }, { "epoch": 2.5051520162829157, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5050058364868164, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8800593018531799, "num_tokens": 747903154.0, "step": 19693 }, { "epoch": 2.5052792265615063, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5912532806396484, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8765792846679688, "num_tokens": 747937849.0, "step": 19694 }, { "epoch": 2.505406436840097, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4472577571868896, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8782541751861572, "num_tokens": 747972387.0, "step": 19695 }, { "epoch": 2.5055336471186873, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.604712724685669, "learning_rate": 1e-05, "loss": 0.4869, "mean_token_accuracy": 0.8638354539871216, "num_tokens": 748007181.0, "step": 19696 }, { "epoch": 2.505660857397278, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.380200147628784, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8846743106842041, "num_tokens": 748044342.0, "step": 19697 }, { "epoch": 2.5057880676758684, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6670844554901123, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8724947571754456, "num_tokens": 748074012.0, "step": 19698 }, { "epoch": 2.5059152779544585, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.301530599594116, "learning_rate": 1e-05, "loss": 0.3962, "mean_token_accuracy": 0.8884861469268799, "num_tokens": 748111055.0, "step": 19699 }, { "epoch": 2.5060424882330494, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7684414386749268, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8603309392929077, "num_tokens": 748151922.0, "step": 19700 }, { "epoch": 2.5061696985116395, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.238816499710083, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8884126543998718, "num_tokens": 748186090.0, "step": 19701 }, { "epoch": 2.5062969087902305, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9333560466766357, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8853496313095093, "num_tokens": 748223281.0, "step": 19702 }, { "epoch": 2.5064241190688206, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003337860107421875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.224752187728882, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8813151717185974, "num_tokens": 748255657.0, "step": 19703 }, { "epoch": 2.5065513293474115, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.136790752410889, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8688373565673828, "num_tokens": 748296954.0, "step": 19704 }, { "epoch": 2.5066785396260016, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3101606369018555, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8731627464294434, "num_tokens": 748334753.0, "step": 19705 }, { "epoch": 2.506805749904592, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.287232398986816, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8812985420227051, "num_tokens": 748377880.0, "step": 19706 }, { "epoch": 2.5069329601831827, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.546980857849121, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.87401282787323, "num_tokens": 748418230.0, "step": 19707 }, { "epoch": 2.507060170461773, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.241506576538086, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8832181692123413, "num_tokens": 748459320.0, "step": 19708 }, { "epoch": 2.5071873807403637, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.687375783920288, "learning_rate": 1e-05, "loss": 0.488, "mean_token_accuracy": 0.8617120981216431, "num_tokens": 748503951.0, "step": 19709 }, { "epoch": 2.5073145910189543, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.072601318359375, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8791635036468506, "num_tokens": 748544736.0, "step": 19710 }, { "epoch": 2.507441801297545, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7253572940826416, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8803012371063232, "num_tokens": 748583174.0, "step": 19711 }, { "epoch": 2.5075690115761353, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.884263515472412, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8804895877838135, "num_tokens": 748624580.0, "step": 19712 }, { "epoch": 2.507696221854726, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6791343688964844, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8643743991851807, "num_tokens": 748660343.0, "step": 19713 }, { "epoch": 2.5078234321333164, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.736189842224121, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8669818639755249, "num_tokens": 748695006.0, "step": 19714 }, { "epoch": 2.507950642411907, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.5993947982788086, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8835759162902832, "num_tokens": 748732142.0, "step": 19715 }, { "epoch": 2.5080778526904974, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.63578462600708, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8692566156387329, "num_tokens": 748771263.0, "step": 19716 }, { "epoch": 2.508205062969088, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5314273834228516, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8743919134140015, "num_tokens": 748812796.0, "step": 19717 }, { "epoch": 2.5083322732476785, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.585487127304077, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.876666247844696, "num_tokens": 748852701.0, "step": 19718 }, { "epoch": 2.508459483526269, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4837377071380615, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8698123097419739, "num_tokens": 748888799.0, "step": 19719 }, { "epoch": 2.5085866938048595, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5761303901672363, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8800133466720581, "num_tokens": 748927056.0, "step": 19720 }, { "epoch": 2.50871390408345, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4355058670043945, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.8642644882202148, "num_tokens": 748961829.0, "step": 19721 }, { "epoch": 2.5088411143620406, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5816380977630615, "learning_rate": 1e-05, "loss": 0.4042, "mean_token_accuracy": 0.8879578113555908, "num_tokens": 748995281.0, "step": 19722 }, { "epoch": 2.508968324640631, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3781380653381348, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8848586082458496, "num_tokens": 749027574.0, "step": 19723 }, { "epoch": 2.509095534919221, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6367006301879883, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8828574419021606, "num_tokens": 749065092.0, "step": 19724 }, { "epoch": 2.509222745197812, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.309598207473755, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.870807409286499, "num_tokens": 749104389.0, "step": 19725 }, { "epoch": 2.5093499554764023, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.722670555114746, "learning_rate": 1e-05, "loss": 0.3865, "mean_token_accuracy": 0.8919812440872192, "num_tokens": 749138433.0, "step": 19726 }, { "epoch": 2.5094771657549932, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2338168621063232, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8850688934326172, "num_tokens": 749178095.0, "step": 19727 }, { "epoch": 2.5096043760335833, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8809282779693604, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8945528864860535, "num_tokens": 749215547.0, "step": 19728 }, { "epoch": 2.5097315863121743, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1919686794281006, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8648963570594788, "num_tokens": 749254935.0, "step": 19729 }, { "epoch": 2.5098587965907644, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.109049320220947, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8896982669830322, "num_tokens": 749290425.0, "step": 19730 }, { "epoch": 2.509986006869355, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2621326446533203, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8827714920043945, "num_tokens": 749319390.0, "step": 19731 }, { "epoch": 2.5101132171479454, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.303829669952393, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.8658626675605774, "num_tokens": 749358517.0, "step": 19732 }, { "epoch": 2.510240427426536, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.4599883556365967, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8782532811164856, "num_tokens": 749402219.0, "step": 19733 }, { "epoch": 2.5103676377051265, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.3744683265686035, "learning_rate": 1e-05, "loss": 0.5261, "mean_token_accuracy": 0.851667582988739, "num_tokens": 749437514.0, "step": 19734 }, { "epoch": 2.510494847983717, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.724374532699585, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8814867734909058, "num_tokens": 749477148.0, "step": 19735 }, { "epoch": 2.5106220582623076, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.195063591003418, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8810834884643555, "num_tokens": 749511174.0, "step": 19736 }, { "epoch": 2.510749268540898, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7873129844665527, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8850514888763428, "num_tokens": 749547117.0, "step": 19737 }, { "epoch": 2.5108764788194886, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.9871952533721924, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8868454098701477, "num_tokens": 749583755.0, "step": 19738 }, { "epoch": 2.511003689098079, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.721972942352295, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8823188543319702, "num_tokens": 749621742.0, "step": 19739 }, { "epoch": 2.5111308993766697, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8136675357818604, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.8907725811004639, "num_tokens": 749659178.0, "step": 19740 }, { "epoch": 2.51125810965526, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6472651958465576, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8705891966819763, "num_tokens": 749698828.0, "step": 19741 }, { "epoch": 2.5113853199338507, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7069435119628906, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8644903898239136, "num_tokens": 749732609.0, "step": 19742 }, { "epoch": 2.5115125302124413, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.559701681137085, "learning_rate": 1e-05, "loss": 0.3773, "mean_token_accuracy": 0.8969943523406982, "num_tokens": 749768240.0, "step": 19743 }, { "epoch": 2.511639740491032, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6478404998779297, "learning_rate": 1e-05, "loss": 0.4045, "mean_token_accuracy": 0.8860937356948853, "num_tokens": 749809435.0, "step": 19744 }, { "epoch": 2.5117669507696223, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4670166969299316, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8681427836418152, "num_tokens": 749852790.0, "step": 19745 }, { "epoch": 2.511894161048213, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6601078510284424, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8753564357757568, "num_tokens": 749895111.0, "step": 19746 }, { "epoch": 2.5120213713268034, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.389251470565796, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8869458436965942, "num_tokens": 749936142.0, "step": 19747 }, { "epoch": 2.512148581605394, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7086598873138428, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8872566223144531, "num_tokens": 749972106.0, "step": 19748 }, { "epoch": 2.512275791883984, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.314049243927002, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8669611811637878, "num_tokens": 750014090.0, "step": 19749 }, { "epoch": 2.512403002162575, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8195643424987793, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8863446116447449, "num_tokens": 750052542.0, "step": 19750 }, { "epoch": 2.512530212441165, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.256974935531616, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8640660047531128, "num_tokens": 750091846.0, "step": 19751 }, { "epoch": 2.512657422719756, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.983689785003662, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8632016181945801, "num_tokens": 750135721.0, "step": 19752 }, { "epoch": 2.512784632998346, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2601332664489746, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8678029775619507, "num_tokens": 750177786.0, "step": 19753 }, { "epoch": 2.5129118432769366, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.1743645668029785, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8755776286125183, "num_tokens": 750209577.0, "step": 19754 }, { "epoch": 2.513039053555527, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.389378547668457, "learning_rate": 1e-05, "loss": 0.3873, "mean_token_accuracy": 0.894311785697937, "num_tokens": 750243168.0, "step": 19755 }, { "epoch": 2.5131662638341177, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.237376689910889, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8744481801986694, "num_tokens": 750288304.0, "step": 19756 }, { "epoch": 2.513293474112708, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5862371921539307, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.875419020652771, "num_tokens": 750329662.0, "step": 19757 }, { "epoch": 2.5134206843912987, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.150177478790283, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8720799088478088, "num_tokens": 750364581.0, "step": 19758 }, { "epoch": 2.5135478946698893, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.705554723739624, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8824231624603271, "num_tokens": 750403129.0, "step": 19759 }, { "epoch": 2.51367510494848, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9420688152313232, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8806767463684082, "num_tokens": 750439738.0, "step": 19760 }, { "epoch": 2.5138023152270703, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6967105865478516, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8635762929916382, "num_tokens": 750479302.0, "step": 19761 }, { "epoch": 2.513929525505661, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.746565818786621, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8804681301116943, "num_tokens": 750517658.0, "step": 19762 }, { "epoch": 2.5140567357842514, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.64090633392334, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8918941617012024, "num_tokens": 750558192.0, "step": 19763 }, { "epoch": 2.514183946062842, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.606149911880493, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.881958544254303, "num_tokens": 750595209.0, "step": 19764 }, { "epoch": 2.5143111563414324, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6211488246917725, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.872689962387085, "num_tokens": 750634560.0, "step": 19765 }, { "epoch": 2.514438366620023, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5005791187286377, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8819550275802612, "num_tokens": 750675223.0, "step": 19766 }, { "epoch": 2.5145655768986135, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5990052223205566, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8775337934494019, "num_tokens": 750717302.0, "step": 19767 }, { "epoch": 2.514692787177204, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4378747940063477, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8758800029754639, "num_tokens": 750754615.0, "step": 19768 }, { "epoch": 2.5148199974557945, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5948989391326904, "learning_rate": 1e-05, "loss": 0.3891, "mean_token_accuracy": 0.8949467539787292, "num_tokens": 750788745.0, "step": 19769 }, { "epoch": 2.514947207734385, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3917746543884277, "learning_rate": 1e-05, "loss": 0.4734, "mean_token_accuracy": 0.863020658493042, "num_tokens": 750829727.0, "step": 19770 }, { "epoch": 2.5150744180129756, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6158666610717773, "learning_rate": 1e-05, "loss": 0.4808, "mean_token_accuracy": 0.8650012016296387, "num_tokens": 750873095.0, "step": 19771 }, { "epoch": 2.5152016282915657, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.347970724105835, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8673443794250488, "num_tokens": 750909513.0, "step": 19772 }, { "epoch": 2.5153288385701567, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6689493656158447, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8779176473617554, "num_tokens": 750947234.0, "step": 19773 }, { "epoch": 2.5154560488487467, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2686526775360107, "learning_rate": 1e-05, "loss": 0.3808, "mean_token_accuracy": 0.8902034759521484, "num_tokens": 750984926.0, "step": 19774 }, { "epoch": 2.5155832591273377, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7857677936553955, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8793334364891052, "num_tokens": 751023700.0, "step": 19775 }, { "epoch": 2.515710469405928, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1944119930267334, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8815304040908813, "num_tokens": 751065444.0, "step": 19776 }, { "epoch": 2.5158376796845188, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9929463863372803, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8630515933036804, "num_tokens": 751101627.0, "step": 19777 }, { "epoch": 2.515964889963109, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.1840128898620605, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8853529095649719, "num_tokens": 751138006.0, "step": 19778 }, { "epoch": 2.5160921002416994, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.232427597045898, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8902427554130554, "num_tokens": 751179187.0, "step": 19779 }, { "epoch": 2.51621931052029, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3250489234924316, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8824176788330078, "num_tokens": 751217086.0, "step": 19780 }, { "epoch": 2.5163465207988804, "ewc_loss": 0.10498046875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 4.417379856109619, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8750593066215515, "num_tokens": 751259330.0, "step": 19781 }, { "epoch": 2.516473731077471, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.615774631500244, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8795862197875977, "num_tokens": 751300873.0, "step": 19782 }, { "epoch": 2.5166009413560615, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.370221138000488, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8862165212631226, "num_tokens": 751330783.0, "step": 19783 }, { "epoch": 2.516728151634652, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.8239715099334717, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8827845454216003, "num_tokens": 751370032.0, "step": 19784 }, { "epoch": 2.5168553619132426, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.141457557678223, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.883036732673645, "num_tokens": 751412561.0, "step": 19785 }, { "epoch": 2.516982572191833, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.8255298137664795, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8777554035186768, "num_tokens": 751442168.0, "step": 19786 }, { "epoch": 2.5171097824704236, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.933988332748413, "learning_rate": 1e-05, "loss": 0.5033, "mean_token_accuracy": 0.8574224710464478, "num_tokens": 751488656.0, "step": 19787 }, { "epoch": 2.517236992749014, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7104225158691406, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8760449886322021, "num_tokens": 751533305.0, "step": 19788 }, { "epoch": 2.5173642030276047, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7908647060394287, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8779484033584595, "num_tokens": 751570693.0, "step": 19789 }, { "epoch": 2.517491413306195, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.619375467300415, "learning_rate": 1e-05, "loss": 0.3956, "mean_token_accuracy": 0.893625020980835, "num_tokens": 751604994.0, "step": 19790 }, { "epoch": 2.5176186235847857, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7150819301605225, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8777363300323486, "num_tokens": 751642552.0, "step": 19791 }, { "epoch": 2.5177458338633762, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5318689346313477, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.886239767074585, "num_tokens": 751677253.0, "step": 19792 }, { "epoch": 2.5178730441419668, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.682368040084839, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8839730024337769, "num_tokens": 751717737.0, "step": 19793 }, { "epoch": 2.5180002544205573, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.450542688369751, "learning_rate": 1e-05, "loss": 0.3724, "mean_token_accuracy": 0.8984197974205017, "num_tokens": 751750453.0, "step": 19794 }, { "epoch": 2.518127464699148, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.681293487548828, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8781747817993164, "num_tokens": 751792179.0, "step": 19795 }, { "epoch": 2.5182546749777384, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3586246967315674, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8811891078948975, "num_tokens": 751833367.0, "step": 19796 }, { "epoch": 2.5183818852563284, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.766122579574585, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.882498562335968, "num_tokens": 751874042.0, "step": 19797 }, { "epoch": 2.5185090955349194, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2960567474365234, "learning_rate": 1e-05, "loss": 0.4979, "mean_token_accuracy": 0.8551621437072754, "num_tokens": 751914990.0, "step": 19798 }, { "epoch": 2.5186363058135095, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.912550687789917, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8641607761383057, "num_tokens": 751946266.0, "step": 19799 }, { "epoch": 2.5187635160921005, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2497398853302, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.8894680738449097, "num_tokens": 751990359.0, "step": 19800 }, { "epoch": 2.5188907263706906, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1022047996521, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.876282274723053, "num_tokens": 752026955.0, "step": 19801 }, { "epoch": 2.5190179366492815, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.33048415184021, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8686416149139404, "num_tokens": 752065187.0, "step": 19802 }, { "epoch": 2.5191451469278716, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.241168975830078, "learning_rate": 1e-05, "loss": 0.3962, "mean_token_accuracy": 0.8925217390060425, "num_tokens": 752100855.0, "step": 19803 }, { "epoch": 2.519272357206462, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.527113437652588, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8819102048873901, "num_tokens": 752136672.0, "step": 19804 }, { "epoch": 2.5193995674850527, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.217187404632568, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8851228952407837, "num_tokens": 752167529.0, "step": 19805 }, { "epoch": 2.519526777763643, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.699007749557495, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8652114868164062, "num_tokens": 752201865.0, "step": 19806 }, { "epoch": 2.5196539880422337, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.022143840789795, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8807277679443359, "num_tokens": 752239315.0, "step": 19807 }, { "epoch": 2.5197811983208243, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7159037590026855, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8862894773483276, "num_tokens": 752278505.0, "step": 19808 }, { "epoch": 2.519908408599415, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.805981397628784, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8866711854934692, "num_tokens": 752319346.0, "step": 19809 }, { "epoch": 2.5200356188780053, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6793413162231445, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8683339357376099, "num_tokens": 752364210.0, "step": 19810 }, { "epoch": 2.520162829156596, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.637552499771118, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.8913618922233582, "num_tokens": 752400679.0, "step": 19811 }, { "epoch": 2.5202900394351864, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6508030891418457, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8795203566551208, "num_tokens": 752436003.0, "step": 19812 }, { "epoch": 2.520417249713777, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5125598907470703, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8695520758628845, "num_tokens": 752478690.0, "step": 19813 }, { "epoch": 2.5205444599923674, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6420798301696777, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8642265200614929, "num_tokens": 752513326.0, "step": 19814 }, { "epoch": 2.520671670270958, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4296202659606934, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.870363712310791, "num_tokens": 752556553.0, "step": 19815 }, { "epoch": 2.5207988805495485, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6217222213745117, "learning_rate": 1e-05, "loss": 0.3981, "mean_token_accuracy": 0.8877688646316528, "num_tokens": 752599219.0, "step": 19816 }, { "epoch": 2.520926090828139, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.36444354057312, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8839684128761292, "num_tokens": 752639308.0, "step": 19817 }, { "epoch": 2.5210533011067295, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6688666343688965, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8630192279815674, "num_tokens": 752679384.0, "step": 19818 }, { "epoch": 2.52118051138532, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.304633617401123, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8725482821464539, "num_tokens": 752722395.0, "step": 19819 }, { "epoch": 2.5213077216639106, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.770813465118408, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8776803612709045, "num_tokens": 752754491.0, "step": 19820 }, { "epoch": 2.521434931942501, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2390050888061523, "learning_rate": 1e-05, "loss": 0.3981, "mean_token_accuracy": 0.8872110843658447, "num_tokens": 752792595.0, "step": 19821 }, { "epoch": 2.521562142221091, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9201221466064453, "learning_rate": 1e-05, "loss": 0.4522, "mean_token_accuracy": 0.8757637739181519, "num_tokens": 752830689.0, "step": 19822 }, { "epoch": 2.521689352499682, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2024686336517334, "learning_rate": 1e-05, "loss": 0.3777, "mean_token_accuracy": 0.8926656246185303, "num_tokens": 752862848.0, "step": 19823 }, { "epoch": 2.5218165627782723, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.126636505126953, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8819930553436279, "num_tokens": 752900566.0, "step": 19824 }, { "epoch": 2.5219437730568632, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.271083116531372, "learning_rate": 1e-05, "loss": 0.4042, "mean_token_accuracy": 0.8878740072250366, "num_tokens": 752939201.0, "step": 19825 }, { "epoch": 2.5220709833354533, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.3292975425720215, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.886489987373352, "num_tokens": 752973820.0, "step": 19826 }, { "epoch": 2.522198193614044, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.575535774230957, "learning_rate": 1e-05, "loss": 0.5064, "mean_token_accuracy": 0.858311653137207, "num_tokens": 753018563.0, "step": 19827 }, { "epoch": 2.5223254038926344, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2793073654174805, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8722783327102661, "num_tokens": 753060097.0, "step": 19828 }, { "epoch": 2.522452614171225, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.685351848602295, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.872333288192749, "num_tokens": 753099200.0, "step": 19829 }, { "epoch": 2.5225798244498154, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.147790431976318, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8816802501678467, "num_tokens": 753132421.0, "step": 19830 }, { "epoch": 2.522707034728406, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7395334243774414, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8685303330421448, "num_tokens": 753165380.0, "step": 19831 }, { "epoch": 2.5228342450069965, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9544856548309326, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8803436756134033, "num_tokens": 753203501.0, "step": 19832 }, { "epoch": 2.522961455285587, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.678131341934204, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.868484616279602, "num_tokens": 753247674.0, "step": 19833 }, { "epoch": 2.5230886655641775, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8151180744171143, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.881731390953064, "num_tokens": 753285967.0, "step": 19834 }, { "epoch": 2.523215875842768, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.619310140609741, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8743885159492493, "num_tokens": 753319080.0, "step": 19835 }, { "epoch": 2.5233430861213586, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.716514825820923, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8697538375854492, "num_tokens": 753357924.0, "step": 19836 }, { "epoch": 2.523470296399949, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.536287546157837, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8845975399017334, "num_tokens": 753397922.0, "step": 19837 }, { "epoch": 2.5235975066785397, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6854100227355957, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.8557555079460144, "num_tokens": 753431268.0, "step": 19838 }, { "epoch": 2.52372471695713, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.471679210662842, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8719779253005981, "num_tokens": 753465169.0, "step": 19839 }, { "epoch": 2.5238519272357207, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6705541610717773, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8804727792739868, "num_tokens": 753501017.0, "step": 19840 }, { "epoch": 2.5239791375143112, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.387298107147217, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.8819305300712585, "num_tokens": 753541039.0, "step": 19841 }, { "epoch": 2.5241063477929018, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7347795963287354, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.85101318359375, "num_tokens": 753585793.0, "step": 19842 }, { "epoch": 2.5242335580714923, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3191261291503906, "learning_rate": 1e-05, "loss": 0.385, "mean_token_accuracy": 0.8905740976333618, "num_tokens": 753623752.0, "step": 19843 }, { "epoch": 2.524360768350083, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.841169595718384, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8721632957458496, "num_tokens": 753662939.0, "step": 19844 }, { "epoch": 2.5244879786286734, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2841625213623047, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8834997415542603, "num_tokens": 753699200.0, "step": 19845 }, { "epoch": 2.524615188907264, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9752016067504883, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8690789937973022, "num_tokens": 753743867.0, "step": 19846 }, { "epoch": 2.524742399185854, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.305663824081421, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8661574125289917, "num_tokens": 753781494.0, "step": 19847 }, { "epoch": 2.524869609464445, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.102025032043457, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8800186514854431, "num_tokens": 753823518.0, "step": 19848 }, { "epoch": 2.524996819743035, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.4264400005340576, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8756236433982849, "num_tokens": 753855058.0, "step": 19849 }, { "epoch": 2.525124030021626, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.121584415435791, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8826298117637634, "num_tokens": 753896721.0, "step": 19850 }, { "epoch": 2.525251240300216, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.5253467559814453, "learning_rate": 1e-05, "loss": 0.397, "mean_token_accuracy": 0.8922731876373291, "num_tokens": 753933225.0, "step": 19851 }, { "epoch": 2.5253784505788066, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.050187110900879, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8722580671310425, "num_tokens": 753970373.0, "step": 19852 }, { "epoch": 2.525505660857397, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6108930110931396, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.8873195052146912, "num_tokens": 754007544.0, "step": 19853 }, { "epoch": 2.5256328711359877, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.885178804397583, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8844645023345947, "num_tokens": 754040858.0, "step": 19854 }, { "epoch": 2.525760081414578, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.630125045776367, "learning_rate": 1e-05, "loss": 0.503, "mean_token_accuracy": 0.8607089519500732, "num_tokens": 754083877.0, "step": 19855 }, { "epoch": 2.5258872916931687, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7021214962005615, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8906323313713074, "num_tokens": 754122792.0, "step": 19856 }, { "epoch": 2.5260145019717593, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.630078077316284, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8792985677719116, "num_tokens": 754157986.0, "step": 19857 }, { "epoch": 2.52614171225035, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.554558277130127, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8780034780502319, "num_tokens": 754198321.0, "step": 19858 }, { "epoch": 2.5262689225289403, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.628811836242676, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8622052669525146, "num_tokens": 754235463.0, "step": 19859 }, { "epoch": 2.526396132807531, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4402575492858887, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.878510594367981, "num_tokens": 754274319.0, "step": 19860 }, { "epoch": 2.5265233430861214, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6229445934295654, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8809511661529541, "num_tokens": 754315206.0, "step": 19861 }, { "epoch": 2.526650553364712, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3761847019195557, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8744770288467407, "num_tokens": 754353300.0, "step": 19862 }, { "epoch": 2.5267777636433024, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.660684823989868, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8693585395812988, "num_tokens": 754392692.0, "step": 19863 }, { "epoch": 2.526904973921893, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.315823793411255, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8773826956748962, "num_tokens": 754428688.0, "step": 19864 }, { "epoch": 2.5270321842004835, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7339088916778564, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8786077499389648, "num_tokens": 754467647.0, "step": 19865 }, { "epoch": 2.527159394479074, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2640788555145264, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8660379648208618, "num_tokens": 754501511.0, "step": 19866 }, { "epoch": 2.5272866047576645, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8353703022003174, "learning_rate": 1e-05, "loss": 0.3943, "mean_token_accuracy": 0.8906725645065308, "num_tokens": 754543000.0, "step": 19867 }, { "epoch": 2.527413815036255, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.213348627090454, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8797867894172668, "num_tokens": 754583485.0, "step": 19868 }, { "epoch": 2.5275410253148456, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.017306327819824, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8823643922805786, "num_tokens": 754620472.0, "step": 19869 }, { "epoch": 2.5276682355934357, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.234050989151001, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8822621703147888, "num_tokens": 754659906.0, "step": 19870 }, { "epoch": 2.5277954458720266, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.219479084014893, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8788554668426514, "num_tokens": 754696128.0, "step": 19871 }, { "epoch": 2.5279226561506167, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.38411021232605, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8793407678604126, "num_tokens": 754729553.0, "step": 19872 }, { "epoch": 2.5280498664292077, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.317692756652832, "learning_rate": 1e-05, "loss": 0.494, "mean_token_accuracy": 0.8641000390052795, "num_tokens": 754773689.0, "step": 19873 }, { "epoch": 2.528177076707798, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6109795570373535, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8744543790817261, "num_tokens": 754809326.0, "step": 19874 }, { "epoch": 2.5283042869863888, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.228994846343994, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.878066897392273, "num_tokens": 754850627.0, "step": 19875 }, { "epoch": 2.528431497264979, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.729093313217163, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8839526176452637, "num_tokens": 754887810.0, "step": 19876 }, { "epoch": 2.5285587075435694, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.051179885864258, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8766946196556091, "num_tokens": 754923778.0, "step": 19877 }, { "epoch": 2.52868591782216, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7215375900268555, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8804889917373657, "num_tokens": 754963486.0, "step": 19878 }, { "epoch": 2.5288131281007504, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8608505725860596, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8904657363891602, "num_tokens": 755000302.0, "step": 19879 }, { "epoch": 2.528940338379341, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6515538692474365, "learning_rate": 1e-05, "loss": 0.3912, "mean_token_accuracy": 0.8927590847015381, "num_tokens": 755040319.0, "step": 19880 }, { "epoch": 2.5290675486579315, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7416934967041016, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8847179412841797, "num_tokens": 755073410.0, "step": 19881 }, { "epoch": 2.529194758936522, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.577620029449463, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8829079270362854, "num_tokens": 755112279.0, "step": 19882 }, { "epoch": 2.5293219692151125, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6707141399383545, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8611806631088257, "num_tokens": 755146904.0, "step": 19883 }, { "epoch": 2.529449179493703, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5206782817840576, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.87105393409729, "num_tokens": 755179881.0, "step": 19884 }, { "epoch": 2.5295763897722936, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6294479370117188, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8905032873153687, "num_tokens": 755214504.0, "step": 19885 }, { "epoch": 2.529703600050884, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.442023277282715, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8772725462913513, "num_tokens": 755254229.0, "step": 19886 }, { "epoch": 2.5298308103294747, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6406784057617188, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.872871994972229, "num_tokens": 755296150.0, "step": 19887 }, { "epoch": 2.529958020608065, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.362912893295288, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8846864104270935, "num_tokens": 755336508.0, "step": 19888 }, { "epoch": 2.5300852308866557, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.701566696166992, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8671883344650269, "num_tokens": 755376863.0, "step": 19889 }, { "epoch": 2.5302124411652462, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.30159330368042, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8866519927978516, "num_tokens": 755414209.0, "step": 19890 }, { "epoch": 2.5303396514438368, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8111486434936523, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.862160861492157, "num_tokens": 755454281.0, "step": 19891 }, { "epoch": 2.5304668617224273, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2436535358428955, "learning_rate": 1e-05, "loss": 0.3719, "mean_token_accuracy": 0.8964220881462097, "num_tokens": 755492441.0, "step": 19892 }, { "epoch": 2.530594072001018, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.984621286392212, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8779042959213257, "num_tokens": 755526264.0, "step": 19893 }, { "epoch": 2.5307212822796084, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2676103115081787, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8831735253334045, "num_tokens": 755560819.0, "step": 19894 }, { "epoch": 2.5308484925581984, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.16377067565918, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8775213360786438, "num_tokens": 755592103.0, "step": 19895 }, { "epoch": 2.5309757028367894, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.3796985149383545, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8884732723236084, "num_tokens": 755631864.0, "step": 19896 }, { "epoch": 2.5311029131153795, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.236146450042725, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8722299337387085, "num_tokens": 755672555.0, "step": 19897 }, { "epoch": 2.5312301233939705, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.5813422203063965, "learning_rate": 1e-05, "loss": 0.4021, "mean_token_accuracy": 0.8903631567955017, "num_tokens": 755705606.0, "step": 19898 }, { "epoch": 2.5313573336725606, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.154133319854736, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8736708760261536, "num_tokens": 755738844.0, "step": 19899 }, { "epoch": 2.5314845439511515, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.686570644378662, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8872666954994202, "num_tokens": 755774164.0, "step": 19900 }, { "epoch": 2.5316117542297416, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.945194721221924, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.87042236328125, "num_tokens": 755816848.0, "step": 19901 }, { "epoch": 2.531738964508332, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6881208419799805, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8745484948158264, "num_tokens": 755854281.0, "step": 19902 }, { "epoch": 2.5318661747869227, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.754244804382324, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8637986183166504, "num_tokens": 755893769.0, "step": 19903 }, { "epoch": 2.531993385065513, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.659991502761841, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8755042552947998, "num_tokens": 755931690.0, "step": 19904 }, { "epoch": 2.5321205953441037, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6173229217529297, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8804123997688293, "num_tokens": 755967177.0, "step": 19905 }, { "epoch": 2.5322478056226942, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6091713905334473, "learning_rate": 1e-05, "loss": 0.3987, "mean_token_accuracy": 0.8877887725830078, "num_tokens": 755998440.0, "step": 19906 }, { "epoch": 2.5323750159012848, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5176589488983154, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8755805492401123, "num_tokens": 756038049.0, "step": 19907 }, { "epoch": 2.5325022261798753, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.582587480545044, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.8893734812736511, "num_tokens": 756073185.0, "step": 19908 }, { "epoch": 2.532629436458466, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4508919715881348, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8711838722229004, "num_tokens": 756113439.0, "step": 19909 }, { "epoch": 2.5327566467370564, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.563248872756958, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8814451694488525, "num_tokens": 756155516.0, "step": 19910 }, { "epoch": 2.532883857015647, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4084906578063965, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8792626857757568, "num_tokens": 756196685.0, "step": 19911 }, { "epoch": 2.5330110672942374, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5796332359313965, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8742300271987915, "num_tokens": 756230150.0, "step": 19912 }, { "epoch": 2.533138277572828, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3565118312835693, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.8892503380775452, "num_tokens": 756270806.0, "step": 19913 }, { "epoch": 2.5332654878514185, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6381402015686035, "learning_rate": 1e-05, "loss": 0.3906, "mean_token_accuracy": 0.8907288312911987, "num_tokens": 756300387.0, "step": 19914 }, { "epoch": 2.533392698130009, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.303884744644165, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8818191289901733, "num_tokens": 756335436.0, "step": 19915 }, { "epoch": 2.5335199084085995, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7274599075317383, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8671865463256836, "num_tokens": 756371942.0, "step": 19916 }, { "epoch": 2.53364711868719, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.225560426712036, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8807214498519897, "num_tokens": 756416102.0, "step": 19917 }, { "epoch": 2.5337743289657806, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8867664337158203, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8657483458518982, "num_tokens": 756457066.0, "step": 19918 }, { "epoch": 2.533901539244371, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.168405055999756, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8795506954193115, "num_tokens": 756493977.0, "step": 19919 }, { "epoch": 2.534028749522961, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.142749786376953, "learning_rate": 1e-05, "loss": 0.4971, "mean_token_accuracy": 0.8588727116584778, "num_tokens": 756535926.0, "step": 19920 }, { "epoch": 2.534155959801552, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.213961601257324, "learning_rate": 1e-05, "loss": 0.3974, "mean_token_accuracy": 0.892513632774353, "num_tokens": 756578406.0, "step": 19921 }, { "epoch": 2.5342831700801423, "ewc_loss": 0.10498046875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.422636032104492, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8612188100814819, "num_tokens": 756615104.0, "step": 19922 }, { "epoch": 2.5344103803587332, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5244405269622803, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8712332248687744, "num_tokens": 756652059.0, "step": 19923 }, { "epoch": 2.5345375906373233, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.4516682624816895, "learning_rate": 1e-05, "loss": 0.3974, "mean_token_accuracy": 0.8969545364379883, "num_tokens": 756688640.0, "step": 19924 }, { "epoch": 2.534664800915914, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.8147759437561035, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.890282392501831, "num_tokens": 756721596.0, "step": 19925 }, { "epoch": 2.5347920111945044, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.251308917999268, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8667250871658325, "num_tokens": 756757663.0, "step": 19926 }, { "epoch": 2.534919221473095, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.845823287963867, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.880328357219696, "num_tokens": 756802021.0, "step": 19927 }, { "epoch": 2.5350464317516854, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0300774574279785, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8701503276824951, "num_tokens": 756842039.0, "step": 19928 }, { "epoch": 2.535173642030276, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7761898040771484, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8797048926353455, "num_tokens": 756875788.0, "step": 19929 }, { "epoch": 2.5353008523088665, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.85654878616333, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8868169784545898, "num_tokens": 756913292.0, "step": 19930 }, { "epoch": 2.535428062587457, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6648032665252686, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8952140212059021, "num_tokens": 756953005.0, "step": 19931 }, { "epoch": 2.5355552728660475, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7532806396484375, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8799585700035095, "num_tokens": 756987656.0, "step": 19932 }, { "epoch": 2.535682483144638, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5641541481018066, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8855720162391663, "num_tokens": 757021040.0, "step": 19933 }, { "epoch": 2.5358096934232286, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7098193168640137, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8918300867080688, "num_tokens": 757053039.0, "step": 19934 }, { "epoch": 2.535936903701819, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4688658714294434, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8728612661361694, "num_tokens": 757090057.0, "step": 19935 }, { "epoch": 2.5360641139804097, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.700045585632324, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8826926350593567, "num_tokens": 757129537.0, "step": 19936 }, { "epoch": 2.536191324259, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3899059295654297, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8885639309883118, "num_tokens": 757164892.0, "step": 19937 }, { "epoch": 2.5363185345375907, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.757201671600342, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8717817068099976, "num_tokens": 757204717.0, "step": 19938 }, { "epoch": 2.5364457448161812, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3221778869628906, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8761300444602966, "num_tokens": 757239725.0, "step": 19939 }, { "epoch": 2.5365729550947718, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8696489334106445, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.86131751537323, "num_tokens": 757280356.0, "step": 19940 }, { "epoch": 2.5367001653733623, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2719569206237793, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.873045802116394, "num_tokens": 757316498.0, "step": 19941 }, { "epoch": 2.536827375651953, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0428266525268555, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8735235333442688, "num_tokens": 757351675.0, "step": 19942 }, { "epoch": 2.5369545859305433, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.305189609527588, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8749985694885254, "num_tokens": 757393310.0, "step": 19943 }, { "epoch": 2.537081796209134, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.193807601928711, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8699449896812439, "num_tokens": 757429800.0, "step": 19944 }, { "epoch": 2.537209006487724, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.4774932861328125, "learning_rate": 1e-05, "loss": 0.3927, "mean_token_accuracy": 0.8928862810134888, "num_tokens": 757463456.0, "step": 19945 }, { "epoch": 2.537336216766315, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.200961589813232, "learning_rate": 1e-05, "loss": 0.4766, "mean_token_accuracy": 0.8655930757522583, "num_tokens": 757496977.0, "step": 19946 }, { "epoch": 2.537463427044905, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6334228515625, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8841309547424316, "num_tokens": 757536003.0, "step": 19947 }, { "epoch": 2.537590637323496, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.0537943840026855, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8705556988716125, "num_tokens": 757575017.0, "step": 19948 }, { "epoch": 2.537717847602086, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.696027994155884, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8782694339752197, "num_tokens": 757617098.0, "step": 19949 }, { "epoch": 2.5378450578806766, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8550333976745605, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8680430054664612, "num_tokens": 757652343.0, "step": 19950 }, { "epoch": 2.537972268159267, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.7676801681518555, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8796319961547852, "num_tokens": 757691021.0, "step": 19951 }, { "epoch": 2.5380994784378577, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6067752838134766, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8842337131500244, "num_tokens": 757727805.0, "step": 19952 }, { "epoch": 2.538226688716448, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6872644424438477, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8717445135116577, "num_tokens": 757767979.0, "step": 19953 }, { "epoch": 2.5383538989950387, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4668898582458496, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.880486786365509, "num_tokens": 757807488.0, "step": 19954 }, { "epoch": 2.5384811092736292, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7276110649108887, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8692096471786499, "num_tokens": 757838298.0, "step": 19955 }, { "epoch": 2.5386083195522198, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.36863112449646, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8749421834945679, "num_tokens": 757875016.0, "step": 19956 }, { "epoch": 2.5387355298308103, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.740511178970337, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8830651640892029, "num_tokens": 757921306.0, "step": 19957 }, { "epoch": 2.538862740109401, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.294581413269043, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8824261426925659, "num_tokens": 757956555.0, "step": 19958 }, { "epoch": 2.5389899503879914, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8584725856781006, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8702190518379211, "num_tokens": 757990433.0, "step": 19959 }, { "epoch": 2.539117160666582, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.245926856994629, "learning_rate": 1e-05, "loss": 0.3986, "mean_token_accuracy": 0.890232264995575, "num_tokens": 758026606.0, "step": 19960 }, { "epoch": 2.5392443709451724, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.006470203399658, "learning_rate": 1e-05, "loss": 0.3972, "mean_token_accuracy": 0.8919649124145508, "num_tokens": 758060995.0, "step": 19961 }, { "epoch": 2.539371581223763, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.267169237136841, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8844168782234192, "num_tokens": 758097479.0, "step": 19962 }, { "epoch": 2.5394987915023535, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.184421539306641, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8789787888526917, "num_tokens": 758134541.0, "step": 19963 }, { "epoch": 2.539626001780944, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.390368938446045, "learning_rate": 1e-05, "loss": 0.4009, "mean_token_accuracy": 0.8884388208389282, "num_tokens": 758173885.0, "step": 19964 }, { "epoch": 2.5397532120595345, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.252198696136475, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8693311810493469, "num_tokens": 758219394.0, "step": 19965 }, { "epoch": 2.539880422338125, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.5783231258392334, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8760450482368469, "num_tokens": 758261146.0, "step": 19966 }, { "epoch": 2.5400076326167156, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.174640655517578, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8966643810272217, "num_tokens": 758297935.0, "step": 19967 }, { "epoch": 2.5401348428953057, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6988747119903564, "learning_rate": 1e-05, "loss": 0.4641, "mean_token_accuracy": 0.871947169303894, "num_tokens": 758328579.0, "step": 19968 }, { "epoch": 2.5402620531738966, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.006340026855469, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8751768469810486, "num_tokens": 758361731.0, "step": 19969 }, { "epoch": 2.5403892634524867, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6908013820648193, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8829676508903503, "num_tokens": 758394651.0, "step": 19970 }, { "epoch": 2.5405164737310777, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8194727897644043, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8792930841445923, "num_tokens": 758436455.0, "step": 19971 }, { "epoch": 2.540643684009668, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6344799995422363, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8831734657287598, "num_tokens": 758471880.0, "step": 19972 }, { "epoch": 2.5407708942882588, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6889588832855225, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8804186582565308, "num_tokens": 758514354.0, "step": 19973 }, { "epoch": 2.540898104566849, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5662333965301514, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8852157592773438, "num_tokens": 758554263.0, "step": 19974 }, { "epoch": 2.5410253148454394, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6259841918945312, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8774616122245789, "num_tokens": 758588907.0, "step": 19975 }, { "epoch": 2.54115252512403, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.53239107131958, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.872400164604187, "num_tokens": 758618573.0, "step": 19976 }, { "epoch": 2.5412797354026204, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5892446041107178, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8802178502082825, "num_tokens": 758654315.0, "step": 19977 }, { "epoch": 2.541406945681211, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.476891279220581, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.873308539390564, "num_tokens": 758690670.0, "step": 19978 }, { "epoch": 2.5415341559598015, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5673859119415283, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8802070617675781, "num_tokens": 758725686.0, "step": 19979 }, { "epoch": 2.541661366238392, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.433372735977173, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8828076720237732, "num_tokens": 758757002.0, "step": 19980 }, { "epoch": 2.5417885765169825, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5754077434539795, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.884152889251709, "num_tokens": 758793695.0, "step": 19981 }, { "epoch": 2.541915786795573, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.400529623031616, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8759804964065552, "num_tokens": 758837278.0, "step": 19982 }, { "epoch": 2.5420429970741636, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.614854574203491, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8757425546646118, "num_tokens": 758871252.0, "step": 19983 }, { "epoch": 2.542170207352754, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.327751398086548, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8790872097015381, "num_tokens": 758912250.0, "step": 19984 }, { "epoch": 2.5422974176313446, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.673980474472046, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8787142038345337, "num_tokens": 758951773.0, "step": 19985 }, { "epoch": 2.542424627909935, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2693932056427, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8877663612365723, "num_tokens": 758989054.0, "step": 19986 }, { "epoch": 2.5425518381885257, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8009259700775146, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.884342610836029, "num_tokens": 759026917.0, "step": 19987 }, { "epoch": 2.5426790484671162, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.20993709564209, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8778637647628784, "num_tokens": 759065752.0, "step": 19988 }, { "epoch": 2.5428062587457068, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.016005516052246, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8870348930358887, "num_tokens": 759100650.0, "step": 19989 }, { "epoch": 2.5429334690242973, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.210460662841797, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.867495059967041, "num_tokens": 759138274.0, "step": 19990 }, { "epoch": 2.543060679302888, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.242189884185791, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.86778724193573, "num_tokens": 759176060.0, "step": 19991 }, { "epoch": 2.5431878895814783, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.3667354583740234, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8864255547523499, "num_tokens": 759208323.0, "step": 19992 }, { "epoch": 2.5433150998600684, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.3821821212768555, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8915470838546753, "num_tokens": 759239180.0, "step": 19993 }, { "epoch": 2.5434423101386594, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6248350143432617, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8826276063919067, "num_tokens": 759278489.0, "step": 19994 }, { "epoch": 2.5435695204172495, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.289663791656494, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.878105878829956, "num_tokens": 759315254.0, "step": 19995 }, { "epoch": 2.5436967306958405, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.7944138050079346, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8821296691894531, "num_tokens": 759345969.0, "step": 19996 }, { "epoch": 2.5438239409744305, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.071009159088135, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8748407363891602, "num_tokens": 759386530.0, "step": 19997 }, { "epoch": 2.5439511512530215, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.755101442337036, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8742205500602722, "num_tokens": 759428755.0, "step": 19998 }, { "epoch": 2.5440783615316116, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.881627321243286, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8777449131011963, "num_tokens": 759471335.0, "step": 19999 }, { "epoch": 2.544205571810202, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6743600368499756, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8790096044540405, "num_tokens": 759512541.0, "step": 20000 }, { "epoch": 2.5443327820887927, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7491793632507324, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.87447589635849, "num_tokens": 759551015.0, "step": 20001 }, { "epoch": 2.544459992367383, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6054491996765137, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8699009418487549, "num_tokens": 759588698.0, "step": 20002 }, { "epoch": 2.5445872026459737, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6684370040893555, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8623775839805603, "num_tokens": 759630180.0, "step": 20003 }, { "epoch": 2.5447144129245642, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.536041498184204, "learning_rate": 1e-05, "loss": 0.495, "mean_token_accuracy": 0.8596092462539673, "num_tokens": 759664416.0, "step": 20004 }, { "epoch": 2.5448416232031548, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6141598224639893, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8686997890472412, "num_tokens": 759708868.0, "step": 20005 }, { "epoch": 2.5449688334817453, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.461681604385376, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8813150525093079, "num_tokens": 759742431.0, "step": 20006 }, { "epoch": 2.545096043760336, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.633232593536377, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8823434710502625, "num_tokens": 759776504.0, "step": 20007 }, { "epoch": 2.5452232540389264, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3846805095672607, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.8918032646179199, "num_tokens": 759808246.0, "step": 20008 }, { "epoch": 2.545350464317517, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.679168939590454, "learning_rate": 1e-05, "loss": 0.3912, "mean_token_accuracy": 0.8932463526725769, "num_tokens": 759846654.0, "step": 20009 }, { "epoch": 2.5454776745961074, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.315023183822632, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8756556510925293, "num_tokens": 759882594.0, "step": 20010 }, { "epoch": 2.545604884874698, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7735843658447266, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8840571641921997, "num_tokens": 759923557.0, "step": 20011 }, { "epoch": 2.5457320951532885, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2528884410858154, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8721559047698975, "num_tokens": 759961152.0, "step": 20012 }, { "epoch": 2.545859305431879, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9303500652313232, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8808519244194031, "num_tokens": 759995505.0, "step": 20013 }, { "epoch": 2.5459865157104695, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.230602979660034, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8776847124099731, "num_tokens": 760033910.0, "step": 20014 }, { "epoch": 2.54611372598906, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.128349304199219, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.892230749130249, "num_tokens": 760067605.0, "step": 20015 }, { "epoch": 2.5462409362676506, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.333488702774048, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8639346361160278, "num_tokens": 760101151.0, "step": 20016 }, { "epoch": 2.546368146546241, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.271543025970459, "learning_rate": 1e-05, "loss": 0.4822, "mean_token_accuracy": 0.8692400455474854, "num_tokens": 760140253.0, "step": 20017 }, { "epoch": 2.546495356824831, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5247209072113037, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8786724805831909, "num_tokens": 760175988.0, "step": 20018 }, { "epoch": 2.546622567103422, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2338738441467285, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.874674916267395, "num_tokens": 760211734.0, "step": 20019 }, { "epoch": 2.5467497773820122, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.685467004776001, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8740503787994385, "num_tokens": 760248887.0, "step": 20020 }, { "epoch": 2.546876987660603, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.0618062019348145, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8766622543334961, "num_tokens": 760284836.0, "step": 20021 }, { "epoch": 2.5470041979391933, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7186694145202637, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8764461278915405, "num_tokens": 760327328.0, "step": 20022 }, { "epoch": 2.547131408217784, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.871910572052002, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8795216083526611, "num_tokens": 760362876.0, "step": 20023 }, { "epoch": 2.5472586184963744, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6740851402282715, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8843079805374146, "num_tokens": 760399652.0, "step": 20024 }, { "epoch": 2.547385828774965, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.713128089904785, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8850563764572144, "num_tokens": 760436346.0, "step": 20025 }, { "epoch": 2.5475130390535554, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.618588924407959, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8773388266563416, "num_tokens": 760468540.0, "step": 20026 }, { "epoch": 2.547640249332146, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.608429193496704, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8807108402252197, "num_tokens": 760504088.0, "step": 20027 }, { "epoch": 2.5477674596107365, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5578675270080566, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8774750232696533, "num_tokens": 760542543.0, "step": 20028 }, { "epoch": 2.547894669889327, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5408904552459717, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8687433004379272, "num_tokens": 760582472.0, "step": 20029 }, { "epoch": 2.5480218801679175, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.52286696434021, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8720530271530151, "num_tokens": 760618410.0, "step": 20030 }, { "epoch": 2.548149090446508, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5043270587921143, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.884685754776001, "num_tokens": 760656223.0, "step": 20031 }, { "epoch": 2.5482763007250986, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4697530269622803, "learning_rate": 1e-05, "loss": 0.3852, "mean_token_accuracy": 0.8941190242767334, "num_tokens": 760700898.0, "step": 20032 }, { "epoch": 2.548403511003689, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.506699800491333, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8782540559768677, "num_tokens": 760739608.0, "step": 20033 }, { "epoch": 2.5485307212822796, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4570841789245605, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8673568964004517, "num_tokens": 760777808.0, "step": 20034 }, { "epoch": 2.54865793156087, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.50231671333313, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8826059103012085, "num_tokens": 760817497.0, "step": 20035 }, { "epoch": 2.5487851418394607, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.425625801086426, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8746601343154907, "num_tokens": 760855699.0, "step": 20036 }, { "epoch": 2.5489123521180512, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.53725004196167, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8611412644386292, "num_tokens": 760890796.0, "step": 20037 }, { "epoch": 2.5490395623966418, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3751180171966553, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8795720338821411, "num_tokens": 760929056.0, "step": 20038 }, { "epoch": 2.5491667726752323, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5764241218566895, "learning_rate": 1e-05, "loss": 0.3932, "mean_token_accuracy": 0.8902119398117065, "num_tokens": 760963941.0, "step": 20039 }, { "epoch": 2.549293982953823, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.315000534057617, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8777916431427002, "num_tokens": 761009703.0, "step": 20040 }, { "epoch": 2.5494211932324133, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.655519723892212, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8818557858467102, "num_tokens": 761051408.0, "step": 20041 }, { "epoch": 2.549548403511004, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.249887228012085, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.870562732219696, "num_tokens": 761093764.0, "step": 20042 }, { "epoch": 2.549675613789594, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000335693359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8110146522521973, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8801510334014893, "num_tokens": 761129975.0, "step": 20043 }, { "epoch": 2.549802824068185, "ewc_loss": 0.080078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.6253204345703125e-05, "grad_norm": 3.1680006980895996, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8713943958282471, "num_tokens": 761169845.0, "step": 20044 }, { "epoch": 2.549930034346775, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.050354480743408, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8762909173965454, "num_tokens": 761207667.0, "step": 20045 }, { "epoch": 2.550057244625366, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.1599740982055664, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8803342580795288, "num_tokens": 761249649.0, "step": 20046 }, { "epoch": 2.550184454903956, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.367822647094727, "learning_rate": 1e-05, "loss": 0.5475, "mean_token_accuracy": 0.8477156162261963, "num_tokens": 761289879.0, "step": 20047 }, { "epoch": 2.5503116651825466, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.3681552410125732, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8718695044517517, "num_tokens": 761331398.0, "step": 20048 }, { "epoch": 2.550438875461137, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.573258876800537, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8806437253952026, "num_tokens": 761363928.0, "step": 20049 }, { "epoch": 2.5505660857397277, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.781048059463501, "learning_rate": 1e-05, "loss": 0.5071, "mean_token_accuracy": 0.8624398708343506, "num_tokens": 761405737.0, "step": 20050 }, { "epoch": 2.550693296018318, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.435125350952148, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8867999315261841, "num_tokens": 761435934.0, "step": 20051 }, { "epoch": 2.5508205062969087, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.960272789001465, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8820568323135376, "num_tokens": 761466700.0, "step": 20052 }, { "epoch": 2.5509477165754992, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.163929462432861, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8793900012969971, "num_tokens": 761505682.0, "step": 20053 }, { "epoch": 2.5510749268540898, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.865114212036133, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8772344589233398, "num_tokens": 761544071.0, "step": 20054 }, { "epoch": 2.5512021371326803, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.972658634185791, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8862767219543457, "num_tokens": 761586595.0, "step": 20055 }, { "epoch": 2.551329347411271, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7297940254211426, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8792856931686401, "num_tokens": 761623924.0, "step": 20056 }, { "epoch": 2.5514565576898613, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.861401081085205, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8748435974121094, "num_tokens": 761660952.0, "step": 20057 }, { "epoch": 2.551583767968452, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.595238447189331, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8895398378372192, "num_tokens": 761701451.0, "step": 20058 }, { "epoch": 2.5517109782470424, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.830777883529663, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8798543214797974, "num_tokens": 761733638.0, "step": 20059 }, { "epoch": 2.551838188525633, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4924755096435547, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8805210590362549, "num_tokens": 761763978.0, "step": 20060 }, { "epoch": 2.5519653988042235, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8473541736602783, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.885823130607605, "num_tokens": 761802653.0, "step": 20061 }, { "epoch": 2.552092609082814, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.3885741233825684, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8808534741401672, "num_tokens": 761840671.0, "step": 20062 }, { "epoch": 2.5522198193614045, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9474854469299316, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8726311922073364, "num_tokens": 761883823.0, "step": 20063 }, { "epoch": 2.552347029639995, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.353349208831787, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8794016242027283, "num_tokens": 761921597.0, "step": 20064 }, { "epoch": 2.5524742399185856, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.074008464813232, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8863475918769836, "num_tokens": 761963400.0, "step": 20065 }, { "epoch": 2.5526014501971757, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3941633701324463, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8772697448730469, "num_tokens": 762004442.0, "step": 20066 }, { "epoch": 2.5527286604757666, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.1992573738098145, "learning_rate": 1e-05, "loss": 0.4855, "mean_token_accuracy": 0.8662567138671875, "num_tokens": 762040679.0, "step": 20067 }, { "epoch": 2.5528558707543567, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.532756805419922, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8763101100921631, "num_tokens": 762085249.0, "step": 20068 }, { "epoch": 2.5529830810329477, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.151820659637451, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8935818672180176, "num_tokens": 762125551.0, "step": 20069 }, { "epoch": 2.5531102913115378, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6734187602996826, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8829222321510315, "num_tokens": 762162503.0, "step": 20070 }, { "epoch": 2.5532375015901287, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9836418628692627, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.883732795715332, "num_tokens": 762197302.0, "step": 20071 }, { "epoch": 2.553364711868719, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.711331844329834, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8844040632247925, "num_tokens": 762235069.0, "step": 20072 }, { "epoch": 2.5534919221473094, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7536168098449707, "learning_rate": 1e-05, "loss": 0.3895, "mean_token_accuracy": 0.8911535143852234, "num_tokens": 762272014.0, "step": 20073 }, { "epoch": 2.5536191324259, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.693943500518799, "learning_rate": 1e-05, "loss": 0.3926, "mean_token_accuracy": 0.8916682004928589, "num_tokens": 762312840.0, "step": 20074 }, { "epoch": 2.5537463427044904, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5780160427093506, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8773432374000549, "num_tokens": 762347997.0, "step": 20075 }, { "epoch": 2.553873552983081, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.68034029006958, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8731687068939209, "num_tokens": 762390706.0, "step": 20076 }, { "epoch": 2.5540007632616715, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.448943853378296, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.877575159072876, "num_tokens": 762425894.0, "step": 20077 }, { "epoch": 2.554127973540262, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6952691078186035, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8641754388809204, "num_tokens": 762461962.0, "step": 20078 }, { "epoch": 2.5542551838188525, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.3360507488250732, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.876375675201416, "num_tokens": 762500128.0, "step": 20079 }, { "epoch": 2.554382394097443, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.763000249862671, "learning_rate": 1e-05, "loss": 0.3835, "mean_token_accuracy": 0.8962396383285522, "num_tokens": 762536256.0, "step": 20080 }, { "epoch": 2.5545096043760336, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2832443714141846, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8573740720748901, "num_tokens": 762569797.0, "step": 20081 }, { "epoch": 2.554636814654624, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.874990940093994, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8760941028594971, "num_tokens": 762606849.0, "step": 20082 }, { "epoch": 2.5547640249332146, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.226816177368164, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8695496320724487, "num_tokens": 762649526.0, "step": 20083 }, { "epoch": 2.554891235211805, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.049830436706543, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8811587691307068, "num_tokens": 762686240.0, "step": 20084 }, { "epoch": 2.5550184454903957, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2677395343780518, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8685548901557922, "num_tokens": 762727908.0, "step": 20085 }, { "epoch": 2.5551456557689862, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.227177619934082, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8917673230171204, "num_tokens": 762765966.0, "step": 20086 }, { "epoch": 2.5552728660475768, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.4377851486206055, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8790527582168579, "num_tokens": 762806031.0, "step": 20087 }, { "epoch": 2.5554000763261673, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.27384090423584, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8878031969070435, "num_tokens": 762847531.0, "step": 20088 }, { "epoch": 2.555527286604758, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.643190860748291, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8812978267669678, "num_tokens": 762878409.0, "step": 20089 }, { "epoch": 2.5556544968833483, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.148820400238037, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8789017200469971, "num_tokens": 762918964.0, "step": 20090 }, { "epoch": 2.5557817071619384, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7170252799987793, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8871027827262878, "num_tokens": 762958962.0, "step": 20091 }, { "epoch": 2.5559089174405294, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.970034122467041, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8657518029212952, "num_tokens": 762996062.0, "step": 20092 }, { "epoch": 2.5560361277191195, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.7053565979003906, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8804913759231567, "num_tokens": 763033387.0, "step": 20093 }, { "epoch": 2.5561633379977104, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.803010940551758, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8746912479400635, "num_tokens": 763067826.0, "step": 20094 }, { "epoch": 2.5562905482763005, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.641288995742798, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8860450983047485, "num_tokens": 763100442.0, "step": 20095 }, { "epoch": 2.5564177585548915, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6597423553466797, "learning_rate": 1e-05, "loss": 0.3919, "mean_token_accuracy": 0.8905504941940308, "num_tokens": 763137659.0, "step": 20096 }, { "epoch": 2.5565449688334816, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.572702646255493, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8882452249526978, "num_tokens": 763180000.0, "step": 20097 }, { "epoch": 2.556672179112072, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.597852945327759, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.867684543132782, "num_tokens": 763218863.0, "step": 20098 }, { "epoch": 2.5567993893906626, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.514787435531616, "learning_rate": 1e-05, "loss": 0.3957, "mean_token_accuracy": 0.8894674181938171, "num_tokens": 763252973.0, "step": 20099 }, { "epoch": 2.556926599669253, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5902793407440186, "learning_rate": 1e-05, "loss": 0.4965, "mean_token_accuracy": 0.8616493940353394, "num_tokens": 763288781.0, "step": 20100 }, { "epoch": 2.5570538099478437, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.465444326400757, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8734332323074341, "num_tokens": 763321839.0, "step": 20101 }, { "epoch": 2.5571810202264342, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5788347721099854, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8747928738594055, "num_tokens": 763357607.0, "step": 20102 }, { "epoch": 2.5573082305050248, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4039509296417236, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8738725185394287, "num_tokens": 763399489.0, "step": 20103 }, { "epoch": 2.5574354407836153, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6179957389831543, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8635742664337158, "num_tokens": 763441256.0, "step": 20104 }, { "epoch": 2.557562651062206, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3630809783935547, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.869687557220459, "num_tokens": 763475019.0, "step": 20105 }, { "epoch": 2.5576898613407963, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6810903549194336, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8775167465209961, "num_tokens": 763510456.0, "step": 20106 }, { "epoch": 2.557817071619387, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2888343334198, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8736443519592285, "num_tokens": 763546181.0, "step": 20107 }, { "epoch": 2.5579442818979774, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8026068210601807, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8828059434890747, "num_tokens": 763578763.0, "step": 20108 }, { "epoch": 2.558071492176568, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.224915027618408, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8668538331985474, "num_tokens": 763615615.0, "step": 20109 }, { "epoch": 2.5581987024551585, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.969136953353882, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8895204663276672, "num_tokens": 763655546.0, "step": 20110 }, { "epoch": 2.558325912733749, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2055344581604004, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8725402355194092, "num_tokens": 763695400.0, "step": 20111 }, { "epoch": 2.5584531230123395, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.214473724365234, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8601959943771362, "num_tokens": 763725968.0, "step": 20112 }, { "epoch": 2.55858033329093, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3393352031707764, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8804015517234802, "num_tokens": 763763708.0, "step": 20113 }, { "epoch": 2.5587075435695206, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.361236095428467, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8665388822555542, "num_tokens": 763800836.0, "step": 20114 }, { "epoch": 2.558834753848111, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.604586601257324, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8714488744735718, "num_tokens": 763842119.0, "step": 20115 }, { "epoch": 2.558961964126701, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.294592380523682, "learning_rate": 1e-05, "loss": 0.4468, "mean_token_accuracy": 0.8774552345275879, "num_tokens": 763876797.0, "step": 20116 }, { "epoch": 2.559089174405292, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.780867576599121, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8742910027503967, "num_tokens": 763913131.0, "step": 20117 }, { "epoch": 2.5592163846838822, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.083102226257324, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8795903921127319, "num_tokens": 763952430.0, "step": 20118 }, { "epoch": 2.559343594962473, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7644565105438232, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8793852925300598, "num_tokens": 763988978.0, "step": 20119 }, { "epoch": 2.5594708052410633, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8781259059906006, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.876152753829956, "num_tokens": 764027473.0, "step": 20120 }, { "epoch": 2.559598015519654, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6924407482147217, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8835026621818542, "num_tokens": 764063681.0, "step": 20121 }, { "epoch": 2.5597252257982444, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7185637950897217, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8826242089271545, "num_tokens": 764102463.0, "step": 20122 }, { "epoch": 2.559852436076835, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.614553928375244, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8877894878387451, "num_tokens": 764141912.0, "step": 20123 }, { "epoch": 2.5599796463554254, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6305136680603027, "learning_rate": 1e-05, "loss": 0.3986, "mean_token_accuracy": 0.8901351690292358, "num_tokens": 764176448.0, "step": 20124 }, { "epoch": 2.560106856634016, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.538332462310791, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8624744415283203, "num_tokens": 764217287.0, "step": 20125 }, { "epoch": 2.5602340669126065, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5989065170288086, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8765122890472412, "num_tokens": 764255593.0, "step": 20126 }, { "epoch": 2.560361277191197, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4756312370300293, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8750196695327759, "num_tokens": 764294624.0, "step": 20127 }, { "epoch": 2.5604884874697875, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.599414110183716, "learning_rate": 1e-05, "loss": 0.3994, "mean_token_accuracy": 0.8886532187461853, "num_tokens": 764328495.0, "step": 20128 }, { "epoch": 2.560615697748378, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.415227174758911, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.887631893157959, "num_tokens": 764359986.0, "step": 20129 }, { "epoch": 2.5607429080269686, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.631403684616089, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8730784058570862, "num_tokens": 764394039.0, "step": 20130 }, { "epoch": 2.560870118305559, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3437395095825195, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8701493740081787, "num_tokens": 764436106.0, "step": 20131 }, { "epoch": 2.5609973285841496, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7111730575561523, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.881846010684967, "num_tokens": 764470004.0, "step": 20132 }, { "epoch": 2.56112453886274, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2603447437286377, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8861512541770935, "num_tokens": 764507679.0, "step": 20133 }, { "epoch": 2.5612517491413307, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.851893424987793, "learning_rate": 1e-05, "loss": 0.4045, "mean_token_accuracy": 0.8923691511154175, "num_tokens": 764548818.0, "step": 20134 }, { "epoch": 2.561378959419921, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.209412097930908, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8741871118545532, "num_tokens": 764590289.0, "step": 20135 }, { "epoch": 2.5615061696985117, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.06786584854126, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8838011026382446, "num_tokens": 764624210.0, "step": 20136 }, { "epoch": 2.5616333799771023, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2568516731262207, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8762138485908508, "num_tokens": 764657071.0, "step": 20137 }, { "epoch": 2.561760590255693, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.279170036315918, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8798165321350098, "num_tokens": 764691439.0, "step": 20138 }, { "epoch": 2.5618878005342833, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.43015456199646, "learning_rate": 1e-05, "loss": 0.3865, "mean_token_accuracy": 0.8946155309677124, "num_tokens": 764731551.0, "step": 20139 }, { "epoch": 2.562015010812874, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.372793674468994, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8943780064582825, "num_tokens": 764763233.0, "step": 20140 }, { "epoch": 2.562142221091464, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6977717876434326, "learning_rate": 1e-05, "loss": 0.3791, "mean_token_accuracy": 0.8987802267074585, "num_tokens": 764796937.0, "step": 20141 }, { "epoch": 2.562269431370055, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.222955703735352, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8825168609619141, "num_tokens": 764831808.0, "step": 20142 }, { "epoch": 2.562396641648645, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.8086822032928467, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8799123764038086, "num_tokens": 764863288.0, "step": 20143 }, { "epoch": 2.562523851927236, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.984422206878662, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8897933959960938, "num_tokens": 764905417.0, "step": 20144 }, { "epoch": 2.562651062205826, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7496049404144287, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8803983926773071, "num_tokens": 764942058.0, "step": 20145 }, { "epoch": 2.5627782724844166, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.803147554397583, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8764369487762451, "num_tokens": 764983652.0, "step": 20146 }, { "epoch": 2.562905482763007, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6659996509552, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8825811147689819, "num_tokens": 765023966.0, "step": 20147 }, { "epoch": 2.5630326930415976, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.692338705062866, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8742277026176453, "num_tokens": 765055847.0, "step": 20148 }, { "epoch": 2.563159903320188, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.595623016357422, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8711773157119751, "num_tokens": 765094387.0, "step": 20149 }, { "epoch": 2.5632871135987787, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.620326280593872, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8700714111328125, "num_tokens": 765128193.0, "step": 20150 }, { "epoch": 2.5634143238773692, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5132434368133545, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8774621486663818, "num_tokens": 765166550.0, "step": 20151 }, { "epoch": 2.5635415341559598, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6010475158691406, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8763735890388489, "num_tokens": 765204228.0, "step": 20152 }, { "epoch": 2.5636687444345503, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4269187450408936, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8827148675918579, "num_tokens": 765247148.0, "step": 20153 }, { "epoch": 2.563795954713141, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.625941276550293, "learning_rate": 1e-05, "loss": 0.3796, "mean_token_accuracy": 0.8916508555412292, "num_tokens": 765287902.0, "step": 20154 }, { "epoch": 2.5639231649917313, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.352104425430298, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8846628665924072, "num_tokens": 765331118.0, "step": 20155 }, { "epoch": 2.564050375270322, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.701523542404175, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8880027532577515, "num_tokens": 765369942.0, "step": 20156 }, { "epoch": 2.5641775855489124, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2836103439331055, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8794147968292236, "num_tokens": 765407664.0, "step": 20157 }, { "epoch": 2.564304795827503, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.828272819519043, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8776171207427979, "num_tokens": 765453303.0, "step": 20158 }, { "epoch": 2.5644320061060935, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.222282648086548, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8796044588088989, "num_tokens": 765492558.0, "step": 20159 }, { "epoch": 2.564559216384684, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.038017749786377, "learning_rate": 1e-05, "loss": 0.5067, "mean_token_accuracy": 0.8551979660987854, "num_tokens": 765529964.0, "step": 20160 }, { "epoch": 2.5646864266632745, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.224492073059082, "learning_rate": 1e-05, "loss": 0.393, "mean_token_accuracy": 0.8921385407447815, "num_tokens": 765570158.0, "step": 20161 }, { "epoch": 2.564813636941865, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.276782035827637, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.8574620485305786, "num_tokens": 765609304.0, "step": 20162 }, { "epoch": 2.5649408472204556, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.4105820655822754, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8776892423629761, "num_tokens": 765651124.0, "step": 20163 }, { "epoch": 2.5650680574990457, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.378849506378174, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8912616968154907, "num_tokens": 765685544.0, "step": 20164 }, { "epoch": 2.5651952677776366, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.683368682861328, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8908054828643799, "num_tokens": 765726120.0, "step": 20165 }, { "epoch": 2.5653224780562267, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.249368667602539, "learning_rate": 1e-05, "loss": 0.4813, "mean_token_accuracy": 0.8660914301872253, "num_tokens": 765764145.0, "step": 20166 }, { "epoch": 2.5654496883348177, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.8116726875305176, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.8907638788223267, "num_tokens": 765795530.0, "step": 20167 }, { "epoch": 2.5655768986134078, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.0162177085876465, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.874947190284729, "num_tokens": 765831793.0, "step": 20168 }, { "epoch": 2.5657041088919987, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.757124900817871, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8724859952926636, "num_tokens": 765876724.0, "step": 20169 }, { "epoch": 2.565831319170589, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8161237239837646, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.891759991645813, "num_tokens": 765917604.0, "step": 20170 }, { "epoch": 2.5659585294491793, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6883790493011475, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8809213638305664, "num_tokens": 765953717.0, "step": 20171 }, { "epoch": 2.56608573972777, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.682852268218994, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8753861784934998, "num_tokens": 765989967.0, "step": 20172 }, { "epoch": 2.5662129500063604, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.600916862487793, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8833805918693542, "num_tokens": 766029276.0, "step": 20173 }, { "epoch": 2.566340160284951, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6015748977661133, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8862712383270264, "num_tokens": 766064819.0, "step": 20174 }, { "epoch": 2.5664673705635415, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.535161256790161, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8737114667892456, "num_tokens": 766103621.0, "step": 20175 }, { "epoch": 2.566594580842132, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5642685890197754, "learning_rate": 1e-05, "loss": 0.3966, "mean_token_accuracy": 0.8876374363899231, "num_tokens": 766138945.0, "step": 20176 }, { "epoch": 2.5667217911207225, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4818480014801025, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8742190599441528, "num_tokens": 766180771.0, "step": 20177 }, { "epoch": 2.566849001399313, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.565816879272461, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8701972961425781, "num_tokens": 766216907.0, "step": 20178 }, { "epoch": 2.5669762116779036, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4254634380340576, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8746740221977234, "num_tokens": 766256302.0, "step": 20179 }, { "epoch": 2.567103421956494, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.594057559967041, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8808116912841797, "num_tokens": 766291255.0, "step": 20180 }, { "epoch": 2.5672306322350846, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.365648031234741, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.8844401836395264, "num_tokens": 766323895.0, "step": 20181 }, { "epoch": 2.567357842513675, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6453053951263428, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8846666812896729, "num_tokens": 766364518.0, "step": 20182 }, { "epoch": 2.5674850527922657, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.306206464767456, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8678702116012573, "num_tokens": 766402256.0, "step": 20183 }, { "epoch": 2.567612263070856, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7526891231536865, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8793613910675049, "num_tokens": 766435186.0, "step": 20184 }, { "epoch": 2.5677394733494467, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2252843379974365, "learning_rate": 1e-05, "loss": 0.4647, "mean_token_accuracy": 0.8725213408470154, "num_tokens": 766476950.0, "step": 20185 }, { "epoch": 2.5678666836280373, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9195096492767334, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8923238515853882, "num_tokens": 766517908.0, "step": 20186 }, { "epoch": 2.567993893906628, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1950175762176514, "learning_rate": 1e-05, "loss": 0.4664, "mean_token_accuracy": 0.866147518157959, "num_tokens": 766556251.0, "step": 20187 }, { "epoch": 2.5681211041852183, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.158068656921387, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8675345778465271, "num_tokens": 766593865.0, "step": 20188 }, { "epoch": 2.5682483144638084, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2712180614471436, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8761717677116394, "num_tokens": 766627594.0, "step": 20189 }, { "epoch": 2.5683755247423994, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.352203845977783, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8849360346794128, "num_tokens": 766669620.0, "step": 20190 }, { "epoch": 2.5685027350209895, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.513457775115967, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8853944540023804, "num_tokens": 766711553.0, "step": 20191 }, { "epoch": 2.5686299452995804, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.3778486251831055, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8787014484405518, "num_tokens": 766751717.0, "step": 20192 }, { "epoch": 2.5687571555781705, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.7764065265655518, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8815896511077881, "num_tokens": 766782207.0, "step": 20193 }, { "epoch": 2.5688843658567615, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.174190044403076, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8743507266044617, "num_tokens": 766822050.0, "step": 20194 }, { "epoch": 2.5690115761353516, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7980759143829346, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8677819967269897, "num_tokens": 766867713.0, "step": 20195 }, { "epoch": 2.569138786413942, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9706878662109375, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8822924494743347, "num_tokens": 766908191.0, "step": 20196 }, { "epoch": 2.5692659966925326, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7317025661468506, "learning_rate": 1e-05, "loss": 0.4743, "mean_token_accuracy": 0.8681547045707703, "num_tokens": 766945774.0, "step": 20197 }, { "epoch": 2.569393206971123, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8070316314697266, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8931430578231812, "num_tokens": 766986255.0, "step": 20198 }, { "epoch": 2.5695204172497137, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.640101432800293, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8776997923851013, "num_tokens": 767020930.0, "step": 20199 }, { "epoch": 2.5696476275283042, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.716719388961792, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8737881183624268, "num_tokens": 767059863.0, "step": 20200 }, { "epoch": 2.5697748378068948, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.541325092315674, "learning_rate": 1e-05, "loss": 0.5138, "mean_token_accuracy": 0.8540052771568298, "num_tokens": 767100566.0, "step": 20201 }, { "epoch": 2.5699020480854853, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6764109134674072, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8710636496543884, "num_tokens": 767137892.0, "step": 20202 }, { "epoch": 2.570029258364076, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.460500955581665, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8592042326927185, "num_tokens": 767175340.0, "step": 20203 }, { "epoch": 2.5701564686426663, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.703206777572632, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8636887073516846, "num_tokens": 767209601.0, "step": 20204 }, { "epoch": 2.570283678921257, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3657097816467285, "learning_rate": 1e-05, "loss": 0.3731, "mean_token_accuracy": 0.896371603012085, "num_tokens": 767247500.0, "step": 20205 }, { "epoch": 2.5704108891998474, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7835779190063477, "learning_rate": 1e-05, "loss": 0.48, "mean_token_accuracy": 0.8650428056716919, "num_tokens": 767285251.0, "step": 20206 }, { "epoch": 2.570538099478438, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2976691722869873, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8880486488342285, "num_tokens": 767321493.0, "step": 20207 }, { "epoch": 2.5706653097570284, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.924776554107666, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8861098289489746, "num_tokens": 767351375.0, "step": 20208 }, { "epoch": 2.570792520035619, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2667834758758545, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8774970173835754, "num_tokens": 767392990.0, "step": 20209 }, { "epoch": 2.5709197303142095, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.0967888832092285, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8831969499588013, "num_tokens": 767431584.0, "step": 20210 }, { "epoch": 2.5710469405928, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3438572883605957, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8755040168762207, "num_tokens": 767468451.0, "step": 20211 }, { "epoch": 2.5711741508713906, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.209572792053223, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8724379539489746, "num_tokens": 767508133.0, "step": 20212 }, { "epoch": 2.571301361149981, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5159783363342285, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8846851587295532, "num_tokens": 767539427.0, "step": 20213 }, { "epoch": 2.571428571428571, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.198903560638428, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.878588080406189, "num_tokens": 767574842.0, "step": 20214 }, { "epoch": 2.571555781707162, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6682376861572266, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8767147660255432, "num_tokens": 767609378.0, "step": 20215 }, { "epoch": 2.5716829919857522, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.031855583190918, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8709653615951538, "num_tokens": 767642283.0, "step": 20216 }, { "epoch": 2.571810202264343, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6969711780548096, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.884519636631012, "num_tokens": 767680019.0, "step": 20217 }, { "epoch": 2.5719374125429333, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.835747003555298, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.872175395488739, "num_tokens": 767715570.0, "step": 20218 }, { "epoch": 2.572064622821524, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6694982051849365, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8794924020767212, "num_tokens": 767757257.0, "step": 20219 }, { "epoch": 2.5721918331001143, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6599485874176025, "learning_rate": 1e-05, "loss": 0.3727, "mean_token_accuracy": 0.8998675346374512, "num_tokens": 767791684.0, "step": 20220 }, { "epoch": 2.572319043378705, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6275906562805176, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8790092468261719, "num_tokens": 767831687.0, "step": 20221 }, { "epoch": 2.5724462536572954, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5363259315490723, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8848206996917725, "num_tokens": 767871349.0, "step": 20222 }, { "epoch": 2.572573463935886, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6128036975860596, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8866685032844543, "num_tokens": 767908301.0, "step": 20223 }, { "epoch": 2.5727006742144765, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.456812858581543, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8819900155067444, "num_tokens": 767944809.0, "step": 20224 }, { "epoch": 2.572827884493067, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.605642557144165, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8731642961502075, "num_tokens": 767984629.0, "step": 20225 }, { "epoch": 2.5729550947716575, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.417182445526123, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8664071559906006, "num_tokens": 768017697.0, "step": 20226 }, { "epoch": 2.573082305050248, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.598618745803833, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8737658262252808, "num_tokens": 768060259.0, "step": 20227 }, { "epoch": 2.5732095153288386, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.366702079772949, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8699307441711426, "num_tokens": 768096248.0, "step": 20228 }, { "epoch": 2.573336725607429, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.645519733428955, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8817099332809448, "num_tokens": 768131730.0, "step": 20229 }, { "epoch": 2.5734639358860196, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3083479404449463, "learning_rate": 1e-05, "loss": 0.4054, "mean_token_accuracy": 0.8874629735946655, "num_tokens": 768167551.0, "step": 20230 }, { "epoch": 2.57359114616461, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7388112545013428, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8787777423858643, "num_tokens": 768198212.0, "step": 20231 }, { "epoch": 2.5737183564432007, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2411093711853027, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.876859724521637, "num_tokens": 768233052.0, "step": 20232 }, { "epoch": 2.573845566721791, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.87736439704895, "learning_rate": 1e-05, "loss": 0.3967, "mean_token_accuracy": 0.8905134201049805, "num_tokens": 768272961.0, "step": 20233 }, { "epoch": 2.5739727770003817, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.184110641479492, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8872430920600891, "num_tokens": 768308748.0, "step": 20234 }, { "epoch": 2.5740999872789723, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.108994483947754, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.869565486907959, "num_tokens": 768346592.0, "step": 20235 }, { "epoch": 2.574227197557563, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.244447946548462, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8820841908454895, "num_tokens": 768394367.0, "step": 20236 }, { "epoch": 2.5743544078361533, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.323200225830078, "learning_rate": 1e-05, "loss": 0.4794, "mean_token_accuracy": 0.871959388256073, "num_tokens": 768437060.0, "step": 20237 }, { "epoch": 2.574481618114744, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.4750242233276367, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8777137398719788, "num_tokens": 768470373.0, "step": 20238 }, { "epoch": 2.574608828393334, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.372881889343262, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8865264654159546, "num_tokens": 768507567.0, "step": 20239 }, { "epoch": 2.574736038671925, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7022855281829834, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8871462345123291, "num_tokens": 768547217.0, "step": 20240 }, { "epoch": 2.574863248950515, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.23065185546875, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8850492835044861, "num_tokens": 768586720.0, "step": 20241 }, { "epoch": 2.574990459229106, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.785381317138672, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8826084136962891, "num_tokens": 768627890.0, "step": 20242 }, { "epoch": 2.575117669507696, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.042301177978516, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8713575005531311, "num_tokens": 768663393.0, "step": 20243 }, { "epoch": 2.5752448797862866, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.738126277923584, "learning_rate": 1e-05, "loss": 0.4134, "mean_token_accuracy": 0.8892531991004944, "num_tokens": 768706950.0, "step": 20244 }, { "epoch": 2.575372090064877, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.86967396736145, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8689367771148682, "num_tokens": 768747391.0, "step": 20245 }, { "epoch": 2.5754993003434676, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.652477264404297, "learning_rate": 1e-05, "loss": 0.3736, "mean_token_accuracy": 0.8969225287437439, "num_tokens": 768781170.0, "step": 20246 }, { "epoch": 2.575626510622058, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.754206418991089, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8746488094329834, "num_tokens": 768820505.0, "step": 20247 }, { "epoch": 2.5757537209006487, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5695438385009766, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8766024708747864, "num_tokens": 768859647.0, "step": 20248 }, { "epoch": 2.575880931179239, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6819422245025635, "learning_rate": 1e-05, "loss": 0.3857, "mean_token_accuracy": 0.8934168815612793, "num_tokens": 768894883.0, "step": 20249 }, { "epoch": 2.5760081414578297, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.485572576522827, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8907172679901123, "num_tokens": 768934648.0, "step": 20250 }, { "epoch": 2.5761353517364203, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.684746265411377, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.883751392364502, "num_tokens": 768969568.0, "step": 20251 }, { "epoch": 2.576262562015011, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4078643321990967, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8689225912094116, "num_tokens": 769009423.0, "step": 20252 }, { "epoch": 2.5763897722936013, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.747377634048462, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8817605376243591, "num_tokens": 769043812.0, "step": 20253 }, { "epoch": 2.576516982572192, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.330897569656372, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8826512694358826, "num_tokens": 769079959.0, "step": 20254 }, { "epoch": 2.5766441928507824, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8449223041534424, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8891019821166992, "num_tokens": 769119375.0, "step": 20255 }, { "epoch": 2.576771403129373, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2693164348602295, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8875938653945923, "num_tokens": 769157085.0, "step": 20256 }, { "epoch": 2.5768986134079634, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.020823001861572, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8747907876968384, "num_tokens": 769190883.0, "step": 20257 }, { "epoch": 2.577025823686554, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.298828363418579, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8806349039077759, "num_tokens": 769228370.0, "step": 20258 }, { "epoch": 2.5771530339651445, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.174386024475098, "learning_rate": 1e-05, "loss": 0.3696, "mean_token_accuracy": 0.9056629538536072, "num_tokens": 769265551.0, "step": 20259 }, { "epoch": 2.577280244243735, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.4526731967926025, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8828271627426147, "num_tokens": 769299482.0, "step": 20260 }, { "epoch": 2.5774074545223256, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.20120906829834, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8799874782562256, "num_tokens": 769339165.0, "step": 20261 }, { "epoch": 2.5775346648009156, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.628786325454712, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8846731185913086, "num_tokens": 769374066.0, "step": 20262 }, { "epoch": 2.5776618750795066, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.042168140411377, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8783676624298096, "num_tokens": 769415536.0, "step": 20263 }, { "epoch": 2.5777890853580967, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6820132732391357, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8757614493370056, "num_tokens": 769454016.0, "step": 20264 }, { "epoch": 2.5779162956366877, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8429713249206543, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8817102909088135, "num_tokens": 769493150.0, "step": 20265 }, { "epoch": 2.5780435059152778, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.692214012145996, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8748118877410889, "num_tokens": 769524247.0, "step": 20266 }, { "epoch": 2.5781707161938687, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.631352424621582, "learning_rate": 1e-05, "loss": 0.3933, "mean_token_accuracy": 0.8921743631362915, "num_tokens": 769566476.0, "step": 20267 }, { "epoch": 2.578297926472459, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6756350994110107, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8763573169708252, "num_tokens": 769603474.0, "step": 20268 }, { "epoch": 2.5784251367510493, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5249500274658203, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8754111528396606, "num_tokens": 769640446.0, "step": 20269 }, { "epoch": 2.57855234702964, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6448824405670166, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.885357141494751, "num_tokens": 769679809.0, "step": 20270 }, { "epoch": 2.5786795573082304, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.403106212615967, "learning_rate": 1e-05, "loss": 0.3806, "mean_token_accuracy": 0.8968645334243774, "num_tokens": 769717221.0, "step": 20271 }, { "epoch": 2.578806767586821, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6721980571746826, "learning_rate": 1e-05, "loss": 0.3808, "mean_token_accuracy": 0.8926015496253967, "num_tokens": 769749350.0, "step": 20272 }, { "epoch": 2.5789339778654115, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3392152786254883, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8701270818710327, "num_tokens": 769786604.0, "step": 20273 }, { "epoch": 2.579061188144002, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7301716804504395, "learning_rate": 1e-05, "loss": 0.3839, "mean_token_accuracy": 0.8936915993690491, "num_tokens": 769827709.0, "step": 20274 }, { "epoch": 2.5791883984225925, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2791149616241455, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8769801259040833, "num_tokens": 769866713.0, "step": 20275 }, { "epoch": 2.579315608701183, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.84999418258667, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8794685006141663, "num_tokens": 769907365.0, "step": 20276 }, { "epoch": 2.5794428189797736, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.232940196990967, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8796616196632385, "num_tokens": 769945573.0, "step": 20277 }, { "epoch": 2.579570029258364, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.02565336227417, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8755229115486145, "num_tokens": 769980936.0, "step": 20278 }, { "epoch": 2.5796972395369546, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2517244815826416, "learning_rate": 1e-05, "loss": 0.4761, "mean_token_accuracy": 0.8657650947570801, "num_tokens": 770022244.0, "step": 20279 }, { "epoch": 2.579824449815545, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.204473972320557, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8859121799468994, "num_tokens": 770059078.0, "step": 20280 }, { "epoch": 2.5799516600941357, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.3798322677612305, "learning_rate": 1e-05, "loss": 0.4741, "mean_token_accuracy": 0.8662208318710327, "num_tokens": 770100816.0, "step": 20281 }, { "epoch": 2.580078870372726, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.299510478973389, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8844116926193237, "num_tokens": 770138912.0, "step": 20282 }, { "epoch": 2.5802060806513167, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.5930614471435547, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8864299058914185, "num_tokens": 770178915.0, "step": 20283 }, { "epoch": 2.5803332909299073, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.226670742034912, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8876473307609558, "num_tokens": 770217707.0, "step": 20284 }, { "epoch": 2.580460501208498, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.726128339767456, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8788039684295654, "num_tokens": 770251659.0, "step": 20285 }, { "epoch": 2.5805877114870883, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.034991264343262, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8747701644897461, "num_tokens": 770291612.0, "step": 20286 }, { "epoch": 2.5807149217656784, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.712934732437134, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8719298839569092, "num_tokens": 770333822.0, "step": 20287 }, { "epoch": 2.5808421320442694, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.86301589012146, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.874419093132019, "num_tokens": 770369850.0, "step": 20288 }, { "epoch": 2.5809693423228595, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6630287170410156, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8669997453689575, "num_tokens": 770408825.0, "step": 20289 }, { "epoch": 2.5810965526014504, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.707390546798706, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.8923928737640381, "num_tokens": 770451320.0, "step": 20290 }, { "epoch": 2.5812237628800405, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.609694004058838, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8738365769386292, "num_tokens": 770488099.0, "step": 20291 }, { "epoch": 2.5813509731586315, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6175425052642822, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8705296516418457, "num_tokens": 770533689.0, "step": 20292 }, { "epoch": 2.5814781834372216, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.534679412841797, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8814464807510376, "num_tokens": 770570998.0, "step": 20293 }, { "epoch": 2.581605393715812, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5712997913360596, "learning_rate": 1e-05, "loss": 0.3917, "mean_token_accuracy": 0.891668438911438, "num_tokens": 770610248.0, "step": 20294 }, { "epoch": 2.5817326039944026, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.482332468032837, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8637124300003052, "num_tokens": 770650230.0, "step": 20295 }, { "epoch": 2.581859814272993, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.556037425994873, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8870798349380493, "num_tokens": 770694740.0, "step": 20296 }, { "epoch": 2.5819870245515837, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.433075428009033, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8859728574752808, "num_tokens": 770734004.0, "step": 20297 }, { "epoch": 2.582114234830174, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5818910598754883, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8759129047393799, "num_tokens": 770770421.0, "step": 20298 }, { "epoch": 2.5822414451087647, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3974661827087402, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8735238313674927, "num_tokens": 770810790.0, "step": 20299 }, { "epoch": 2.5823686553873553, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.623145580291748, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.875212550163269, "num_tokens": 770848614.0, "step": 20300 }, { "epoch": 2.582495865665946, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3256099224090576, "learning_rate": 1e-05, "loss": 0.3698, "mean_token_accuracy": 0.8982385396957397, "num_tokens": 770889888.0, "step": 20301 }, { "epoch": 2.5826230759445363, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.721632957458496, "learning_rate": 1e-05, "loss": 0.5089, "mean_token_accuracy": 0.8559331893920898, "num_tokens": 770925918.0, "step": 20302 }, { "epoch": 2.582750286223127, "ewc_loss": 0.08056640625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2375080585479736, "learning_rate": 1e-05, "loss": 0.3618, "mean_token_accuracy": 0.8964246511459351, "num_tokens": 770963130.0, "step": 20303 }, { "epoch": 2.5828774965017174, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8943536281585693, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8697094917297363, "num_tokens": 770998308.0, "step": 20304 }, { "epoch": 2.583004706780308, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1981568336486816, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.872352123260498, "num_tokens": 771032532.0, "step": 20305 }, { "epoch": 2.5831319170588984, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.118653774261475, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8869376182556152, "num_tokens": 771070859.0, "step": 20306 }, { "epoch": 2.583259127337489, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2424376010894775, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8729711771011353, "num_tokens": 771111216.0, "step": 20307 }, { "epoch": 2.5833863376160795, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.373338222503662, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8803761005401611, "num_tokens": 771146703.0, "step": 20308 }, { "epoch": 2.58351354789467, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5109434127807617, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8817434310913086, "num_tokens": 771178778.0, "step": 20309 }, { "epoch": 2.5836407581732606, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.393436908721924, "learning_rate": 1e-05, "loss": 0.4758, "mean_token_accuracy": 0.8691059947013855, "num_tokens": 771217272.0, "step": 20310 }, { "epoch": 2.583767968451851, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.767362594604492, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.87325119972229, "num_tokens": 771252085.0, "step": 20311 }, { "epoch": 2.583895178730441, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.2069220542907715, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8828680515289307, "num_tokens": 771285367.0, "step": 20312 }, { "epoch": 2.584022389009032, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.80340838432312, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8850716948509216, "num_tokens": 771324388.0, "step": 20313 }, { "epoch": 2.5841495992876222, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.9985809326171875, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8857188820838928, "num_tokens": 771361307.0, "step": 20314 }, { "epoch": 2.584276809566213, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7359838485717773, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.867308497428894, "num_tokens": 771400383.0, "step": 20315 }, { "epoch": 2.5844040198448033, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.831592321395874, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8898395299911499, "num_tokens": 771435065.0, "step": 20316 }, { "epoch": 2.584531230123394, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.646413803100586, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.8632354140281677, "num_tokens": 771480573.0, "step": 20317 }, { "epoch": 2.5846584404019843, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7287676334381104, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8845964670181274, "num_tokens": 771517172.0, "step": 20318 }, { "epoch": 2.584785650680575, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5806970596313477, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8678606152534485, "num_tokens": 771554347.0, "step": 20319 }, { "epoch": 2.5849128609591654, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6589479446411133, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.8850865960121155, "num_tokens": 771591209.0, "step": 20320 }, { "epoch": 2.585040071237756, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4835307598114014, "learning_rate": 1e-05, "loss": 0.4001, "mean_token_accuracy": 0.8902326226234436, "num_tokens": 771626604.0, "step": 20321 }, { "epoch": 2.5851672815163464, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.665919542312622, "learning_rate": 1e-05, "loss": 0.439, "mean_token_accuracy": 0.8767104744911194, "num_tokens": 771662525.0, "step": 20322 }, { "epoch": 2.585294491794937, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.3983426094055176, "learning_rate": 1e-05, "loss": 0.4015, "mean_token_accuracy": 0.8890854120254517, "num_tokens": 771701616.0, "step": 20323 }, { "epoch": 2.5854217020735275, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.713083505630493, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8714132308959961, "num_tokens": 771736864.0, "step": 20324 }, { "epoch": 2.585548912352118, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3251683712005615, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8871451616287231, "num_tokens": 771774900.0, "step": 20325 }, { "epoch": 2.5856761226307086, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.830925464630127, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8752435445785522, "num_tokens": 771812078.0, "step": 20326 }, { "epoch": 2.585803332909299, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.256258487701416, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8788570761680603, "num_tokens": 771852247.0, "step": 20327 }, { "epoch": 2.5859305431878896, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.004631042480469, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8749914765357971, "num_tokens": 771891264.0, "step": 20328 }, { "epoch": 2.58605775346648, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2681500911712646, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.866981029510498, "num_tokens": 771935356.0, "step": 20329 }, { "epoch": 2.5861849637450707, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.213702201843262, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8561614751815796, "num_tokens": 771967284.0, "step": 20330 }, { "epoch": 2.586312174023661, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.411207437515259, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8848477602005005, "num_tokens": 772013465.0, "step": 20331 }, { "epoch": 2.5864393843022517, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.293360233306885, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8740955591201782, "num_tokens": 772046978.0, "step": 20332 }, { "epoch": 2.5865665945808423, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6452972888946533, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8872133493423462, "num_tokens": 772081630.0, "step": 20333 }, { "epoch": 2.586693804859433, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.160720348358154, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8690173625946045, "num_tokens": 772122197.0, "step": 20334 }, { "epoch": 2.5868210151380233, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.734443187713623, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8810908198356628, "num_tokens": 772162744.0, "step": 20335 }, { "epoch": 2.586948225416614, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9376471042633057, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8729650974273682, "num_tokens": 772207520.0, "step": 20336 }, { "epoch": 2.587075435695204, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.725067377090454, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8775720596313477, "num_tokens": 772244869.0, "step": 20337 }, { "epoch": 2.587202645973795, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7554726600646973, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8801693320274353, "num_tokens": 772279132.0, "step": 20338 }, { "epoch": 2.587329856252385, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6627299785614014, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.8904292583465576, "num_tokens": 772323668.0, "step": 20339 }, { "epoch": 2.587457066530976, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6003949642181396, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8780023455619812, "num_tokens": 772367834.0, "step": 20340 }, { "epoch": 2.587584276809566, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6288325786590576, "learning_rate": 1e-05, "loss": 0.3778, "mean_token_accuracy": 0.8957962989807129, "num_tokens": 772398777.0, "step": 20341 }, { "epoch": 2.5877114870881566, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5172953605651855, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8781031966209412, "num_tokens": 772432330.0, "step": 20342 }, { "epoch": 2.587838697366747, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.58912992477417, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8713498115539551, "num_tokens": 772472412.0, "step": 20343 }, { "epoch": 2.5879659076453376, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4551267623901367, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8780758380889893, "num_tokens": 772509896.0, "step": 20344 }, { "epoch": 2.588093117923928, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5697765350341797, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8873306512832642, "num_tokens": 772549443.0, "step": 20345 }, { "epoch": 2.5882203282025187, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.414712429046631, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8678858280181885, "num_tokens": 772592086.0, "step": 20346 }, { "epoch": 2.588347538481109, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.583493709564209, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8654555678367615, "num_tokens": 772632805.0, "step": 20347 }, { "epoch": 2.5884747487596997, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.365037202835083, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8676983118057251, "num_tokens": 772673806.0, "step": 20348 }, { "epoch": 2.5886019590382903, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.625046730041504, "learning_rate": 1e-05, "loss": 0.3748, "mean_token_accuracy": 0.9013323187828064, "num_tokens": 772709434.0, "step": 20349 }, { "epoch": 2.588729169316881, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3054652214050293, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8808860778808594, "num_tokens": 772748702.0, "step": 20350 }, { "epoch": 2.5888563795954713, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.721064329147339, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8615444302558899, "num_tokens": 772792235.0, "step": 20351 }, { "epoch": 2.588983589874062, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.232011079788208, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.872616171836853, "num_tokens": 772827253.0, "step": 20352 }, { "epoch": 2.5891108001526524, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.872337818145752, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8912378549575806, "num_tokens": 772871951.0, "step": 20353 }, { "epoch": 2.589238010431243, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1678824424743652, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8871338367462158, "num_tokens": 772908585.0, "step": 20354 }, { "epoch": 2.5893652207098334, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.130892276763916, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8719020485877991, "num_tokens": 772946883.0, "step": 20355 }, { "epoch": 2.589492430988424, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.21343994140625, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8831219673156738, "num_tokens": 772985143.0, "step": 20356 }, { "epoch": 2.5896196412670145, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.390370845794678, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8843218088150024, "num_tokens": 773027004.0, "step": 20357 }, { "epoch": 2.589746851545605, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.478316068649292, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8676676154136658, "num_tokens": 773067807.0, "step": 20358 }, { "epoch": 2.5898740618241956, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.473805904388428, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.883942723274231, "num_tokens": 773103160.0, "step": 20359 }, { "epoch": 2.5900012721027856, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.7998154163360596, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8856394290924072, "num_tokens": 773143036.0, "step": 20360 }, { "epoch": 2.5901284823813766, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.286632061004639, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8843312859535217, "num_tokens": 773180930.0, "step": 20361 }, { "epoch": 2.5902556926599667, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.881260633468628, "learning_rate": 1e-05, "loss": 0.5034, "mean_token_accuracy": 0.8613067865371704, "num_tokens": 773220056.0, "step": 20362 }, { "epoch": 2.5903829029385577, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.052224159240723, "learning_rate": 1e-05, "loss": 0.3878, "mean_token_accuracy": 0.8963928818702698, "num_tokens": 773255890.0, "step": 20363 }, { "epoch": 2.5905101132171477, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7792739868164062, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8827939629554749, "num_tokens": 773298727.0, "step": 20364 }, { "epoch": 2.5906373234957387, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8877034187316895, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8604933023452759, "num_tokens": 773338137.0, "step": 20365 }, { "epoch": 2.590764533774329, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.66428279876709, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8861899375915527, "num_tokens": 773381532.0, "step": 20366 }, { "epoch": 2.5908917440529193, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7974610328674316, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8673826456069946, "num_tokens": 773416354.0, "step": 20367 }, { "epoch": 2.59101895433151, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5644118785858154, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8776439428329468, "num_tokens": 773456130.0, "step": 20368 }, { "epoch": 2.5911461646101004, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7442357540130615, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8863173723220825, "num_tokens": 773493812.0, "step": 20369 }, { "epoch": 2.591273374888691, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.462547779083252, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8815051317214966, "num_tokens": 773533000.0, "step": 20370 }, { "epoch": 2.5914005851672814, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7722084522247314, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8744907379150391, "num_tokens": 773572355.0, "step": 20371 }, { "epoch": 2.591527795445872, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.373201608657837, "learning_rate": 1e-05, "loss": 0.3946, "mean_token_accuracy": 0.8900988698005676, "num_tokens": 773609154.0, "step": 20372 }, { "epoch": 2.5916550057244625, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.853780746459961, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8876805305480957, "num_tokens": 773654430.0, "step": 20373 }, { "epoch": 2.591782216003053, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3030128479003906, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8846988677978516, "num_tokens": 773693710.0, "step": 20374 }, { "epoch": 2.5919094262816436, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.021912097930908, "learning_rate": 1e-05, "loss": 0.5027, "mean_token_accuracy": 0.853405237197876, "num_tokens": 773728858.0, "step": 20375 }, { "epoch": 2.592036636560234, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3421735763549805, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8717736005783081, "num_tokens": 773761080.0, "step": 20376 }, { "epoch": 2.5921638468388246, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.148925304412842, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.891032338142395, "num_tokens": 773799084.0, "step": 20377 }, { "epoch": 2.592291057117415, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4407432079315186, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.884247899055481, "num_tokens": 773840378.0, "step": 20378 }, { "epoch": 2.5924182673960057, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.187742710113525, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8822145462036133, "num_tokens": 773878627.0, "step": 20379 }, { "epoch": 2.592545477674596, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.612290143966675, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8695818781852722, "num_tokens": 773921492.0, "step": 20380 }, { "epoch": 2.5926726879531867, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.056288242340088, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8633825182914734, "num_tokens": 773959603.0, "step": 20381 }, { "epoch": 2.5927998982317773, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.696605920791626, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8685486912727356, "num_tokens": 773996550.0, "step": 20382 }, { "epoch": 2.592927108510368, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8342599868774414, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8791015148162842, "num_tokens": 774033815.0, "step": 20383 }, { "epoch": 2.5930543187889583, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.6970055103302, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8786706924438477, "num_tokens": 774068930.0, "step": 20384 }, { "epoch": 2.5931815290675484, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6296305656433105, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8766505718231201, "num_tokens": 774112014.0, "step": 20385 }, { "epoch": 2.5933087393461394, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.680861234664917, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8781935572624207, "num_tokens": 774149676.0, "step": 20386 }, { "epoch": 2.5934359496247295, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4727365970611572, "learning_rate": 1e-05, "loss": 0.3749, "mean_token_accuracy": 0.8978426456451416, "num_tokens": 774190160.0, "step": 20387 }, { "epoch": 2.5935631599033204, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6905176639556885, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8769630193710327, "num_tokens": 774235133.0, "step": 20388 }, { "epoch": 2.5936903701819105, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3900322914123535, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.8839391469955444, "num_tokens": 774277962.0, "step": 20389 }, { "epoch": 2.5938175804605015, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7239556312561035, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8819249868392944, "num_tokens": 774316946.0, "step": 20390 }, { "epoch": 2.5939447907390916, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.301151752471924, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8839598894119263, "num_tokens": 774356376.0, "step": 20391 }, { "epoch": 2.594072001017682, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.813448905944824, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.889382004737854, "num_tokens": 774392543.0, "step": 20392 }, { "epoch": 2.5941992112962726, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2583909034729004, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8735663294792175, "num_tokens": 774432114.0, "step": 20393 }, { "epoch": 2.594326421574863, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9572415351867676, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8705854415893555, "num_tokens": 774468758.0, "step": 20394 }, { "epoch": 2.5944536318534537, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.259056329727173, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.866054892539978, "num_tokens": 774503136.0, "step": 20395 }, { "epoch": 2.594580842132044, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.124812602996826, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8876203894615173, "num_tokens": 774545162.0, "step": 20396 }, { "epoch": 2.5947080524106347, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3339123725891113, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8800005316734314, "num_tokens": 774581831.0, "step": 20397 }, { "epoch": 2.5948352626892253, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.253780364990234, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8847531676292419, "num_tokens": 774619258.0, "step": 20398 }, { "epoch": 2.594962472967816, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5341274738311768, "learning_rate": 1e-05, "loss": 0.4746, "mean_token_accuracy": 0.8652457594871521, "num_tokens": 774650244.0, "step": 20399 }, { "epoch": 2.5950896832464063, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.211668491363525, "learning_rate": 1e-05, "loss": 0.4921, "mean_token_accuracy": 0.865240216255188, "num_tokens": 774690060.0, "step": 20400 }, { "epoch": 2.595216893524997, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6511871814727783, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8761113286018372, "num_tokens": 774733925.0, "step": 20401 }, { "epoch": 2.5953441038035874, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.078527927398682, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8831421732902527, "num_tokens": 774769344.0, "step": 20402 }, { "epoch": 2.595471314082178, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6815683841705322, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8948774337768555, "num_tokens": 774809548.0, "step": 20403 }, { "epoch": 2.5955985243607684, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9126968383789062, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8761811852455139, "num_tokens": 774848042.0, "step": 20404 }, { "epoch": 2.595725734639359, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.669455051422119, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8727914094924927, "num_tokens": 774884226.0, "step": 20405 }, { "epoch": 2.5958529449179495, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7719743251800537, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8765179514884949, "num_tokens": 774913697.0, "step": 20406 }, { "epoch": 2.59598015519654, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.609755754470825, "learning_rate": 1e-05, "loss": 0.3871, "mean_token_accuracy": 0.894688606262207, "num_tokens": 774948369.0, "step": 20407 }, { "epoch": 2.5961073654751305, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6520907878875732, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8827703595161438, "num_tokens": 774992260.0, "step": 20408 }, { "epoch": 2.596234575753721, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5874991416931152, "learning_rate": 1e-05, "loss": 0.545, "mean_token_accuracy": 0.8387153148651123, "num_tokens": 775023613.0, "step": 20409 }, { "epoch": 2.596361786032311, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5765786170959473, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8747068643569946, "num_tokens": 775063874.0, "step": 20410 }, { "epoch": 2.596488996310902, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5121700763702393, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8738187551498413, "num_tokens": 775104447.0, "step": 20411 }, { "epoch": 2.596616206589492, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.541076183319092, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8867731094360352, "num_tokens": 775143472.0, "step": 20412 }, { "epoch": 2.596743416868083, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4784624576568604, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8851246237754822, "num_tokens": 775176274.0, "step": 20413 }, { "epoch": 2.5968706271466733, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5438902378082275, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.872172474861145, "num_tokens": 775216485.0, "step": 20414 }, { "epoch": 2.596997837425264, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.415102243423462, "learning_rate": 1e-05, "loss": 0.3957, "mean_token_accuracy": 0.8885798454284668, "num_tokens": 775256828.0, "step": 20415 }, { "epoch": 2.5971250477038543, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5855188369750977, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8699073791503906, "num_tokens": 775296577.0, "step": 20416 }, { "epoch": 2.597252257982445, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3729934692382812, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8782378435134888, "num_tokens": 775329567.0, "step": 20417 }, { "epoch": 2.5973794682610354, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6458382606506348, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8832143545150757, "num_tokens": 775363729.0, "step": 20418 }, { "epoch": 2.597506678539626, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3079750537872314, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8783571124076843, "num_tokens": 775397547.0, "step": 20419 }, { "epoch": 2.5976338888182164, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7463369369506836, "learning_rate": 1e-05, "loss": 0.4588, "mean_token_accuracy": 0.8696991801261902, "num_tokens": 775441033.0, "step": 20420 }, { "epoch": 2.597761099096807, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2236034870147705, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.884412407875061, "num_tokens": 775477516.0, "step": 20421 }, { "epoch": 2.5978883093753975, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.927157163619995, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8839322328567505, "num_tokens": 775515573.0, "step": 20422 }, { "epoch": 2.598015519653988, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1775081157684326, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8828452229499817, "num_tokens": 775552970.0, "step": 20423 }, { "epoch": 2.5981427299325786, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.183868408203125, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8755514621734619, "num_tokens": 775590180.0, "step": 20424 }, { "epoch": 2.598269940211169, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.2727885246276855, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8835364580154419, "num_tokens": 775627346.0, "step": 20425 }, { "epoch": 2.5983971504897596, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.405745983123779, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8772400617599487, "num_tokens": 775667173.0, "step": 20426 }, { "epoch": 2.59852436076835, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.568610668182373, "learning_rate": 1e-05, "loss": 0.4754, "mean_token_accuracy": 0.8664759993553162, "num_tokens": 775702710.0, "step": 20427 }, { "epoch": 2.5986515710469407, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.39235258102417, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8783658742904663, "num_tokens": 775743934.0, "step": 20428 }, { "epoch": 2.598778781325531, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.7863948345184326, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8827404975891113, "num_tokens": 775783184.0, "step": 20429 }, { "epoch": 2.5989059916041217, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.205705165863037, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8715298771858215, "num_tokens": 775820019.0, "step": 20430 }, { "epoch": 2.5990332018827123, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.823800802230835, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8791064620018005, "num_tokens": 775858740.0, "step": 20431 }, { "epoch": 2.599160412161303, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9929285049438477, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8778025507926941, "num_tokens": 775900615.0, "step": 20432 }, { "epoch": 2.5992876224398933, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.729123115539551, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8887125849723816, "num_tokens": 775938953.0, "step": 20433 }, { "epoch": 2.599414832718484, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8517813682556152, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.875465452671051, "num_tokens": 775973203.0, "step": 20434 }, { "epoch": 2.599542042997074, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6320059299468994, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8673977255821228, "num_tokens": 776014984.0, "step": 20435 }, { "epoch": 2.599669253275665, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.753119707107544, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.863063633441925, "num_tokens": 776059883.0, "step": 20436 }, { "epoch": 2.599796463554255, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5417299270629883, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8660762310028076, "num_tokens": 776101025.0, "step": 20437 }, { "epoch": 2.599923673832846, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.72509765625, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8837231397628784, "num_tokens": 776134356.0, "step": 20438 }, { "epoch": 2.600050884111436, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.435434103012085, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8844648599624634, "num_tokens": 776176158.0, "step": 20439 }, { "epoch": 2.6001780943900266, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7683403491973877, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8800193071365356, "num_tokens": 776209586.0, "step": 20440 }, { "epoch": 2.600305304668617, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3738315105438232, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8706141710281372, "num_tokens": 776240707.0, "step": 20441 }, { "epoch": 2.6004325149472076, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8388824462890625, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8754273056983948, "num_tokens": 776280448.0, "step": 20442 }, { "epoch": 2.600559725225798, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2988579273223877, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8915456533432007, "num_tokens": 776320112.0, "step": 20443 }, { "epoch": 2.6006869355043887, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.977231025695801, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8763355612754822, "num_tokens": 776362720.0, "step": 20444 }, { "epoch": 2.600814145782979, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.3122918605804443, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8754034638404846, "num_tokens": 776403006.0, "step": 20445 }, { "epoch": 2.6009413560615697, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1260151863098145, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8748399019241333, "num_tokens": 776436039.0, "step": 20446 }, { "epoch": 2.6010685663401603, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.418057441711426, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8787185549736023, "num_tokens": 776476337.0, "step": 20447 }, { "epoch": 2.601195776618751, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.169666767120361, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.888789176940918, "num_tokens": 776512552.0, "step": 20448 }, { "epoch": 2.6013229868973413, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.5835046768188477, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8755714297294617, "num_tokens": 776549727.0, "step": 20449 }, { "epoch": 2.601450197175932, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.064607620239258, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8751523494720459, "num_tokens": 776586701.0, "step": 20450 }, { "epoch": 2.6015774074545224, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.679574489593506, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8759364485740662, "num_tokens": 776627386.0, "step": 20451 }, { "epoch": 2.601704617733113, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.858118772506714, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.883583664894104, "num_tokens": 776668389.0, "step": 20452 }, { "epoch": 2.6018318280117034, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6711642742156982, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8742152452468872, "num_tokens": 776708106.0, "step": 20453 }, { "epoch": 2.601959038290294, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6649270057678223, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8788459300994873, "num_tokens": 776744577.0, "step": 20454 }, { "epoch": 2.6020862485688845, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6622273921966553, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8657249808311462, "num_tokens": 776784350.0, "step": 20455 }, { "epoch": 2.602213458847475, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.512856960296631, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8868207931518555, "num_tokens": 776823728.0, "step": 20456 }, { "epoch": 2.6023406691260655, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6672494411468506, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.879205584526062, "num_tokens": 776860097.0, "step": 20457 }, { "epoch": 2.6024678794046556, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.410464286804199, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8824530839920044, "num_tokens": 776895876.0, "step": 20458 }, { "epoch": 2.6025950896832466, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.7057600021362305, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8728238344192505, "num_tokens": 776930003.0, "step": 20459 }, { "epoch": 2.6027222999618367, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3144783973693848, "learning_rate": 1e-05, "loss": 0.3653, "mean_token_accuracy": 0.897503137588501, "num_tokens": 776969699.0, "step": 20460 }, { "epoch": 2.6028495102404277, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8187782764434814, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8812753558158875, "num_tokens": 777006517.0, "step": 20461 }, { "epoch": 2.6029767205190177, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.255401372909546, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8605051040649414, "num_tokens": 777044160.0, "step": 20462 }, { "epoch": 2.6031039307976087, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9443297386169434, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.881621241569519, "num_tokens": 777079890.0, "step": 20463 }, { "epoch": 2.603231141076199, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.229823589324951, "learning_rate": 1e-05, "loss": 0.3912, "mean_token_accuracy": 0.8909596800804138, "num_tokens": 777120928.0, "step": 20464 }, { "epoch": 2.6033583513547893, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.114737510681152, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8667929172515869, "num_tokens": 777166363.0, "step": 20465 }, { "epoch": 2.60348556163338, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.3353960514068604, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8782160878181458, "num_tokens": 777201146.0, "step": 20466 }, { "epoch": 2.6036127719119704, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.246410846710205, "learning_rate": 1e-05, "loss": 0.5124, "mean_token_accuracy": 0.8597599864006042, "num_tokens": 777239884.0, "step": 20467 }, { "epoch": 2.603739982190561, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.50709867477417, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.864188015460968, "num_tokens": 777279898.0, "step": 20468 }, { "epoch": 2.6038671924691514, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.232685089111328, "learning_rate": 1e-05, "loss": 0.4755, "mean_token_accuracy": 0.8739066123962402, "num_tokens": 777320614.0, "step": 20469 }, { "epoch": 2.603994402747742, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6682851314544678, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8750841617584229, "num_tokens": 777364782.0, "step": 20470 }, { "epoch": 2.6041216130263325, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1036529541015625, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8708973526954651, "num_tokens": 777398276.0, "step": 20471 }, { "epoch": 2.604248823304923, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7037017345428467, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8850134611129761, "num_tokens": 777434510.0, "step": 20472 }, { "epoch": 2.6043760335835135, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.919447660446167, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8801225423812866, "num_tokens": 777469266.0, "step": 20473 }, { "epoch": 2.604503243862104, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.670322895050049, "learning_rate": 1e-05, "loss": 0.3921, "mean_token_accuracy": 0.8914422392845154, "num_tokens": 777500770.0, "step": 20474 }, { "epoch": 2.6046304541406946, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.765692710876465, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8791984915733337, "num_tokens": 777537977.0, "step": 20475 }, { "epoch": 2.604757664419285, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.618664264678955, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8819090127944946, "num_tokens": 777574065.0, "step": 20476 }, { "epoch": 2.6048848746978757, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6545164585113525, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8861532807350159, "num_tokens": 777613366.0, "step": 20477 }, { "epoch": 2.605012084976466, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.581110954284668, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8747186064720154, "num_tokens": 777650954.0, "step": 20478 }, { "epoch": 2.6051392952550567, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5853772163391113, "learning_rate": 1e-05, "loss": 0.4676, "mean_token_accuracy": 0.8661609888076782, "num_tokens": 777694370.0, "step": 20479 }, { "epoch": 2.6052665055336472, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5076944828033447, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8757064342498779, "num_tokens": 777734039.0, "step": 20480 }, { "epoch": 2.6053937158122378, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5501813888549805, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8770200610160828, "num_tokens": 777776237.0, "step": 20481 }, { "epoch": 2.6055209260908283, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.462923526763916, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8827239274978638, "num_tokens": 777813767.0, "step": 20482 }, { "epoch": 2.6056481363694184, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5569446086883545, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8814066648483276, "num_tokens": 777848642.0, "step": 20483 }, { "epoch": 2.6057753466480094, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4102392196655273, "learning_rate": 1e-05, "loss": 0.4063, "mean_token_accuracy": 0.8840134739875793, "num_tokens": 777885748.0, "step": 20484 }, { "epoch": 2.6059025569265994, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.726977586746216, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8656466007232666, "num_tokens": 777922366.0, "step": 20485 }, { "epoch": 2.6060297672051904, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.284813642501831, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.877295970916748, "num_tokens": 777965130.0, "step": 20486 }, { "epoch": 2.6061569774837805, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7965264320373535, "learning_rate": 1e-05, "loss": 0.3897, "mean_token_accuracy": 0.8920833468437195, "num_tokens": 777999425.0, "step": 20487 }, { "epoch": 2.6062841877623715, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.197970151901245, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8660658597946167, "num_tokens": 778042378.0, "step": 20488 }, { "epoch": 2.6064113980409616, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.015475749969482, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8794663548469543, "num_tokens": 778083858.0, "step": 20489 }, { "epoch": 2.606538608319552, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.199430227279663, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8859392404556274, "num_tokens": 778122103.0, "step": 20490 }, { "epoch": 2.6066658185981426, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.269564151763916, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8801636099815369, "num_tokens": 778158207.0, "step": 20491 }, { "epoch": 2.606793028876733, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.3834879398345947, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8843884468078613, "num_tokens": 778191542.0, "step": 20492 }, { "epoch": 2.6069202391553237, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.386612415313721, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8809784650802612, "num_tokens": 778234929.0, "step": 20493 }, { "epoch": 2.607047449433914, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.646880626678467, "learning_rate": 1e-05, "loss": 0.3828, "mean_token_accuracy": 0.8966310620307922, "num_tokens": 778273225.0, "step": 20494 }, { "epoch": 2.6071746597125047, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.2929911613464355, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8743736743927002, "num_tokens": 778310998.0, "step": 20495 }, { "epoch": 2.6073018699910953, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.8140664100646973, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8713498711585999, "num_tokens": 778348379.0, "step": 20496 }, { "epoch": 2.607429080269686, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.067641258239746, "learning_rate": 1e-05, "loss": 0.4617, "mean_token_accuracy": 0.8741455674171448, "num_tokens": 778390742.0, "step": 20497 }, { "epoch": 2.6075562905482763, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.776184320449829, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8836036920547485, "num_tokens": 778429239.0, "step": 20498 }, { "epoch": 2.607683500826867, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.8727643489837646, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8881409764289856, "num_tokens": 778467606.0, "step": 20499 }, { "epoch": 2.6078107111054574, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.7084784507751465, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.870392918586731, "num_tokens": 778500790.0, "step": 20500 }, { "epoch": 2.607937921384048, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.722020387649536, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8814343214035034, "num_tokens": 778536412.0, "step": 20501 }, { "epoch": 2.6080651316626384, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.617344379425049, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8725464344024658, "num_tokens": 778579812.0, "step": 20502 }, { "epoch": 2.608192341941229, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6225340366363525, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8864427208900452, "num_tokens": 778626135.0, "step": 20503 }, { "epoch": 2.6083195522198195, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5391764640808105, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8839988112449646, "num_tokens": 778666163.0, "step": 20504 }, { "epoch": 2.60844676249841, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5907013416290283, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8819841742515564, "num_tokens": 778703168.0, "step": 20505 }, { "epoch": 2.6085739727770005, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4929425716400146, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8695870637893677, "num_tokens": 778741696.0, "step": 20506 }, { "epoch": 2.608701183055591, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5789811611175537, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8781726360321045, "num_tokens": 778777450.0, "step": 20507 }, { "epoch": 2.608828393334181, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.441328525543213, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8797892928123474, "num_tokens": 778811448.0, "step": 20508 }, { "epoch": 2.608955603612772, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6536436080932617, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8771255612373352, "num_tokens": 778849333.0, "step": 20509 }, { "epoch": 2.609082813891362, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3489840030670166, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8854918479919434, "num_tokens": 778885264.0, "step": 20510 }, { "epoch": 2.609210024169953, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.718613862991333, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8770222067832947, "num_tokens": 778918160.0, "step": 20511 }, { "epoch": 2.6093372344485433, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2547085285186768, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8852196335792542, "num_tokens": 778955938.0, "step": 20512 }, { "epoch": 2.609464444727134, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8703207969665527, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8736433982849121, "num_tokens": 778994412.0, "step": 20513 }, { "epoch": 2.6095916550057243, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2076969146728516, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8667620420455933, "num_tokens": 779030818.0, "step": 20514 }, { "epoch": 2.609718865284315, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.082433223724365, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8881027102470398, "num_tokens": 779070217.0, "step": 20515 }, { "epoch": 2.6098460755629054, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2348012924194336, "learning_rate": 1e-05, "loss": 0.3913, "mean_token_accuracy": 0.8905712962150574, "num_tokens": 779105668.0, "step": 20516 }, { "epoch": 2.609973285841496, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.319822311401367, "learning_rate": 1e-05, "loss": 0.3913, "mean_token_accuracy": 0.8992472887039185, "num_tokens": 779140005.0, "step": 20517 }, { "epoch": 2.6101004961200864, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.46635365486145, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8804357051849365, "num_tokens": 779176269.0, "step": 20518 }, { "epoch": 2.610227706398677, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.373254299163818, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8908716440200806, "num_tokens": 779209914.0, "step": 20519 }, { "epoch": 2.6103549166772675, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.724277973175049, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8886082172393799, "num_tokens": 779244994.0, "step": 20520 }, { "epoch": 2.610482126955858, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.209347248077393, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8853695392608643, "num_tokens": 779283700.0, "step": 20521 }, { "epoch": 2.6106093372344485, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.793837070465088, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8889737725257874, "num_tokens": 779324827.0, "step": 20522 }, { "epoch": 2.610736547513039, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.9912314414978027, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8699315786361694, "num_tokens": 779363593.0, "step": 20523 }, { "epoch": 2.6108637577916296, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7462337017059326, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8838357925415039, "num_tokens": 779401295.0, "step": 20524 }, { "epoch": 2.61099096807022, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.813521146774292, "learning_rate": 1e-05, "loss": 0.4819, "mean_token_accuracy": 0.8633261919021606, "num_tokens": 779440375.0, "step": 20525 }, { "epoch": 2.6111181783488107, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.67905330657959, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8764793872833252, "num_tokens": 779475881.0, "step": 20526 }, { "epoch": 2.611245388627401, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.674034833908081, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8690260648727417, "num_tokens": 779520155.0, "step": 20527 }, { "epoch": 2.6113725989059917, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.589155673980713, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8761760592460632, "num_tokens": 779563952.0, "step": 20528 }, { "epoch": 2.6114998091845822, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5978586673736572, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8891203999519348, "num_tokens": 779601618.0, "step": 20529 }, { "epoch": 2.6116270194631728, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5204885005950928, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8732162714004517, "num_tokens": 779645228.0, "step": 20530 }, { "epoch": 2.6117542297417633, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5862224102020264, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.877461850643158, "num_tokens": 779681780.0, "step": 20531 }, { "epoch": 2.611881440020354, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4608280658721924, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8807978630065918, "num_tokens": 779717783.0, "step": 20532 }, { "epoch": 2.612008650298944, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.584620475769043, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8847038149833679, "num_tokens": 779752542.0, "step": 20533 }, { "epoch": 2.612135860577535, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3975539207458496, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8791065216064453, "num_tokens": 779795134.0, "step": 20534 }, { "epoch": 2.612263070856125, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.631309986114502, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8851293325424194, "num_tokens": 779828589.0, "step": 20535 }, { "epoch": 2.612390281134716, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3433432579040527, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8663800358772278, "num_tokens": 779865904.0, "step": 20536 }, { "epoch": 2.612517491413306, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.694709062576294, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8801148533821106, "num_tokens": 779910204.0, "step": 20537 }, { "epoch": 2.6126447016918966, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2611637115478516, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8861784338951111, "num_tokens": 779951032.0, "step": 20538 }, { "epoch": 2.612771911970487, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.845099449157715, "learning_rate": 1e-05, "loss": 0.4033, "mean_token_accuracy": 0.889889121055603, "num_tokens": 779983316.0, "step": 20539 }, { "epoch": 2.6128991222490776, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.211534261703491, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8795396089553833, "num_tokens": 780021167.0, "step": 20540 }, { "epoch": 2.613026332527668, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.05156135559082, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8787769079208374, "num_tokens": 780061276.0, "step": 20541 }, { "epoch": 2.6131535428062587, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.234036922454834, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8827624320983887, "num_tokens": 780097204.0, "step": 20542 }, { "epoch": 2.613280753084849, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.280282497406006, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8706473112106323, "num_tokens": 780137208.0, "step": 20543 }, { "epoch": 2.6134079633634397, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.411299705505371, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8723280429840088, "num_tokens": 780177815.0, "step": 20544 }, { "epoch": 2.6135351736420303, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.382662296295166, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.8935341835021973, "num_tokens": 780215798.0, "step": 20545 }, { "epoch": 2.613662383920621, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.697477340698242, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8623368740081787, "num_tokens": 780252252.0, "step": 20546 }, { "epoch": 2.6137895941992113, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.24055290222168, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8954554200172424, "num_tokens": 780288100.0, "step": 20547 }, { "epoch": 2.613916804477802, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.80452823638916, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8788408041000366, "num_tokens": 780331922.0, "step": 20548 }, { "epoch": 2.6140440147563924, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.028926372528076, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8721208572387695, "num_tokens": 780369930.0, "step": 20549 }, { "epoch": 2.614171225034983, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.750434398651123, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8792939186096191, "num_tokens": 780403893.0, "step": 20550 }, { "epoch": 2.6142984353135734, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8401215076446533, "learning_rate": 1e-05, "loss": 0.4022, "mean_token_accuracy": 0.8905160427093506, "num_tokens": 780438426.0, "step": 20551 }, { "epoch": 2.614425645592164, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6574337482452393, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.88229900598526, "num_tokens": 780483267.0, "step": 20552 }, { "epoch": 2.6145528558707545, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7200822830200195, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8817030191421509, "num_tokens": 780520355.0, "step": 20553 }, { "epoch": 2.614680066149345, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.571265459060669, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.879177451133728, "num_tokens": 780563449.0, "step": 20554 }, { "epoch": 2.6148072764279355, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6591079235076904, "learning_rate": 1e-05, "loss": 0.3908, "mean_token_accuracy": 0.8911343216896057, "num_tokens": 780598607.0, "step": 20555 }, { "epoch": 2.6149344867065256, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4866201877593994, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8805063366889954, "num_tokens": 780638425.0, "step": 20556 }, { "epoch": 2.6150616969851166, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6479592323303223, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8687276840209961, "num_tokens": 780683441.0, "step": 20557 }, { "epoch": 2.6151889072637067, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4154441356658936, "learning_rate": 1e-05, "loss": 0.3838, "mean_token_accuracy": 0.8943206071853638, "num_tokens": 780720281.0, "step": 20558 }, { "epoch": 2.6153161175422976, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6724836826324463, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8897948265075684, "num_tokens": 780761038.0, "step": 20559 }, { "epoch": 2.6154433278208877, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3544468879699707, "learning_rate": 1e-05, "loss": 0.4962, "mean_token_accuracy": 0.8561890721321106, "num_tokens": 780802358.0, "step": 20560 }, { "epoch": 2.6155705380994787, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7533187866210938, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.882089376449585, "num_tokens": 780839820.0, "step": 20561 }, { "epoch": 2.615697748378069, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.280184507369995, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8789178133010864, "num_tokens": 780879969.0, "step": 20562 }, { "epoch": 2.6158249586566593, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8931100368499756, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8805350065231323, "num_tokens": 780916119.0, "step": 20563 }, { "epoch": 2.61595216893525, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2401862144470215, "learning_rate": 1e-05, "loss": 0.4245, "mean_token_accuracy": 0.877516508102417, "num_tokens": 780957311.0, "step": 20564 }, { "epoch": 2.6160793792138404, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.071883201599121, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8738538026809692, "num_tokens": 780998322.0, "step": 20565 }, { "epoch": 2.616206589492431, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3025357723236084, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8647632002830505, "num_tokens": 781036249.0, "step": 20566 }, { "epoch": 2.6163337997710214, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.223926544189453, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.881065845489502, "num_tokens": 781075228.0, "step": 20567 }, { "epoch": 2.616461010049612, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.469912052154541, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.886596143245697, "num_tokens": 781110907.0, "step": 20568 }, { "epoch": 2.6165882203282025, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.254833221435547, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.869911789894104, "num_tokens": 781148615.0, "step": 20569 }, { "epoch": 2.616715430606793, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6590988636016846, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8866264820098877, "num_tokens": 781181159.0, "step": 20570 }, { "epoch": 2.6168426408853835, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1074748039245605, "learning_rate": 1e-05, "loss": 0.5175, "mean_token_accuracy": 0.8562232255935669, "num_tokens": 781221563.0, "step": 20571 }, { "epoch": 2.616969851163974, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7202768325805664, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.867343544960022, "num_tokens": 781257606.0, "step": 20572 }, { "epoch": 2.6170970614425646, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8926663398742676, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8764875531196594, "num_tokens": 781296653.0, "step": 20573 }, { "epoch": 2.617224271721155, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6877694129943848, "learning_rate": 1e-05, "loss": 0.3875, "mean_token_accuracy": 0.8939738869667053, "num_tokens": 781331596.0, "step": 20574 }, { "epoch": 2.6173514819997457, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7150259017944336, "learning_rate": 1e-05, "loss": 0.3672, "mean_token_accuracy": 0.9021867513656616, "num_tokens": 781365599.0, "step": 20575 }, { "epoch": 2.617478692278336, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6404621601104736, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8780187368392944, "num_tokens": 781406169.0, "step": 20576 }, { "epoch": 2.6176059025569267, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.591480255126953, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8699864149093628, "num_tokens": 781444570.0, "step": 20577 }, { "epoch": 2.6177331128355172, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5997445583343506, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8678332567214966, "num_tokens": 781482752.0, "step": 20578 }, { "epoch": 2.6178603231141078, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.519597053527832, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8664827942848206, "num_tokens": 781524549.0, "step": 20579 }, { "epoch": 2.6179875333926983, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5750343799591064, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.866559624671936, "num_tokens": 781560152.0, "step": 20580 }, { "epoch": 2.6181147436712884, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4433774948120117, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8781774044036865, "num_tokens": 781606399.0, "step": 20581 }, { "epoch": 2.6182419539498794, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5617876052856445, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8798643946647644, "num_tokens": 781646386.0, "step": 20582 }, { "epoch": 2.6183691642284694, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4304707050323486, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8731313943862915, "num_tokens": 781685629.0, "step": 20583 }, { "epoch": 2.6184963745070604, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5702075958251953, "learning_rate": 1e-05, "loss": 0.371, "mean_token_accuracy": 0.8958362936973572, "num_tokens": 781717198.0, "step": 20584 }, { "epoch": 2.6186235847856505, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.397514581680298, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8665918111801147, "num_tokens": 781749281.0, "step": 20585 }, { "epoch": 2.6187507950642415, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5682244300842285, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8828050494194031, "num_tokens": 781791544.0, "step": 20586 }, { "epoch": 2.6188780053428315, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3438313007354736, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8728111386299133, "num_tokens": 781832864.0, "step": 20587 }, { "epoch": 2.619005215621422, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.623882532119751, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8787634372711182, "num_tokens": 781875757.0, "step": 20588 }, { "epoch": 2.6191324259000126, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2950048446655273, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8818613290786743, "num_tokens": 781911200.0, "step": 20589 }, { "epoch": 2.619259636178603, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.725320816040039, "learning_rate": 1e-05, "loss": 0.4001, "mean_token_accuracy": 0.8917807340621948, "num_tokens": 781948372.0, "step": 20590 }, { "epoch": 2.6193868464571937, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.226823091506958, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8765662908554077, "num_tokens": 781985360.0, "step": 20591 }, { "epoch": 2.619514056735784, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.000339508056640625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8901736736297607, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8831373453140259, "num_tokens": 782019439.0, "step": 20592 }, { "epoch": 2.6196412670143747, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.149473190307617, "learning_rate": 1e-05, "loss": 0.3455, "mean_token_accuracy": 0.9047914743423462, "num_tokens": 782055805.0, "step": 20593 }, { "epoch": 2.6197684772929652, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.166934013366699, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8747919201850891, "num_tokens": 782092879.0, "step": 20594 }, { "epoch": 2.6198956875715558, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.2164599895477295, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8753830194473267, "num_tokens": 782133123.0, "step": 20595 }, { "epoch": 2.6200228978501463, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 4.462924480438232, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8664214611053467, "num_tokens": 782169999.0, "step": 20596 }, { "epoch": 2.620150108128737, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.5275611877441406, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8832715749740601, "num_tokens": 782208269.0, "step": 20597 }, { "epoch": 2.6202773184073274, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 7.2479248046875e-05, "grad_norm": 4.530134201049805, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.875497579574585, "num_tokens": 782242927.0, "step": 20598 }, { "epoch": 2.620404528685918, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.887883186340332, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8760757446289062, "num_tokens": 782281018.0, "step": 20599 }, { "epoch": 2.6205317389645084, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.301401138305664, "learning_rate": 1e-05, "loss": 0.538, "mean_token_accuracy": 0.8535954356193542, "num_tokens": 782315336.0, "step": 20600 }, { "epoch": 2.620658949243099, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.9103496074676514, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8926556706428528, "num_tokens": 782351163.0, "step": 20601 }, { "epoch": 2.6207861595216895, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.060685157775879, "learning_rate": 1e-05, "loss": 0.4995, "mean_token_accuracy": 0.8622426986694336, "num_tokens": 782396955.0, "step": 20602 }, { "epoch": 2.62091336980028, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.8052828311920166, "learning_rate": 1e-05, "loss": 0.4054, "mean_token_accuracy": 0.8920581936836243, "num_tokens": 782432005.0, "step": 20603 }, { "epoch": 2.6210405800788705, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8979716300964355, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.883267879486084, "num_tokens": 782467659.0, "step": 20604 }, { "epoch": 2.621167790357461, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.695988416671753, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8781948089599609, "num_tokens": 782502621.0, "step": 20605 }, { "epoch": 2.621295000636051, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7956631183624268, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8864403367042542, "num_tokens": 782535803.0, "step": 20606 }, { "epoch": 2.621422210914642, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5633604526519775, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8778742551803589, "num_tokens": 782574720.0, "step": 20607 }, { "epoch": 2.621549421193232, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.760793685913086, "learning_rate": 1e-05, "loss": 0.4156, "mean_token_accuracy": 0.8835338354110718, "num_tokens": 782615477.0, "step": 20608 }, { "epoch": 2.621676631471823, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.459601879119873, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8709813952445984, "num_tokens": 782653153.0, "step": 20609 }, { "epoch": 2.6218038417504133, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.811011791229248, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8694168329238892, "num_tokens": 782691024.0, "step": 20610 }, { "epoch": 2.621931052029004, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3627045154571533, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8796775341033936, "num_tokens": 782732031.0, "step": 20611 }, { "epoch": 2.6220582623075943, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.911482572555542, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8762527704238892, "num_tokens": 782773430.0, "step": 20612 }, { "epoch": 2.622185472586185, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.332437515258789, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8735740184783936, "num_tokens": 782815852.0, "step": 20613 }, { "epoch": 2.6223126828647754, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.05361270904541, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8869437575340271, "num_tokens": 782851327.0, "step": 20614 }, { "epoch": 2.622439893143366, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.355714797973633, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8677445650100708, "num_tokens": 782887382.0, "step": 20615 }, { "epoch": 2.6225671034219564, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.189723968505859, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8895982503890991, "num_tokens": 782923667.0, "step": 20616 }, { "epoch": 2.622694313700547, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4920847415924072, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8851938247680664, "num_tokens": 782961200.0, "step": 20617 }, { "epoch": 2.6228215239791375, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.205423355102539, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8782019019126892, "num_tokens": 782995832.0, "step": 20618 }, { "epoch": 2.622948734257728, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6452345848083496, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8790143132209778, "num_tokens": 783040880.0, "step": 20619 }, { "epoch": 2.6230759445363185, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0535101890563965, "learning_rate": 1e-05, "loss": 0.4121, "mean_token_accuracy": 0.8846861124038696, "num_tokens": 783075068.0, "step": 20620 }, { "epoch": 2.623203154814909, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7221262454986572, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8817394971847534, "num_tokens": 783109321.0, "step": 20621 }, { "epoch": 2.6233303650934996, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.829378843307495, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8836402893066406, "num_tokens": 783145222.0, "step": 20622 }, { "epoch": 2.62345757537209, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.714662551879883, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8715828657150269, "num_tokens": 783185017.0, "step": 20623 }, { "epoch": 2.6235847856506807, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.617257595062256, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.873296856880188, "num_tokens": 783228137.0, "step": 20624 }, { "epoch": 2.623711995929271, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6917672157287598, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.87651526927948, "num_tokens": 783266525.0, "step": 20625 }, { "epoch": 2.6238392062078617, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.483232021331787, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.875063419342041, "num_tokens": 783301090.0, "step": 20626 }, { "epoch": 2.6239664164864522, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.698573350906372, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8804565072059631, "num_tokens": 783336816.0, "step": 20627 }, { "epoch": 2.6240936267650428, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.36905574798584, "learning_rate": 1e-05, "loss": 0.3976, "mean_token_accuracy": 0.8909541964530945, "num_tokens": 783376766.0, "step": 20628 }, { "epoch": 2.6242208370436333, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.737689733505249, "learning_rate": 1e-05, "loss": 0.3986, "mean_token_accuracy": 0.890025794506073, "num_tokens": 783416780.0, "step": 20629 }, { "epoch": 2.624348047322224, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2889277935028076, "learning_rate": 1e-05, "loss": 0.3921, "mean_token_accuracy": 0.8912644386291504, "num_tokens": 783449957.0, "step": 20630 }, { "epoch": 2.624475257600814, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.865009307861328, "learning_rate": 1e-05, "loss": 0.3798, "mean_token_accuracy": 0.8957474231719971, "num_tokens": 783480488.0, "step": 20631 }, { "epoch": 2.624602467879405, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.234036445617676, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.875719428062439, "num_tokens": 783521296.0, "step": 20632 }, { "epoch": 2.624729678157995, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.032135009765625, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8698596954345703, "num_tokens": 783561937.0, "step": 20633 }, { "epoch": 2.624856888436586, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.267988681793213, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8854198455810547, "num_tokens": 783598928.0, "step": 20634 }, { "epoch": 2.624984098715176, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.207794666290283, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8843225240707397, "num_tokens": 783635538.0, "step": 20635 }, { "epoch": 2.6251113089937665, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.405299425125122, "learning_rate": 1e-05, "loss": 0.3965, "mean_token_accuracy": 0.8905086517333984, "num_tokens": 783675712.0, "step": 20636 }, { "epoch": 2.625238519272357, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.302642345428467, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8786002397537231, "num_tokens": 783715091.0, "step": 20637 }, { "epoch": 2.6253657295509476, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6238601207733154, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8886862993240356, "num_tokens": 783753263.0, "step": 20638 }, { "epoch": 2.625492939829538, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.206207275390625, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8728150129318237, "num_tokens": 783786251.0, "step": 20639 }, { "epoch": 2.6256201501081287, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7336881160736084, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8793778419494629, "num_tokens": 783829751.0, "step": 20640 }, { "epoch": 2.625747360386719, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.018984794616699, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8703888058662415, "num_tokens": 783867786.0, "step": 20641 }, { "epoch": 2.6258745706653097, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.721585512161255, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8715653419494629, "num_tokens": 783904245.0, "step": 20642 }, { "epoch": 2.6260017809439002, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8259382247924805, "learning_rate": 1e-05, "loss": 0.3976, "mean_token_accuracy": 0.8907274603843689, "num_tokens": 783946892.0, "step": 20643 }, { "epoch": 2.6261289912224908, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.661612033843994, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8672163486480713, "num_tokens": 783988132.0, "step": 20644 }, { "epoch": 2.6262562015010813, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7080562114715576, "learning_rate": 1e-05, "loss": 0.488, "mean_token_accuracy": 0.869123637676239, "num_tokens": 784024321.0, "step": 20645 }, { "epoch": 2.626383411779672, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5959110260009766, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8725563287734985, "num_tokens": 784061266.0, "step": 20646 }, { "epoch": 2.6265106220582624, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.616757869720459, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8753623962402344, "num_tokens": 784101991.0, "step": 20647 }, { "epoch": 2.626637832336853, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5308034420013428, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8793209791183472, "num_tokens": 784137726.0, "step": 20648 }, { "epoch": 2.6267650426154434, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5774197578430176, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8742341995239258, "num_tokens": 784177773.0, "step": 20649 }, { "epoch": 2.626892252894034, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.473886489868164, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8777441382408142, "num_tokens": 784217429.0, "step": 20650 }, { "epoch": 2.6270194631726245, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5854830741882324, "learning_rate": 1e-05, "loss": 0.3962, "mean_token_accuracy": 0.8906410932540894, "num_tokens": 784252289.0, "step": 20651 }, { "epoch": 2.627146673451215, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4206039905548096, "learning_rate": 1e-05, "loss": 0.5067, "mean_token_accuracy": 0.8548139333724976, "num_tokens": 784293472.0, "step": 20652 }, { "epoch": 2.6272738837298055, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.613116502761841, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8750975131988525, "num_tokens": 784339423.0, "step": 20653 }, { "epoch": 2.6274010940083956, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.36617374420166, "learning_rate": 1e-05, "loss": 0.3805, "mean_token_accuracy": 0.8903541564941406, "num_tokens": 784367026.0, "step": 20654 }, { "epoch": 2.6275283042869866, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.674609422683716, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8790105581283569, "num_tokens": 784404004.0, "step": 20655 }, { "epoch": 2.6276555145655767, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2884292602539062, "learning_rate": 1e-05, "loss": 0.3862, "mean_token_accuracy": 0.8917343616485596, "num_tokens": 784439335.0, "step": 20656 }, { "epoch": 2.6277827248441676, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7954063415527344, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8724834322929382, "num_tokens": 784480141.0, "step": 20657 }, { "epoch": 2.6279099351227577, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.233536720275879, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8823156356811523, "num_tokens": 784514785.0, "step": 20658 }, { "epoch": 2.6280371454013487, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.985734701156616, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8695629239082336, "num_tokens": 784549776.0, "step": 20659 }, { "epoch": 2.628164355679939, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.255796194076538, "learning_rate": 1e-05, "loss": 0.4933, "mean_token_accuracy": 0.8635821342468262, "num_tokens": 784578558.0, "step": 20660 }, { "epoch": 2.6282915659585293, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1659393310546875, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8831062912940979, "num_tokens": 784614056.0, "step": 20661 }, { "epoch": 2.62841877623712, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.328617572784424, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8804177045822144, "num_tokens": 784651663.0, "step": 20662 }, { "epoch": 2.6285459865157104, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.31802225112915, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8707589507102966, "num_tokens": 784690343.0, "step": 20663 }, { "epoch": 2.628673196794301, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.555882692337036, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8805081248283386, "num_tokens": 784728569.0, "step": 20664 }, { "epoch": 2.6288004070728914, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.276844024658203, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8879821300506592, "num_tokens": 784762376.0, "step": 20665 }, { "epoch": 2.628927617351482, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.73496675491333, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8798103332519531, "num_tokens": 784793154.0, "step": 20666 }, { "epoch": 2.6290548276300725, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.0902628898620605, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8811894655227661, "num_tokens": 784829640.0, "step": 20667 }, { "epoch": 2.629182037908663, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.740957736968994, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.888563871383667, "num_tokens": 784867045.0, "step": 20668 }, { "epoch": 2.6293092481872535, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.895815134048462, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8809975385665894, "num_tokens": 784907157.0, "step": 20669 }, { "epoch": 2.629436458465844, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.694835901260376, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.881854236125946, "num_tokens": 784942247.0, "step": 20670 }, { "epoch": 2.6295636687444346, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.762500524520874, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8767036199569702, "num_tokens": 784976910.0, "step": 20671 }, { "epoch": 2.629690879023025, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6097195148468018, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.861750066280365, "num_tokens": 785017969.0, "step": 20672 }, { "epoch": 2.6298180893016156, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6655774116516113, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.883036732673645, "num_tokens": 785056182.0, "step": 20673 }, { "epoch": 2.629945299580206, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.534003257751465, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8877926468849182, "num_tokens": 785091996.0, "step": 20674 }, { "epoch": 2.6300725098587967, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.626121997833252, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8852077126502991, "num_tokens": 785129268.0, "step": 20675 }, { "epoch": 2.6301997201373872, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.454868793487549, "learning_rate": 1e-05, "loss": 0.3949, "mean_token_accuracy": 0.8897631764411926, "num_tokens": 785168418.0, "step": 20676 }, { "epoch": 2.6303269304159778, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6243767738342285, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8846789598464966, "num_tokens": 785209981.0, "step": 20677 }, { "epoch": 2.6304541406945683, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3859455585479736, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8822234272956848, "num_tokens": 785250354.0, "step": 20678 }, { "epoch": 2.6305813509731584, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.689645528793335, "learning_rate": 1e-05, "loss": 0.4363, "mean_token_accuracy": 0.8782743215560913, "num_tokens": 785288594.0, "step": 20679 }, { "epoch": 2.6307085612517493, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.317578077316284, "learning_rate": 1e-05, "loss": 0.3818, "mean_token_accuracy": 0.8932055234909058, "num_tokens": 785323885.0, "step": 20680 }, { "epoch": 2.6308357715303394, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7829372882843018, "learning_rate": 1e-05, "loss": 0.397, "mean_token_accuracy": 0.893108606338501, "num_tokens": 785361950.0, "step": 20681 }, { "epoch": 2.6309629818089304, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2584388256073, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.872017502784729, "num_tokens": 785398650.0, "step": 20682 }, { "epoch": 2.6310901920875205, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 4.054911136627197, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8722692728042603, "num_tokens": 785451273.0, "step": 20683 }, { "epoch": 2.6312174023661115, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2413201332092285, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8751843571662903, "num_tokens": 785491244.0, "step": 20684 }, { "epoch": 2.6313446126447015, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.208752632141113, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8797038197517395, "num_tokens": 785536558.0, "step": 20685 }, { "epoch": 2.631471822923292, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3793444633483887, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8747455477714539, "num_tokens": 785572635.0, "step": 20686 }, { "epoch": 2.6315990332018826, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.333523273468018, "learning_rate": 1e-05, "loss": 0.4626, "mean_token_accuracy": 0.876020073890686, "num_tokens": 785615765.0, "step": 20687 }, { "epoch": 2.631726243480473, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6237170696258545, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8796454668045044, "num_tokens": 785652934.0, "step": 20688 }, { "epoch": 2.6318534537590637, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.238411903381348, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8743717670440674, "num_tokens": 785689018.0, "step": 20689 }, { "epoch": 2.631980664037654, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.753603935241699, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8716944456100464, "num_tokens": 785728257.0, "step": 20690 }, { "epoch": 2.6321078743162447, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.031160354614258, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8708062767982483, "num_tokens": 785768381.0, "step": 20691 }, { "epoch": 2.6322350845948352, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7438371181488037, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.876318097114563, "num_tokens": 785806700.0, "step": 20692 }, { "epoch": 2.6323622948734258, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.839374542236328, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8645312190055847, "num_tokens": 785848602.0, "step": 20693 }, { "epoch": 2.6324895051520163, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.680220365524292, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8803448677062988, "num_tokens": 785888714.0, "step": 20694 }, { "epoch": 2.632616715430607, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.695084571838379, "learning_rate": 1e-05, "loss": 0.4008, "mean_token_accuracy": 0.889576256275177, "num_tokens": 785925163.0, "step": 20695 }, { "epoch": 2.6327439257091974, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.620624542236328, "learning_rate": 1e-05, "loss": 0.4778, "mean_token_accuracy": 0.861392080783844, "num_tokens": 785963566.0, "step": 20696 }, { "epoch": 2.632871135987788, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.612156629562378, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8687459230422974, "num_tokens": 785997654.0, "step": 20697 }, { "epoch": 2.6329983462663784, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.545624017715454, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8751848936080933, "num_tokens": 786038761.0, "step": 20698 }, { "epoch": 2.633125556544969, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.569471597671509, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.88216233253479, "num_tokens": 786072380.0, "step": 20699 }, { "epoch": 2.6332527668235595, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4895403385162354, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8783696293830872, "num_tokens": 786112751.0, "step": 20700 }, { "epoch": 2.63337997710215, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5552148818969727, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8785896301269531, "num_tokens": 786156841.0, "step": 20701 }, { "epoch": 2.6335071873807405, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4509782791137695, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8723119497299194, "num_tokens": 786191015.0, "step": 20702 }, { "epoch": 2.633634397659331, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.56514310836792, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8612120151519775, "num_tokens": 786231420.0, "step": 20703 }, { "epoch": 2.633761607937921, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.393648624420166, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.885454535484314, "num_tokens": 786266586.0, "step": 20704 }, { "epoch": 2.633888818216512, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5988917350769043, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8770794868469238, "num_tokens": 786307489.0, "step": 20705 }, { "epoch": 2.634016028495102, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3374786376953125, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8839824795722961, "num_tokens": 786340960.0, "step": 20706 }, { "epoch": 2.634143238773693, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.663217067718506, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8755030632019043, "num_tokens": 786383086.0, "step": 20707 }, { "epoch": 2.6342704490522832, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.262025833129883, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.885821521282196, "num_tokens": 786421093.0, "step": 20708 }, { "epoch": 2.6343976593308738, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7945334911346436, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8765904903411865, "num_tokens": 786462712.0, "step": 20709 }, { "epoch": 2.6345248696094643, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2074978351593018, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8753782510757446, "num_tokens": 786501607.0, "step": 20710 }, { "epoch": 2.634652079888055, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.993785858154297, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8804619312286377, "num_tokens": 786537163.0, "step": 20711 }, { "epoch": 2.6347792901666454, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.1911001205444336, "learning_rate": 1e-05, "loss": 0.3948, "mean_token_accuracy": 0.888192355632782, "num_tokens": 786572880.0, "step": 20712 }, { "epoch": 2.634906500445236, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2448883056640625, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8798218965530396, "num_tokens": 786610082.0, "step": 20713 }, { "epoch": 2.6350337107238264, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6599464416503906, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8763139843940735, "num_tokens": 786648875.0, "step": 20714 }, { "epoch": 2.635160921002417, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.103171348571777, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8737220168113708, "num_tokens": 786688076.0, "step": 20715 }, { "epoch": 2.6352881312810075, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4359898567199707, "learning_rate": 1e-05, "loss": 0.4429, "mean_token_accuracy": 0.8754558563232422, "num_tokens": 786723969.0, "step": 20716 }, { "epoch": 2.635415341559598, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.106963157653809, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8727120161056519, "num_tokens": 786764018.0, "step": 20717 }, { "epoch": 2.6355425518381885, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.497316360473633, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8846670985221863, "num_tokens": 786796586.0, "step": 20718 }, { "epoch": 2.635669762116779, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.094911575317383, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.865827202796936, "num_tokens": 786833797.0, "step": 20719 }, { "epoch": 2.6357969723953696, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5660042762756348, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8684508204460144, "num_tokens": 786869971.0, "step": 20720 }, { "epoch": 2.63592418267396, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.021849155426025, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.881318211555481, "num_tokens": 786905854.0, "step": 20721 }, { "epoch": 2.6360513929525506, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.5772953033447266, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8791128396987915, "num_tokens": 786940535.0, "step": 20722 }, { "epoch": 2.636178603231141, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9320385456085205, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.868135929107666, "num_tokens": 786974976.0, "step": 20723 }, { "epoch": 2.6363058135097317, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.5658915042877197, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8763436675071716, "num_tokens": 787013389.0, "step": 20724 }, { "epoch": 2.6364330237883222, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8394925594329834, "learning_rate": 1e-05, "loss": 0.3929, "mean_token_accuracy": 0.8938383460044861, "num_tokens": 787044463.0, "step": 20725 }, { "epoch": 2.6365602340669128, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5504579544067383, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8806142807006836, "num_tokens": 787075123.0, "step": 20726 }, { "epoch": 2.636687444345503, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.7601170539855957, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8739250898361206, "num_tokens": 787114790.0, "step": 20727 }, { "epoch": 2.636814654624094, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.494741439819336, "learning_rate": 1e-05, "loss": 0.3981, "mean_token_accuracy": 0.8896200060844421, "num_tokens": 787153263.0, "step": 20728 }, { "epoch": 2.636941864902684, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.716064453125, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.8873831033706665, "num_tokens": 787191955.0, "step": 20729 }, { "epoch": 2.637069075181275, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.459984064102173, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8803846836090088, "num_tokens": 787228330.0, "step": 20730 }, { "epoch": 2.637196285459865, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7075881958007812, "learning_rate": 1e-05, "loss": 0.4983, "mean_token_accuracy": 0.8574469685554504, "num_tokens": 787271927.0, "step": 20731 }, { "epoch": 2.637323495738456, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4409823417663574, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8900715112686157, "num_tokens": 787310515.0, "step": 20732 }, { "epoch": 2.637450706017046, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7114405632019043, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8777990937232971, "num_tokens": 787346720.0, "step": 20733 }, { "epoch": 2.6375779162956365, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3902435302734375, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8747960925102234, "num_tokens": 787387591.0, "step": 20734 }, { "epoch": 2.637705126574227, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7383055686950684, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8761870861053467, "num_tokens": 787427762.0, "step": 20735 }, { "epoch": 2.6378323368528176, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3603971004486084, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8682249784469604, "num_tokens": 787463047.0, "step": 20736 }, { "epoch": 2.637959547131408, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7967729568481445, "learning_rate": 1e-05, "loss": 0.398, "mean_token_accuracy": 0.8910370469093323, "num_tokens": 787500862.0, "step": 20737 }, { "epoch": 2.6380867574099987, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.335163116455078, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8877186179161072, "num_tokens": 787529625.0, "step": 20738 }, { "epoch": 2.638213967688589, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.86053729057312, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8826417922973633, "num_tokens": 787567248.0, "step": 20739 }, { "epoch": 2.6383411779671797, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3276658058166504, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8758558630943298, "num_tokens": 787606763.0, "step": 20740 }, { "epoch": 2.6384683882457702, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.944155693054199, "learning_rate": 1e-05, "loss": 0.4427, "mean_token_accuracy": 0.8761041164398193, "num_tokens": 787638795.0, "step": 20741 }, { "epoch": 2.6385955985243608, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.3406643867492676, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8723751902580261, "num_tokens": 787679495.0, "step": 20742 }, { "epoch": 2.6387228088029513, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003414154052734375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0096755027771, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8802275657653809, "num_tokens": 787718583.0, "step": 20743 }, { "epoch": 2.638850019081542, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.4084060192108154, "learning_rate": 1e-05, "loss": 0.3806, "mean_token_accuracy": 0.8960691094398499, "num_tokens": 787756558.0, "step": 20744 }, { "epoch": 2.6389772293601323, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.016762733459473, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8811262845993042, "num_tokens": 787789796.0, "step": 20745 }, { "epoch": 2.639104439638723, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.4966256618499756, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.867063045501709, "num_tokens": 787832054.0, "step": 20746 }, { "epoch": 2.6392316499173134, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9207661151885986, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8782274723052979, "num_tokens": 787876723.0, "step": 20747 }, { "epoch": 2.639358860195904, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.5526890754699707, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8675550222396851, "num_tokens": 787912183.0, "step": 20748 }, { "epoch": 2.6394860704744945, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.792921304702759, "learning_rate": 1e-05, "loss": 0.4803, "mean_token_accuracy": 0.8668767809867859, "num_tokens": 787946865.0, "step": 20749 }, { "epoch": 2.639613280753085, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.5618577003479004, "learning_rate": 1e-05, "loss": 0.4123, "mean_token_accuracy": 0.8870254158973694, "num_tokens": 787984286.0, "step": 20750 }, { "epoch": 2.6397404910316755, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.642573833465576, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8911079168319702, "num_tokens": 788020565.0, "step": 20751 }, { "epoch": 2.6398677013102656, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5842695236206055, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8876423835754395, "num_tokens": 788051833.0, "step": 20752 }, { "epoch": 2.6399949115888566, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5010883808135986, "learning_rate": 1e-05, "loss": 0.3907, "mean_token_accuracy": 0.8907056450843811, "num_tokens": 788092780.0, "step": 20753 }, { "epoch": 2.6401221218674467, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5890696048736572, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.871680736541748, "num_tokens": 788140088.0, "step": 20754 }, { "epoch": 2.6402493321460376, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4089767932891846, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8834320306777954, "num_tokens": 788177951.0, "step": 20755 }, { "epoch": 2.6403765424246277, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6412854194641113, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8851327300071716, "num_tokens": 788212124.0, "step": 20756 }, { "epoch": 2.6405037527032187, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.331566572189331, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8767159581184387, "num_tokens": 788255860.0, "step": 20757 }, { "epoch": 2.6406309629818088, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6956214904785156, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.88818359375, "num_tokens": 788296481.0, "step": 20758 }, { "epoch": 2.6407581732603993, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2769081592559814, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8900729417800903, "num_tokens": 788328818.0, "step": 20759 }, { "epoch": 2.64088538353899, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8087334632873535, "learning_rate": 1e-05, "loss": 0.3834, "mean_token_accuracy": 0.8965718746185303, "num_tokens": 788370359.0, "step": 20760 }, { "epoch": 2.6410125938175804, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2060956954956055, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.878669023513794, "num_tokens": 788412425.0, "step": 20761 }, { "epoch": 2.641139804096171, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.01119327545166, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8639844059944153, "num_tokens": 788445137.0, "step": 20762 }, { "epoch": 2.6412670143747614, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2200558185577393, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8870413303375244, "num_tokens": 788478185.0, "step": 20763 }, { "epoch": 2.641394224653352, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.22095251083374, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8777484893798828, "num_tokens": 788514329.0, "step": 20764 }, { "epoch": 2.6415214349319425, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.3387391567230225, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8817651867866516, "num_tokens": 788554234.0, "step": 20765 }, { "epoch": 2.641648645210533, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.373171329498291, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8828513026237488, "num_tokens": 788591684.0, "step": 20766 }, { "epoch": 2.6417758554891235, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6169326305389404, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8817145824432373, "num_tokens": 788626802.0, "step": 20767 }, { "epoch": 2.641903065767714, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.299015045166016, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8837272524833679, "num_tokens": 788673826.0, "step": 20768 }, { "epoch": 2.6420302760463046, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.7574141025543213, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8881893754005432, "num_tokens": 788708393.0, "step": 20769 }, { "epoch": 2.642157486324895, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.121558666229248, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8867196440696716, "num_tokens": 788746480.0, "step": 20770 }, { "epoch": 2.6422846966034856, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.755950450897217, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8778859972953796, "num_tokens": 788785295.0, "step": 20771 }, { "epoch": 2.642411906882076, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9473087787628174, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.8842852115631104, "num_tokens": 788818803.0, "step": 20772 }, { "epoch": 2.6425391171606667, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.691725730895996, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8759802579879761, "num_tokens": 788857266.0, "step": 20773 }, { "epoch": 2.6426663274392572, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.7997360229492188, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8609444499015808, "num_tokens": 788898433.0, "step": 20774 }, { "epoch": 2.6427935377178478, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.609314441680908, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8731423616409302, "num_tokens": 788942259.0, "step": 20775 }, { "epoch": 2.6429207479964383, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.714057683944702, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8699122071266174, "num_tokens": 788981484.0, "step": 20776 }, { "epoch": 2.6430479582750284, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.524350643157959, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8811032176017761, "num_tokens": 789019863.0, "step": 20777 }, { "epoch": 2.6431751685536193, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6843621730804443, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8808073997497559, "num_tokens": 789059139.0, "step": 20778 }, { "epoch": 2.6433023788322094, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4443585872650146, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8830925822257996, "num_tokens": 789100429.0, "step": 20779 }, { "epoch": 2.6434295891108004, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.703888416290283, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8828397989273071, "num_tokens": 789136219.0, "step": 20780 }, { "epoch": 2.6435567993893905, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.368809700012207, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8829024434089661, "num_tokens": 789177541.0, "step": 20781 }, { "epoch": 2.6436840096679814, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7701621055603027, "learning_rate": 1e-05, "loss": 0.3865, "mean_token_accuracy": 0.8951835632324219, "num_tokens": 789217165.0, "step": 20782 }, { "epoch": 2.6438112199465715, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.305788040161133, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8830888271331787, "num_tokens": 789255384.0, "step": 20783 }, { "epoch": 2.643938430225162, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.905461072921753, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8833234906196594, "num_tokens": 789294153.0, "step": 20784 }, { "epoch": 2.6440656405037526, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2842884063720703, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8617719411849976, "num_tokens": 789334933.0, "step": 20785 }, { "epoch": 2.644192850782343, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.0754780769348145, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8589380979537964, "num_tokens": 789370068.0, "step": 20786 }, { "epoch": 2.6443200610609336, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3205654621124268, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8847468495368958, "num_tokens": 789413887.0, "step": 20787 }, { "epoch": 2.644447271339524, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.210143089294434, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8889933824539185, "num_tokens": 789452319.0, "step": 20788 }, { "epoch": 2.6445744816181147, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.4987316131591797, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8880525827407837, "num_tokens": 789488988.0, "step": 20789 }, { "epoch": 2.6447016918967052, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1915764808654785, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8730394840240479, "num_tokens": 789529849.0, "step": 20790 }, { "epoch": 2.6448289021752958, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.6960299015045166, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8769558668136597, "num_tokens": 789572877.0, "step": 20791 }, { "epoch": 2.6449561124538863, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.004701137542725, "learning_rate": 1e-05, "loss": 0.4731, "mean_token_accuracy": 0.8684806823730469, "num_tokens": 789605189.0, "step": 20792 }, { "epoch": 2.645083322732477, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.699347496032715, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8857812285423279, "num_tokens": 789641589.0, "step": 20793 }, { "epoch": 2.6452105330110673, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7991552352905273, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.869280219078064, "num_tokens": 789673869.0, "step": 20794 }, { "epoch": 2.645337743289658, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.668825626373291, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.879265546798706, "num_tokens": 789710605.0, "step": 20795 }, { "epoch": 2.6454649535682484, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6231725215911865, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8875759840011597, "num_tokens": 789747641.0, "step": 20796 }, { "epoch": 2.645592163846839, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6534595489501953, "learning_rate": 1e-05, "loss": 0.3846, "mean_token_accuracy": 0.8936561942100525, "num_tokens": 789781345.0, "step": 20797 }, { "epoch": 2.6457193741254295, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4977478981018066, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8771393299102783, "num_tokens": 789819582.0, "step": 20798 }, { "epoch": 2.64584658440402, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.636604070663452, "learning_rate": 1e-05, "loss": 0.3921, "mean_token_accuracy": 0.8890392780303955, "num_tokens": 789856281.0, "step": 20799 }, { "epoch": 2.6459737946826105, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4123663902282715, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8885430693626404, "num_tokens": 789895711.0, "step": 20800 }, { "epoch": 2.646101004961201, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6587769985198975, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8931471109390259, "num_tokens": 789931751.0, "step": 20801 }, { "epoch": 2.646228215239791, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.35456919670105, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8732292652130127, "num_tokens": 789969207.0, "step": 20802 }, { "epoch": 2.646355425518382, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7088754177093506, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.876454770565033, "num_tokens": 790005706.0, "step": 20803 }, { "epoch": 2.646482635796972, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.292426109313965, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8770477771759033, "num_tokens": 790047389.0, "step": 20804 }, { "epoch": 2.646609846075563, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8161799907684326, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.881795346736908, "num_tokens": 790085430.0, "step": 20805 }, { "epoch": 2.6467370563541532, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2267141342163086, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8754346370697021, "num_tokens": 790126620.0, "step": 20806 }, { "epoch": 2.6468642666327438, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.987298011779785, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8832406997680664, "num_tokens": 790166033.0, "step": 20807 }, { "epoch": 2.6469914769113343, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2258806228637695, "learning_rate": 1e-05, "loss": 0.3959, "mean_token_accuracy": 0.8889074921607971, "num_tokens": 790202463.0, "step": 20808 }, { "epoch": 2.647118687189925, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.198758602142334, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8839980363845825, "num_tokens": 790241600.0, "step": 20809 }, { "epoch": 2.6472458974685154, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.3652920722961426, "learning_rate": 1e-05, "loss": 0.5019, "mean_token_accuracy": 0.8574471473693848, "num_tokens": 790277669.0, "step": 20810 }, { "epoch": 2.647373107747106, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.318365097045898, "learning_rate": 1e-05, "loss": 0.4679, "mean_token_accuracy": 0.8713762164115906, "num_tokens": 790317453.0, "step": 20811 }, { "epoch": 2.6475003180256964, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.571300745010376, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8784698247909546, "num_tokens": 790354460.0, "step": 20812 }, { "epoch": 2.647627528304287, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.265986442565918, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8789036273956299, "num_tokens": 790393933.0, "step": 20813 }, { "epoch": 2.6477547385828775, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.735715627670288, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8816736936569214, "num_tokens": 790425477.0, "step": 20814 }, { "epoch": 2.647881948861468, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.083733081817627, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8576009273529053, "num_tokens": 790470501.0, "step": 20815 }, { "epoch": 2.6480091591400585, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7353060245513916, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8756203651428223, "num_tokens": 790508869.0, "step": 20816 }, { "epoch": 2.648136369418649, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.8899354934692383, "learning_rate": 1e-05, "loss": 0.3911, "mean_token_accuracy": 0.8939743638038635, "num_tokens": 790545675.0, "step": 20817 }, { "epoch": 2.6482635796972396, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.673997402191162, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8828516602516174, "num_tokens": 790586832.0, "step": 20818 }, { "epoch": 2.64839078997583, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7491307258605957, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8758736252784729, "num_tokens": 790633587.0, "step": 20819 }, { "epoch": 2.6485180002544206, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.616929054260254, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8693444132804871, "num_tokens": 790668768.0, "step": 20820 }, { "epoch": 2.648645210533011, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.648695468902588, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8888998031616211, "num_tokens": 790708393.0, "step": 20821 }, { "epoch": 2.6487724208116017, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5540499687194824, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8687598705291748, "num_tokens": 790742632.0, "step": 20822 }, { "epoch": 2.648899631090192, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6190052032470703, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8657772541046143, "num_tokens": 790779740.0, "step": 20823 }, { "epoch": 2.6490268413687827, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4680330753326416, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8874586820602417, "num_tokens": 790819040.0, "step": 20824 }, { "epoch": 2.649154051647373, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6180522441864014, "learning_rate": 1e-05, "loss": 0.3926, "mean_token_accuracy": 0.8930601477622986, "num_tokens": 790853382.0, "step": 20825 }, { "epoch": 2.649281261925964, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.409661293029785, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.870871901512146, "num_tokens": 790890606.0, "step": 20826 }, { "epoch": 2.649408472204554, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.652392864227295, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8857853412628174, "num_tokens": 790929469.0, "step": 20827 }, { "epoch": 2.649535682483145, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3343305587768555, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8776909112930298, "num_tokens": 790970745.0, "step": 20828 }, { "epoch": 2.649662892761735, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7494423389434814, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.869957447052002, "num_tokens": 791010365.0, "step": 20829 }, { "epoch": 2.649790103040326, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.266727924346924, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.8810701966285706, "num_tokens": 791047012.0, "step": 20830 }, { "epoch": 2.649917313318916, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.883091688156128, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8825646638870239, "num_tokens": 791092836.0, "step": 20831 }, { "epoch": 2.6500445235975065, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2295241355895996, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8898528218269348, "num_tokens": 791127725.0, "step": 20832 }, { "epoch": 2.650171733876097, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.090941905975342, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8871910572052002, "num_tokens": 791164530.0, "step": 20833 }, { "epoch": 2.6502989441546876, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.268127202987671, "learning_rate": 1e-05, "loss": 0.382, "mean_token_accuracy": 0.8945111036300659, "num_tokens": 791202384.0, "step": 20834 }, { "epoch": 2.650426154433278, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.289279937744141, "learning_rate": 1e-05, "loss": 0.4009, "mean_token_accuracy": 0.8937987089157104, "num_tokens": 791237998.0, "step": 20835 }, { "epoch": 2.6505533647118686, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.488820791244507, "learning_rate": 1e-05, "loss": 0.4882, "mean_token_accuracy": 0.8578139543533325, "num_tokens": 791277755.0, "step": 20836 }, { "epoch": 2.650680574990459, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.309532642364502, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8684239387512207, "num_tokens": 791315779.0, "step": 20837 }, { "epoch": 2.6508077852690497, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.687199115753174, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8753538727760315, "num_tokens": 791359212.0, "step": 20838 }, { "epoch": 2.6509349955476402, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1502766609191895, "learning_rate": 1e-05, "loss": 0.3875, "mean_token_accuracy": 0.8964131474494934, "num_tokens": 791388308.0, "step": 20839 }, { "epoch": 2.6510622058262308, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7587127685546875, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8761480450630188, "num_tokens": 791425645.0, "step": 20840 }, { "epoch": 2.6511894161048213, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.940103530883789, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8784151077270508, "num_tokens": 791462632.0, "step": 20841 }, { "epoch": 2.651316626383412, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6996071338653564, "learning_rate": 1e-05, "loss": 0.3806, "mean_token_accuracy": 0.8964942693710327, "num_tokens": 791502834.0, "step": 20842 }, { "epoch": 2.6514438366620023, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7757482528686523, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8764592409133911, "num_tokens": 791538973.0, "step": 20843 }, { "epoch": 2.651571046940593, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.645625352859497, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.881111741065979, "num_tokens": 791577804.0, "step": 20844 }, { "epoch": 2.6516982572191834, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6394236087799072, "learning_rate": 1e-05, "loss": 0.3871, "mean_token_accuracy": 0.8913929462432861, "num_tokens": 791618100.0, "step": 20845 }, { "epoch": 2.651825467497774, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.583547592163086, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8753224611282349, "num_tokens": 791660112.0, "step": 20846 }, { "epoch": 2.6519526777763645, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5763773918151855, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8784946799278259, "num_tokens": 791698133.0, "step": 20847 }, { "epoch": 2.652079888054955, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5321388244628906, "learning_rate": 1e-05, "loss": 0.4153, "mean_token_accuracy": 0.8857944011688232, "num_tokens": 791737810.0, "step": 20848 }, { "epoch": 2.6522070983335455, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5460567474365234, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8750492334365845, "num_tokens": 791775187.0, "step": 20849 }, { "epoch": 2.6523343086121356, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4914045333862305, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8862795829772949, "num_tokens": 791810673.0, "step": 20850 }, { "epoch": 2.6524615188907266, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.517657995223999, "learning_rate": 1e-05, "loss": 0.3634, "mean_token_accuracy": 0.901252269744873, "num_tokens": 791848144.0, "step": 20851 }, { "epoch": 2.6525887291693167, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4553821086883545, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8794541358947754, "num_tokens": 791885025.0, "step": 20852 }, { "epoch": 2.6527159394479076, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.539846420288086, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8772943019866943, "num_tokens": 791922885.0, "step": 20853 }, { "epoch": 2.6528431497264977, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4132227897644043, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8607345819473267, "num_tokens": 791959791.0, "step": 20854 }, { "epoch": 2.6529703600050887, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5660440921783447, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8813846111297607, "num_tokens": 791998559.0, "step": 20855 }, { "epoch": 2.6530975702836788, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.362027883529663, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.879580020904541, "num_tokens": 792038428.0, "step": 20856 }, { "epoch": 2.6532247805622693, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.6365251541137695, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8675640821456909, "num_tokens": 792074405.0, "step": 20857 }, { "epoch": 2.65335199084086, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.296635150909424, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8680606484413147, "num_tokens": 792112831.0, "step": 20858 }, { "epoch": 2.6534792011194503, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7457611560821533, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8750097155570984, "num_tokens": 792154122.0, "step": 20859 }, { "epoch": 2.653606411398041, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.223846435546875, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8610243201255798, "num_tokens": 792191818.0, "step": 20860 }, { "epoch": 2.6537336216766314, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.930294990539551, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8796085715293884, "num_tokens": 792226753.0, "step": 20861 }, { "epoch": 2.653860831955222, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.169816493988037, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8793485164642334, "num_tokens": 792263941.0, "step": 20862 }, { "epoch": 2.6539880422338125, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.188146114349365, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8817762136459351, "num_tokens": 792301433.0, "step": 20863 }, { "epoch": 2.654115252512403, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.246546983718872, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8792164921760559, "num_tokens": 792340188.0, "step": 20864 }, { "epoch": 2.6542424627909935, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.456168174743652, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8731052279472351, "num_tokens": 792383498.0, "step": 20865 }, { "epoch": 2.654369673069584, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.571491241455078, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8892956972122192, "num_tokens": 792417160.0, "step": 20866 }, { "epoch": 2.6544968833481746, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.475139141082764, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8751357793807983, "num_tokens": 792451609.0, "step": 20867 }, { "epoch": 2.654624093626765, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.8469595909118652, "learning_rate": 1e-05, "loss": 0.493, "mean_token_accuracy": 0.866622805595398, "num_tokens": 792498309.0, "step": 20868 }, { "epoch": 2.6547513039053556, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.260099411010742, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8695118427276611, "num_tokens": 792538954.0, "step": 20869 }, { "epoch": 2.654878514183946, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.882068634033203, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8662497401237488, "num_tokens": 792577037.0, "step": 20870 }, { "epoch": 2.6550057244625367, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.054245948791504, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8686137199401855, "num_tokens": 792609538.0, "step": 20871 }, { "epoch": 2.655132934741127, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7772998809814453, "learning_rate": 1e-05, "loss": 0.4535, "mean_token_accuracy": 0.8795114755630493, "num_tokens": 792647793.0, "step": 20872 }, { "epoch": 2.6552601450197177, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8825910091400146, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8828496336936951, "num_tokens": 792686746.0, "step": 20873 }, { "epoch": 2.6553873552983083, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.673457145690918, "learning_rate": 1e-05, "loss": 0.4123, "mean_token_accuracy": 0.881010890007019, "num_tokens": 792716324.0, "step": 20874 }, { "epoch": 2.6555145655768984, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7699925899505615, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8841030597686768, "num_tokens": 792757242.0, "step": 20875 }, { "epoch": 2.6556417758554893, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.571272373199463, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8736464977264404, "num_tokens": 792795545.0, "step": 20876 }, { "epoch": 2.6557689861340794, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.740466356277466, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8769693374633789, "num_tokens": 792835772.0, "step": 20877 }, { "epoch": 2.6558961964126704, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.4602761268615723, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8693247437477112, "num_tokens": 792877901.0, "step": 20878 }, { "epoch": 2.6560234066912605, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.766587495803833, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8788152933120728, "num_tokens": 792918528.0, "step": 20879 }, { "epoch": 2.6561506169698514, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3606362342834473, "learning_rate": 1e-05, "loss": 0.3822, "mean_token_accuracy": 0.8921464681625366, "num_tokens": 792954251.0, "step": 20880 }, { "epoch": 2.6562778272484415, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.875253915786743, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8821409344673157, "num_tokens": 792991452.0, "step": 20881 }, { "epoch": 2.656405037527032, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3158528804779053, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8739126920700073, "num_tokens": 793024488.0, "step": 20882 }, { "epoch": 2.6565322478056226, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.01984977722168, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8865841627120972, "num_tokens": 793064070.0, "step": 20883 }, { "epoch": 2.656659458084213, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.319070339202881, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8691429495811462, "num_tokens": 793103088.0, "step": 20884 }, { "epoch": 2.6567866683628036, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.176661968231201, "learning_rate": 1e-05, "loss": 0.4949, "mean_token_accuracy": 0.8652123808860779, "num_tokens": 793146415.0, "step": 20885 }, { "epoch": 2.656913878641394, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4573974609375, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8816534280776978, "num_tokens": 793181588.0, "step": 20886 }, { "epoch": 2.6570410889199847, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.227412700653076, "learning_rate": 1e-05, "loss": 0.4851, "mean_token_accuracy": 0.8627834320068359, "num_tokens": 793217764.0, "step": 20887 }, { "epoch": 2.6571682991985752, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.626138687133789, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8845888376235962, "num_tokens": 793255945.0, "step": 20888 }, { "epoch": 2.6572955094771658, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.101004600524902, "learning_rate": 1e-05, "loss": 0.46, "mean_token_accuracy": 0.8759128451347351, "num_tokens": 793288981.0, "step": 20889 }, { "epoch": 2.6574227197557563, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.719101667404175, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8818602561950684, "num_tokens": 793327635.0, "step": 20890 }, { "epoch": 2.657549930034347, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.900338888168335, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8784553408622742, "num_tokens": 793357077.0, "step": 20891 }, { "epoch": 2.6576771403129373, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.718473196029663, "learning_rate": 1e-05, "loss": 0.3907, "mean_token_accuracy": 0.8945378065109253, "num_tokens": 793389664.0, "step": 20892 }, { "epoch": 2.657804350591528, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6700122356414795, "learning_rate": 1e-05, "loss": 0.3986, "mean_token_accuracy": 0.8922573328018188, "num_tokens": 793429951.0, "step": 20893 }, { "epoch": 2.6579315608701184, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.689398765563965, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8675928711891174, "num_tokens": 793471404.0, "step": 20894 }, { "epoch": 2.658058771148709, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.516065835952759, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8773596286773682, "num_tokens": 793510140.0, "step": 20895 }, { "epoch": 2.6581859814272994, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.691162586212158, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8792833685874939, "num_tokens": 793553927.0, "step": 20896 }, { "epoch": 2.65831319170589, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.40427565574646, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8849387168884277, "num_tokens": 793590413.0, "step": 20897 }, { "epoch": 2.6584404019844805, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.744044542312622, "learning_rate": 1e-05, "loss": 0.4929, "mean_token_accuracy": 0.8594976663589478, "num_tokens": 793626320.0, "step": 20898 }, { "epoch": 2.658567612263071, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3062331676483154, "learning_rate": 1e-05, "loss": 0.4097, "mean_token_accuracy": 0.8810049891471863, "num_tokens": 793665733.0, "step": 20899 }, { "epoch": 2.658694822541661, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8365132808685303, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8785266876220703, "num_tokens": 793705593.0, "step": 20900 }, { "epoch": 2.658822032820252, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.247312068939209, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8826667070388794, "num_tokens": 793744874.0, "step": 20901 }, { "epoch": 2.658949243098842, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.028207778930664, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.8662145137786865, "num_tokens": 793786598.0, "step": 20902 }, { "epoch": 2.659076453377433, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.24853253364563, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8790478706359863, "num_tokens": 793826548.0, "step": 20903 }, { "epoch": 2.6592036636560232, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.21290922164917, "learning_rate": 1e-05, "loss": 0.4951, "mean_token_accuracy": 0.8645947575569153, "num_tokens": 793868411.0, "step": 20904 }, { "epoch": 2.6593308739346138, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.402599334716797, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8732237815856934, "num_tokens": 793905931.0, "step": 20905 }, { "epoch": 2.6594580842132043, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.313252925872803, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8758984804153442, "num_tokens": 793939711.0, "step": 20906 }, { "epoch": 2.659585294491795, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6143956184387207, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.88215172290802, "num_tokens": 793976746.0, "step": 20907 }, { "epoch": 2.6597125047703853, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.230257034301758, "learning_rate": 1e-05, "loss": 0.4989, "mean_token_accuracy": 0.860971212387085, "num_tokens": 794014132.0, "step": 20908 }, { "epoch": 2.659839715048976, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7516932487487793, "learning_rate": 1e-05, "loss": 0.4963, "mean_token_accuracy": 0.8633576035499573, "num_tokens": 794051958.0, "step": 20909 }, { "epoch": 2.6599669253275664, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0263824462890625, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8816977739334106, "num_tokens": 794087494.0, "step": 20910 }, { "epoch": 2.660094135606157, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7383017539978027, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8664904832839966, "num_tokens": 794123156.0, "step": 20911 }, { "epoch": 2.6602213458847475, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.828592300415039, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8657993078231812, "num_tokens": 794169023.0, "step": 20912 }, { "epoch": 2.660348556163338, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6685662269592285, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.891049861907959, "num_tokens": 794202145.0, "step": 20913 }, { "epoch": 2.6604757664419285, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.696760654449463, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8916832804679871, "num_tokens": 794239791.0, "step": 20914 }, { "epoch": 2.660602976720519, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5888733863830566, "learning_rate": 1e-05, "loss": 0.4133, "mean_token_accuracy": 0.8858937621116638, "num_tokens": 794285184.0, "step": 20915 }, { "epoch": 2.6607301869991096, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6285934448242188, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.893520176410675, "num_tokens": 794324705.0, "step": 20916 }, { "epoch": 2.6608573972777, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.536010265350342, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8727107048034668, "num_tokens": 794360712.0, "step": 20917 }, { "epoch": 2.6609846075562906, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.590865135192871, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8849717974662781, "num_tokens": 794401634.0, "step": 20918 }, { "epoch": 2.661111817834881, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.483583450317383, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8821102380752563, "num_tokens": 794433460.0, "step": 20919 }, { "epoch": 2.6612390281134717, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5863499641418457, "learning_rate": 1e-05, "loss": 0.378, "mean_token_accuracy": 0.8945070505142212, "num_tokens": 794468258.0, "step": 20920 }, { "epoch": 2.661366238392062, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.439528703689575, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8749310970306396, "num_tokens": 794505493.0, "step": 20921 }, { "epoch": 2.6614934486706527, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6026201248168945, "learning_rate": 1e-05, "loss": 0.3606, "mean_token_accuracy": 0.9013153314590454, "num_tokens": 794539370.0, "step": 20922 }, { "epoch": 2.661620658949243, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3573100566864014, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8766303658485413, "num_tokens": 794580820.0, "step": 20923 }, { "epoch": 2.661747869227834, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.677490711212158, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8862869739532471, "num_tokens": 794617503.0, "step": 20924 }, { "epoch": 2.661875079506424, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.306333065032959, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8800327181816101, "num_tokens": 794653701.0, "step": 20925 }, { "epoch": 2.662002289785015, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7948055267333984, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8764858245849609, "num_tokens": 794689938.0, "step": 20926 }, { "epoch": 2.662129500063605, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2365357875823975, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8699253797531128, "num_tokens": 794730460.0, "step": 20927 }, { "epoch": 2.662256710342196, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9648215770721436, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8865802884101868, "num_tokens": 794770686.0, "step": 20928 }, { "epoch": 2.662383920620786, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2091009616851807, "learning_rate": 1e-05, "loss": 0.4759, "mean_token_accuracy": 0.8668003678321838, "num_tokens": 794814396.0, "step": 20929 }, { "epoch": 2.6625111308993765, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.198399543762207, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8832237720489502, "num_tokens": 794857425.0, "step": 20930 }, { "epoch": 2.662638341177967, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3352138996124268, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.884354829788208, "num_tokens": 794894115.0, "step": 20931 }, { "epoch": 2.6627655514565576, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.365053176879883, "learning_rate": 1e-05, "loss": 0.4604, "mean_token_accuracy": 0.8750991821289062, "num_tokens": 794929423.0, "step": 20932 }, { "epoch": 2.662892761735148, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.59390926361084, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8780397176742554, "num_tokens": 794966019.0, "step": 20933 }, { "epoch": 2.6630199720137386, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.303894519805908, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8762567043304443, "num_tokens": 795006501.0, "step": 20934 }, { "epoch": 2.663147182292329, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.754901170730591, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8797000646591187, "num_tokens": 795048068.0, "step": 20935 }, { "epoch": 2.6632743925709197, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.120718479156494, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8683807849884033, "num_tokens": 795084918.0, "step": 20936 }, { "epoch": 2.66340160284951, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7697596549987793, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8840968608856201, "num_tokens": 795119261.0, "step": 20937 }, { "epoch": 2.6635288131281007, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9247496128082275, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.887046217918396, "num_tokens": 795162559.0, "step": 20938 }, { "epoch": 2.6636560234066913, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.697720766067505, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8836970329284668, "num_tokens": 795200853.0, "step": 20939 }, { "epoch": 2.663783233685282, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.777324676513672, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8863898515701294, "num_tokens": 795238405.0, "step": 20940 }, { "epoch": 2.6639104439638723, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6245720386505127, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8735383152961731, "num_tokens": 795278341.0, "step": 20941 }, { "epoch": 2.664037654242463, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.682297945022583, "learning_rate": 1e-05, "loss": 0.457, "mean_token_accuracy": 0.8724507093429565, "num_tokens": 795319108.0, "step": 20942 }, { "epoch": 2.6641648645210534, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.546558141708374, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8779032230377197, "num_tokens": 795354641.0, "step": 20943 }, { "epoch": 2.664292074799644, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.630218029022217, "learning_rate": 1e-05, "loss": 0.4834, "mean_token_accuracy": 0.8647053837776184, "num_tokens": 795399098.0, "step": 20944 }, { "epoch": 2.6644192850782344, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4625139236450195, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8843262195587158, "num_tokens": 795444191.0, "step": 20945 }, { "epoch": 2.664546495356825, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.666398525238037, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8772456645965576, "num_tokens": 795476450.0, "step": 20946 }, { "epoch": 2.6646737056354155, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3809940814971924, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.876564621925354, "num_tokens": 795518740.0, "step": 20947 }, { "epoch": 2.6648009159140056, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.727968692779541, "learning_rate": 1e-05, "loss": 0.4593, "mean_token_accuracy": 0.8741322755813599, "num_tokens": 795558944.0, "step": 20948 }, { "epoch": 2.6649281261925966, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3157079219818115, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8799183964729309, "num_tokens": 795592685.0, "step": 20949 }, { "epoch": 2.6650553364711866, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.842872142791748, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8725903034210205, "num_tokens": 795627878.0, "step": 20950 }, { "epoch": 2.6651825467497776, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.26296067237854, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8672401905059814, "num_tokens": 795666784.0, "step": 20951 }, { "epoch": 2.6653097570283677, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.0151872634887695, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8790995478630066, "num_tokens": 795701921.0, "step": 20952 }, { "epoch": 2.6654369673069587, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.264753818511963, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8726679086685181, "num_tokens": 795744497.0, "step": 20953 }, { "epoch": 2.6655641775855488, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.195526599884033, "learning_rate": 1e-05, "loss": 0.4854, "mean_token_accuracy": 0.8654812574386597, "num_tokens": 795786871.0, "step": 20954 }, { "epoch": 2.6656913878641393, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.404494047164917, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8876663446426392, "num_tokens": 795829403.0, "step": 20955 }, { "epoch": 2.66581859814273, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.283709526062012, "learning_rate": 1e-05, "loss": 0.5018, "mean_token_accuracy": 0.8624221086502075, "num_tokens": 795867352.0, "step": 20956 }, { "epoch": 2.6659458084213203, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6431069374084473, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8700644969940186, "num_tokens": 795907209.0, "step": 20957 }, { "epoch": 2.666073018699911, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.14967679977417, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8747962713241577, "num_tokens": 795946515.0, "step": 20958 }, { "epoch": 2.6662002289785014, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.721332550048828, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8876965641975403, "num_tokens": 795983328.0, "step": 20959 }, { "epoch": 2.666327439257092, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9530279636383057, "learning_rate": 1e-05, "loss": 0.4774, "mean_token_accuracy": 0.8626604676246643, "num_tokens": 796018609.0, "step": 20960 }, { "epoch": 2.6664546495356825, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.711507797241211, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8725129961967468, "num_tokens": 796059676.0, "step": 20961 }, { "epoch": 2.666581859814273, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.751281499862671, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8750842809677124, "num_tokens": 796100896.0, "step": 20962 }, { "epoch": 2.6667090700928635, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.651975631713867, "learning_rate": 1e-05, "loss": 0.3943, "mean_token_accuracy": 0.892723798751831, "num_tokens": 796140513.0, "step": 20963 }, { "epoch": 2.666836280371454, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.618623733520508, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8898868560791016, "num_tokens": 796174339.0, "step": 20964 }, { "epoch": 2.6669634906500446, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6145405769348145, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8826233148574829, "num_tokens": 796216624.0, "step": 20965 }, { "epoch": 2.667090700928635, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5427021980285645, "learning_rate": 1e-05, "loss": 0.4669, "mean_token_accuracy": 0.8639692664146423, "num_tokens": 796250681.0, "step": 20966 }, { "epoch": 2.6672179112072256, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.587764263153076, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8806032538414001, "num_tokens": 796284860.0, "step": 20967 }, { "epoch": 2.667345121485816, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4730966091156006, "learning_rate": 1e-05, "loss": 0.3903, "mean_token_accuracy": 0.8921945095062256, "num_tokens": 796325833.0, "step": 20968 }, { "epoch": 2.6674723317644067, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5879361629486084, "learning_rate": 1e-05, "loss": 0.401, "mean_token_accuracy": 0.8873851299285889, "num_tokens": 796354955.0, "step": 20969 }, { "epoch": 2.667599542042997, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4332191944122314, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8778077363967896, "num_tokens": 796390505.0, "step": 20970 }, { "epoch": 2.6677267523215877, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5960233211517334, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.886199414730072, "num_tokens": 796428003.0, "step": 20971 }, { "epoch": 2.6678539626001783, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.368410587310791, "learning_rate": 1e-05, "loss": 0.3998, "mean_token_accuracy": 0.8878511190414429, "num_tokens": 796467396.0, "step": 20972 }, { "epoch": 2.6679811728787683, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6431374549865723, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8764786124229431, "num_tokens": 796505320.0, "step": 20973 }, { "epoch": 2.6681083831573593, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3166422843933105, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8880767822265625, "num_tokens": 796542109.0, "step": 20974 }, { "epoch": 2.6682355934359494, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7131502628326416, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8760140538215637, "num_tokens": 796588523.0, "step": 20975 }, { "epoch": 2.6683628037145404, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.263413667678833, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8623117208480835, "num_tokens": 796624493.0, "step": 20976 }, { "epoch": 2.6684900139931305, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.839646100997925, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8900960683822632, "num_tokens": 796662147.0, "step": 20977 }, { "epoch": 2.668617224271721, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1953604221343994, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8763360977172852, "num_tokens": 796700024.0, "step": 20978 }, { "epoch": 2.6687444345503115, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.058961391448975, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8839218616485596, "num_tokens": 796735058.0, "step": 20979 }, { "epoch": 2.668871644828902, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.189445972442627, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8709745407104492, "num_tokens": 796778419.0, "step": 20980 }, { "epoch": 2.6689988551074926, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.325888156890869, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8854614496231079, "num_tokens": 796817495.0, "step": 20981 }, { "epoch": 2.669126065386083, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.419105052947998, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8758906722068787, "num_tokens": 796850432.0, "step": 20982 }, { "epoch": 2.6692532756646736, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.44612455368042, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.872828483581543, "num_tokens": 796889584.0, "step": 20983 }, { "epoch": 2.669380485943264, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7083892822265625, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.879004716873169, "num_tokens": 796924052.0, "step": 20984 }, { "epoch": 2.6695076962218547, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.319042682647705, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8700987696647644, "num_tokens": 796963795.0, "step": 20985 }, { "epoch": 2.669634906500445, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.835536241531372, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8888447880744934, "num_tokens": 797002621.0, "step": 20986 }, { "epoch": 2.6697621167790357, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1042351722717285, "learning_rate": 1e-05, "loss": 0.5081, "mean_token_accuracy": 0.8562297224998474, "num_tokens": 797047173.0, "step": 20987 }, { "epoch": 2.6698893270576263, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.794536828994751, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.877187967300415, "num_tokens": 797088394.0, "step": 20988 }, { "epoch": 2.670016537336217, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9211740493774414, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8862855434417725, "num_tokens": 797126108.0, "step": 20989 }, { "epoch": 2.6701437476148073, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.705714464187622, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8846699595451355, "num_tokens": 797162718.0, "step": 20990 }, { "epoch": 2.670270957893398, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.80192494392395, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8684316873550415, "num_tokens": 797197552.0, "step": 20991 }, { "epoch": 2.6703981681719884, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6234543323516846, "learning_rate": 1e-05, "loss": 0.4964, "mean_token_accuracy": 0.8577848672866821, "num_tokens": 797232523.0, "step": 20992 }, { "epoch": 2.670525378450579, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.71201229095459, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8818482160568237, "num_tokens": 797266620.0, "step": 20993 }, { "epoch": 2.6706525887291694, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.514960527420044, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8848832845687866, "num_tokens": 797305378.0, "step": 20994 }, { "epoch": 2.67077979900776, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.691012382507324, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8845593929290771, "num_tokens": 797343269.0, "step": 20995 }, { "epoch": 2.6709070092863505, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.434115409851074, "learning_rate": 1e-05, "loss": 0.3549, "mean_token_accuracy": 0.9021462202072144, "num_tokens": 797375034.0, "step": 20996 }, { "epoch": 2.671034219564941, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.740523338317871, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8740199208259583, "num_tokens": 797410869.0, "step": 20997 }, { "epoch": 2.671161429843531, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.356114625930786, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8873883485794067, "num_tokens": 797449189.0, "step": 20998 }, { "epoch": 2.671288640122122, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.821301221847534, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8830119967460632, "num_tokens": 797492182.0, "step": 20999 }, { "epoch": 2.671415850400712, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2865219116210938, "learning_rate": 1e-05, "loss": 0.3687, "mean_token_accuracy": 0.8966099619865417, "num_tokens": 797529043.0, "step": 21000 }, { "epoch": 2.671543060679303, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9958136081695557, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8723451495170593, "num_tokens": 797560630.0, "step": 21001 }, { "epoch": 2.6716702709578932, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2940757274627686, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.8734660148620605, "num_tokens": 797598667.0, "step": 21002 }, { "epoch": 2.6717974812364838, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.158468246459961, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8882859349250793, "num_tokens": 797637420.0, "step": 21003 }, { "epoch": 2.6719246915150743, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4221928119659424, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8851984739303589, "num_tokens": 797670969.0, "step": 21004 }, { "epoch": 2.672051901793665, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.21455192565918, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.880325436592102, "num_tokens": 797711915.0, "step": 21005 }, { "epoch": 2.6721791120722553, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.591219663619995, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8784289360046387, "num_tokens": 797751036.0, "step": 21006 }, { "epoch": 2.672306322350846, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1107096672058105, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8711090087890625, "num_tokens": 797793999.0, "step": 21007 }, { "epoch": 2.6724335326294364, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6841282844543457, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8769382834434509, "num_tokens": 797834836.0, "step": 21008 }, { "epoch": 2.672560742908027, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.927853584289551, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8684467077255249, "num_tokens": 797872855.0, "step": 21009 }, { "epoch": 2.6726879531866174, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.7321622371673584, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8759869337081909, "num_tokens": 797910925.0, "step": 21010 }, { "epoch": 2.672815163465208, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.704597234725952, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8710461258888245, "num_tokens": 797952201.0, "step": 21011 }, { "epoch": 2.6729423737437985, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6891746520996094, "learning_rate": 1e-05, "loss": 0.4923, "mean_token_accuracy": 0.8604185581207275, "num_tokens": 797988706.0, "step": 21012 }, { "epoch": 2.673069584022389, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5439558029174805, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8724807500839233, "num_tokens": 798028438.0, "step": 21013 }, { "epoch": 2.6731967943009796, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6702945232391357, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8886327743530273, "num_tokens": 798062864.0, "step": 21014 }, { "epoch": 2.67332400457957, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4344327449798584, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8826172351837158, "num_tokens": 798098271.0, "step": 21015 }, { "epoch": 2.6734512148581606, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.688732385635376, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8840923309326172, "num_tokens": 798131286.0, "step": 21016 }, { "epoch": 2.673578425136751, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3575098514556885, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.879256010055542, "num_tokens": 798164711.0, "step": 21017 }, { "epoch": 2.6737056354153417, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7451529502868652, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8796671628952026, "num_tokens": 798205280.0, "step": 21018 }, { "epoch": 2.673832845693932, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.284015655517578, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8797511458396912, "num_tokens": 798246561.0, "step": 21019 }, { "epoch": 2.6739600559725227, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8464083671569824, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8808729648590088, "num_tokens": 798287494.0, "step": 21020 }, { "epoch": 2.674087266251113, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2420356273651123, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8807468414306641, "num_tokens": 798325305.0, "step": 21021 }, { "epoch": 2.674214476529704, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9948809146881104, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8890782594680786, "num_tokens": 798363083.0, "step": 21022 }, { "epoch": 2.674341686808294, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2613649368286133, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.860551118850708, "num_tokens": 798403324.0, "step": 21023 }, { "epoch": 2.674468897086885, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.169699192047119, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8851196765899658, "num_tokens": 798441548.0, "step": 21024 }, { "epoch": 2.674596107365475, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.359809398651123, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.873864471912384, "num_tokens": 798486038.0, "step": 21025 }, { "epoch": 2.674723317644066, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.281179428100586, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8876367807388306, "num_tokens": 798530657.0, "step": 21026 }, { "epoch": 2.674850527922656, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.5644876956939697, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8808121681213379, "num_tokens": 798571027.0, "step": 21027 }, { "epoch": 2.6749777382012465, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.233053684234619, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8870387077331543, "num_tokens": 798609874.0, "step": 21028 }, { "epoch": 2.675104948479837, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.712803363800049, "learning_rate": 1e-05, "loss": 0.4651, "mean_token_accuracy": 0.87046217918396, "num_tokens": 798645191.0, "step": 21029 }, { "epoch": 2.6752321587584276, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.05356502532959, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8865009546279907, "num_tokens": 798688445.0, "step": 21030 }, { "epoch": 2.675359369037018, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.714735507965088, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8806635737419128, "num_tokens": 798729582.0, "step": 21031 }, { "epoch": 2.6754865793156086, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8720102310180664, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8863739967346191, "num_tokens": 798765178.0, "step": 21032 }, { "epoch": 2.675613789594199, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.663642168045044, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8785707354545593, "num_tokens": 798806048.0, "step": 21033 }, { "epoch": 2.6757409998727897, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.730518102645874, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.884528636932373, "num_tokens": 798845554.0, "step": 21034 }, { "epoch": 2.67586821015138, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.598400592803955, "learning_rate": 1e-05, "loss": 0.3872, "mean_token_accuracy": 0.8955105543136597, "num_tokens": 798883124.0, "step": 21035 }, { "epoch": 2.6759954204299707, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6512093544006348, "learning_rate": 1e-05, "loss": 0.3993, "mean_token_accuracy": 0.8924952745437622, "num_tokens": 798916040.0, "step": 21036 }, { "epoch": 2.6761226307085613, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.53714919090271, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8781630992889404, "num_tokens": 798952562.0, "step": 21037 }, { "epoch": 2.676249840987152, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6154768466949463, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.873583972454071, "num_tokens": 798990689.0, "step": 21038 }, { "epoch": 2.6763770512657423, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4825267791748047, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.883610725402832, "num_tokens": 799029398.0, "step": 21039 }, { "epoch": 2.676504261544333, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.579967498779297, "learning_rate": 1e-05, "loss": 0.3944, "mean_token_accuracy": 0.8922054767608643, "num_tokens": 799073200.0, "step": 21040 }, { "epoch": 2.6766314718229234, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4214603900909424, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8866406679153442, "num_tokens": 799115020.0, "step": 21041 }, { "epoch": 2.676758682101514, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.615701198577881, "learning_rate": 1e-05, "loss": 0.3976, "mean_token_accuracy": 0.8894936442375183, "num_tokens": 799152348.0, "step": 21042 }, { "epoch": 2.6768858923801044, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.370835304260254, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8768644332885742, "num_tokens": 799191120.0, "step": 21043 }, { "epoch": 2.677013102658695, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6882708072662354, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8798210620880127, "num_tokens": 799229044.0, "step": 21044 }, { "epoch": 2.6771403129372855, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.29707932472229, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8864145278930664, "num_tokens": 799269148.0, "step": 21045 }, { "epoch": 2.6772675232158756, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7985095977783203, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.8880674242973328, "num_tokens": 799305188.0, "step": 21046 }, { "epoch": 2.6773947334944665, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2414941787719727, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8708724975585938, "num_tokens": 799344156.0, "step": 21047 }, { "epoch": 2.6775219437730566, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9962406158447266, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8682537078857422, "num_tokens": 799374792.0, "step": 21048 }, { "epoch": 2.6776491540516476, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2335336208343506, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.879764199256897, "num_tokens": 799408425.0, "step": 21049 }, { "epoch": 2.6777763643302377, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.20250129699707, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8841387033462524, "num_tokens": 799448362.0, "step": 21050 }, { "epoch": 2.6779035746088287, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3477485179901123, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8853271007537842, "num_tokens": 799482225.0, "step": 21051 }, { "epoch": 2.6780307848874187, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.352794170379639, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.865552544593811, "num_tokens": 799523053.0, "step": 21052 }, { "epoch": 2.6781579951660093, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5959882736206055, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8801864981651306, "num_tokens": 799563070.0, "step": 21053 }, { "epoch": 2.6782852054446, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.294961929321289, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8872500658035278, "num_tokens": 799595465.0, "step": 21054 }, { "epoch": 2.6784124157231903, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7619171142578125, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8878800868988037, "num_tokens": 799632308.0, "step": 21055 }, { "epoch": 2.678539626001781, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.093813419342041, "learning_rate": 1e-05, "loss": 0.4046, "mean_token_accuracy": 0.8931413292884827, "num_tokens": 799669563.0, "step": 21056 }, { "epoch": 2.6786668362803714, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.75604510307312, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8771161437034607, "num_tokens": 799714575.0, "step": 21057 }, { "epoch": 2.678794046558962, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9150807857513428, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8825312852859497, "num_tokens": 799752730.0, "step": 21058 }, { "epoch": 2.6789212568375524, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6958401203155518, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8675115704536438, "num_tokens": 799788110.0, "step": 21059 }, { "epoch": 2.679048467116143, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.762519359588623, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8878329396247864, "num_tokens": 799825561.0, "step": 21060 }, { "epoch": 2.6791756773947335, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6120851039886475, "learning_rate": 1e-05, "loss": 0.4071, "mean_token_accuracy": 0.8859747648239136, "num_tokens": 799865412.0, "step": 21061 }, { "epoch": 2.679302887673324, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.688124418258667, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8677032589912415, "num_tokens": 799900238.0, "step": 21062 }, { "epoch": 2.6794300979519146, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5402097702026367, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8734008073806763, "num_tokens": 799936724.0, "step": 21063 }, { "epoch": 2.679557308230505, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.640995502471924, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8756091594696045, "num_tokens": 799978347.0, "step": 21064 }, { "epoch": 2.6796845185090956, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.468895673751831, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.875515341758728, "num_tokens": 800015419.0, "step": 21065 }, { "epoch": 2.679811728787686, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.638155937194824, "learning_rate": 1e-05, "loss": 0.3749, "mean_token_accuracy": 0.9000592827796936, "num_tokens": 800048435.0, "step": 21066 }, { "epoch": 2.6799389390662767, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4153311252593994, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8810255527496338, "num_tokens": 800083317.0, "step": 21067 }, { "epoch": 2.680066149344867, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.665468692779541, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8767894506454468, "num_tokens": 800123303.0, "step": 21068 }, { "epoch": 2.6801933596234577, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3328874111175537, "learning_rate": 1e-05, "loss": 0.4216, "mean_token_accuracy": 0.882479727268219, "num_tokens": 800165157.0, "step": 21069 }, { "epoch": 2.6803205699020483, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.771233558654785, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8728256225585938, "num_tokens": 800203657.0, "step": 21070 }, { "epoch": 2.6804477801806383, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2611312866210938, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.885638952255249, "num_tokens": 800245008.0, "step": 21071 }, { "epoch": 2.6805749904592293, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9267935752868652, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8800508975982666, "num_tokens": 800285438.0, "step": 21072 }, { "epoch": 2.6807022007378194, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.254197597503662, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8736393451690674, "num_tokens": 800321345.0, "step": 21073 }, { "epoch": 2.6808294110164104, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.11529016494751, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8720368146896362, "num_tokens": 800361657.0, "step": 21074 }, { "epoch": 2.6809566212950005, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3265721797943115, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8590964078903198, "num_tokens": 800393809.0, "step": 21075 }, { "epoch": 2.681083831573591, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2577667236328125, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8810818791389465, "num_tokens": 800432541.0, "step": 21076 }, { "epoch": 2.6812110418521815, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5037682056427, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8765748143196106, "num_tokens": 800468552.0, "step": 21077 }, { "epoch": 2.681338252130772, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.260660648345947, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8854023814201355, "num_tokens": 800503911.0, "step": 21078 }, { "epoch": 2.6814654624093626, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.663762331008911, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8789178133010864, "num_tokens": 800549755.0, "step": 21079 }, { "epoch": 2.681592672687953, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.115042686462402, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.8863503336906433, "num_tokens": 800583240.0, "step": 21080 }, { "epoch": 2.6817198829665436, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7346739768981934, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8637673258781433, "num_tokens": 800624016.0, "step": 21081 }, { "epoch": 2.681847093245134, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.919229030609131, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8620902299880981, "num_tokens": 800661052.0, "step": 21082 }, { "epoch": 2.6819743035237247, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6924004554748535, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8814854621887207, "num_tokens": 800701549.0, "step": 21083 }, { "epoch": 2.682101513802315, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.747159004211426, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8768283724784851, "num_tokens": 800738650.0, "step": 21084 }, { "epoch": 2.6822287240809057, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.63244891166687, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8816367983818054, "num_tokens": 800773687.0, "step": 21085 }, { "epoch": 2.6823559343594963, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.628030300140381, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.8879980444908142, "num_tokens": 800812523.0, "step": 21086 }, { "epoch": 2.682483144638087, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5939230918884277, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8805477619171143, "num_tokens": 800850838.0, "step": 21087 }, { "epoch": 2.6826103549166773, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5667238235473633, "learning_rate": 1e-05, "loss": 0.3959, "mean_token_accuracy": 0.8885396718978882, "num_tokens": 800886734.0, "step": 21088 }, { "epoch": 2.682737565195268, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.536397933959961, "learning_rate": 1e-05, "loss": 0.4435, "mean_token_accuracy": 0.8737603425979614, "num_tokens": 800930816.0, "step": 21089 }, { "epoch": 2.6828647754738584, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5512447357177734, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8760006427764893, "num_tokens": 800962982.0, "step": 21090 }, { "epoch": 2.682991985752449, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4873337745666504, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8663877248764038, "num_tokens": 801003608.0, "step": 21091 }, { "epoch": 2.6831191960310394, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5587964057922363, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8741576075553894, "num_tokens": 801032889.0, "step": 21092 }, { "epoch": 2.68324640630963, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4386916160583496, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8705509901046753, "num_tokens": 801068717.0, "step": 21093 }, { "epoch": 2.6833736165882205, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5735578536987305, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8788408041000366, "num_tokens": 801105120.0, "step": 21094 }, { "epoch": 2.683500826866811, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.39473557472229, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8728125095367432, "num_tokens": 801140301.0, "step": 21095 }, { "epoch": 2.683628037145401, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.602247476577759, "learning_rate": 1e-05, "loss": 0.3687, "mean_token_accuracy": 0.8992457389831543, "num_tokens": 801176314.0, "step": 21096 }, { "epoch": 2.683755247423992, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.330709457397461, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.884048581123352, "num_tokens": 801212906.0, "step": 21097 }, { "epoch": 2.683882457702582, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.694408416748047, "learning_rate": 1e-05, "loss": 0.4656, "mean_token_accuracy": 0.8677889108657837, "num_tokens": 801243084.0, "step": 21098 }, { "epoch": 2.684009667981173, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2506463527679443, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8771277666091919, "num_tokens": 801282701.0, "step": 21099 }, { "epoch": 2.684136878259763, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8622641563415527, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8863873481750488, "num_tokens": 801315363.0, "step": 21100 }, { "epoch": 2.6842640885383537, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1774795055389404, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8751208186149597, "num_tokens": 801354103.0, "step": 21101 }, { "epoch": 2.6843912988169443, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.095134258270264, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8782216310501099, "num_tokens": 801393351.0, "step": 21102 }, { "epoch": 2.684518509095535, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2031137943267822, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8735361695289612, "num_tokens": 801432829.0, "step": 21103 }, { "epoch": 2.6846457193741253, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.373528003692627, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8758977651596069, "num_tokens": 801468735.0, "step": 21104 }, { "epoch": 2.684772929652716, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.4432032108306885, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8896795511245728, "num_tokens": 801507238.0, "step": 21105 }, { "epoch": 2.6849001399313064, "ewc_loss": 0.107421875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.4965033531188965, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8860286474227905, "num_tokens": 801543040.0, "step": 21106 }, { "epoch": 2.685027350209897, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.7784125804901123, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8783187866210938, "num_tokens": 801584466.0, "step": 21107 }, { "epoch": 2.6851545604884874, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.336694240570068, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8868939876556396, "num_tokens": 801617466.0, "step": 21108 }, { "epoch": 2.685281770767078, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.902031183242798, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8688473701477051, "num_tokens": 801653408.0, "step": 21109 }, { "epoch": 2.6854089810456685, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.104265213012695, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8765036463737488, "num_tokens": 801687200.0, "step": 21110 }, { "epoch": 2.685536191324259, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.812206268310547, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8759440779685974, "num_tokens": 801728547.0, "step": 21111 }, { "epoch": 2.6856634016028496, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9229865074157715, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8716620802879333, "num_tokens": 801765425.0, "step": 21112 }, { "epoch": 2.68579061188144, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.693635940551758, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8817600011825562, "num_tokens": 801800694.0, "step": 21113 }, { "epoch": 2.6859178221600306, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8067212104797363, "learning_rate": 1e-05, "loss": 0.3964, "mean_token_accuracy": 0.8915446996688843, "num_tokens": 801835748.0, "step": 21114 }, { "epoch": 2.686045032438621, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.589989423751831, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8735668659210205, "num_tokens": 801873527.0, "step": 21115 }, { "epoch": 2.6861722427172117, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.763664960861206, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8878500461578369, "num_tokens": 801910630.0, "step": 21116 }, { "epoch": 2.686299452995802, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.472289800643921, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8838770389556885, "num_tokens": 801946136.0, "step": 21117 }, { "epoch": 2.6864266632743927, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.795989751815796, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8752976059913635, "num_tokens": 801982283.0, "step": 21118 }, { "epoch": 2.686553873552983, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.3856115341186523, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.875002384185791, "num_tokens": 802022263.0, "step": 21119 }, { "epoch": 2.686681083831574, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.902329683303833, "learning_rate": 1e-05, "loss": 0.5455, "mean_token_accuracy": 0.8409976959228516, "num_tokens": 802061837.0, "step": 21120 }, { "epoch": 2.686808294110164, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.324517250061035, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8864386677742004, "num_tokens": 802094528.0, "step": 21121 }, { "epoch": 2.686935504388755, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.042852401733398, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8737480640411377, "num_tokens": 802135157.0, "step": 21122 }, { "epoch": 2.687062714667345, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.3450427055358887, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.8891528844833374, "num_tokens": 802168818.0, "step": 21123 }, { "epoch": 2.687189924945936, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.181888103485107, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8765087723731995, "num_tokens": 802209486.0, "step": 21124 }, { "epoch": 2.687317135224526, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.492788791656494, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8669403195381165, "num_tokens": 802248534.0, "step": 21125 }, { "epoch": 2.6874443455031165, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.191339015960693, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8797553777694702, "num_tokens": 802289330.0, "step": 21126 }, { "epoch": 2.687571555781707, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6556503772735596, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8795545101165771, "num_tokens": 802322065.0, "step": 21127 }, { "epoch": 2.6876987660602976, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.037353515625, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8748632669448853, "num_tokens": 802361820.0, "step": 21128 }, { "epoch": 2.687825976338888, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.710355043411255, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8803356885910034, "num_tokens": 802397748.0, "step": 21129 }, { "epoch": 2.6879531866174786, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.822782278060913, "learning_rate": 1e-05, "loss": 0.4703, "mean_token_accuracy": 0.8651349544525146, "num_tokens": 802432152.0, "step": 21130 }, { "epoch": 2.688080396896069, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6983063220977783, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8833098411560059, "num_tokens": 802471464.0, "step": 21131 }, { "epoch": 2.6882076071746597, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.639840841293335, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8786071538925171, "num_tokens": 802503096.0, "step": 21132 }, { "epoch": 2.68833481745325, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.690976619720459, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.869238018989563, "num_tokens": 802540036.0, "step": 21133 }, { "epoch": 2.6884620277318407, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.487609386444092, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8773822784423828, "num_tokens": 802580798.0, "step": 21134 }, { "epoch": 2.6885892380104313, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.682943105697632, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8726681470870972, "num_tokens": 802619472.0, "step": 21135 }, { "epoch": 2.688716448289022, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3882153034210205, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8800981044769287, "num_tokens": 802656834.0, "step": 21136 }, { "epoch": 2.6888436585676123, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.722482442855835, "learning_rate": 1e-05, "loss": 0.3691, "mean_token_accuracy": 0.898549497127533, "num_tokens": 802690269.0, "step": 21137 }, { "epoch": 2.688970868846203, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.314739942550659, "learning_rate": 1e-05, "loss": 0.3956, "mean_token_accuracy": 0.8896308541297913, "num_tokens": 802726822.0, "step": 21138 }, { "epoch": 2.6890980791247934, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8178553581237793, "learning_rate": 1e-05, "loss": 0.4123, "mean_token_accuracy": 0.887290358543396, "num_tokens": 802764479.0, "step": 21139 }, { "epoch": 2.689225289403384, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2655656337738037, "learning_rate": 1e-05, "loss": 0.4111, "mean_token_accuracy": 0.8825832605361938, "num_tokens": 802800938.0, "step": 21140 }, { "epoch": 2.6893524996819744, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9493937492370605, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8827953338623047, "num_tokens": 802841593.0, "step": 21141 }, { "epoch": 2.689479709960565, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2717132568359375, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8760918378829956, "num_tokens": 802875847.0, "step": 21142 }, { "epoch": 2.6896069202391555, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.101742267608643, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8829327821731567, "num_tokens": 802915786.0, "step": 21143 }, { "epoch": 2.6897341305177456, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3082573413848877, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8749183416366577, "num_tokens": 802958774.0, "step": 21144 }, { "epoch": 2.6898613407963365, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.261233806610107, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8928279876708984, "num_tokens": 802991360.0, "step": 21145 }, { "epoch": 2.6899885510749266, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.487305164337158, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8860993981361389, "num_tokens": 803030852.0, "step": 21146 }, { "epoch": 2.6901157613535176, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.277966499328613, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8839432597160339, "num_tokens": 803068965.0, "step": 21147 }, { "epoch": 2.6902429716321077, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.693126916885376, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8734996318817139, "num_tokens": 803103331.0, "step": 21148 }, { "epoch": 2.6903701819106987, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.111143589019775, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8826165795326233, "num_tokens": 803144242.0, "step": 21149 }, { "epoch": 2.6904973921892887, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7416887283325195, "learning_rate": 1e-05, "loss": 0.488, "mean_token_accuracy": 0.8633072972297668, "num_tokens": 803178355.0, "step": 21150 }, { "epoch": 2.6906246024678793, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9106967449188232, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8763724565505981, "num_tokens": 803218783.0, "step": 21151 }, { "epoch": 2.69075181274647, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00034332275390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.7068402767181396, "learning_rate": 1e-05, "loss": 0.507, "mean_token_accuracy": 0.8596179485321045, "num_tokens": 803256391.0, "step": 21152 }, { "epoch": 2.6908790230250603, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7475037574768066, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.877697229385376, "num_tokens": 803285708.0, "step": 21153 }, { "epoch": 2.691006233303651, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6485471725463867, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.880776047706604, "num_tokens": 803319711.0, "step": 21154 }, { "epoch": 2.6911334435822414, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.635491371154785, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8648768067359924, "num_tokens": 803353916.0, "step": 21155 }, { "epoch": 2.691260653860832, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.576939344406128, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.8904281854629517, "num_tokens": 803388579.0, "step": 21156 }, { "epoch": 2.6913878641394224, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.558056354522705, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8889436721801758, "num_tokens": 803427715.0, "step": 21157 }, { "epoch": 2.691515074418013, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5256290435791016, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8687998056411743, "num_tokens": 803470363.0, "step": 21158 }, { "epoch": 2.6916422846966035, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.533555030822754, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8746922016143799, "num_tokens": 803515009.0, "step": 21159 }, { "epoch": 2.691769494975194, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4824376106262207, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8817664384841919, "num_tokens": 803554773.0, "step": 21160 }, { "epoch": 2.6918967052537845, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.550874948501587, "learning_rate": 1e-05, "loss": 0.4048, "mean_token_accuracy": 0.8841527700424194, "num_tokens": 803590404.0, "step": 21161 }, { "epoch": 2.692023915532375, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4329285621643066, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8876231908798218, "num_tokens": 803634388.0, "step": 21162 }, { "epoch": 2.6921511258109656, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.572477102279663, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8815988302230835, "num_tokens": 803673164.0, "step": 21163 }, { "epoch": 2.692278336089556, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.384267807006836, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8785990476608276, "num_tokens": 803715271.0, "step": 21164 }, { "epoch": 2.6924055463681467, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6316921710968018, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8851114511489868, "num_tokens": 803748295.0, "step": 21165 }, { "epoch": 2.692532756646737, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.314066171646118, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8705034852027893, "num_tokens": 803787591.0, "step": 21166 }, { "epoch": 2.6926599669253277, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.743488311767578, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8801612854003906, "num_tokens": 803817640.0, "step": 21167 }, { "epoch": 2.6927871772039182, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.247999668121338, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8709940910339355, "num_tokens": 803847932.0, "step": 21168 }, { "epoch": 2.6929143874825083, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9034101963043213, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8710715770721436, "num_tokens": 803884206.0, "step": 21169 }, { "epoch": 2.6930415977610993, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.173617124557495, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8824828863143921, "num_tokens": 803923225.0, "step": 21170 }, { "epoch": 2.6931688080396894, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.169580459594727, "learning_rate": 1e-05, "loss": 0.4897, "mean_token_accuracy": 0.8634414672851562, "num_tokens": 803961880.0, "step": 21171 }, { "epoch": 2.6932960183182804, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2360005378723145, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8825759291648865, "num_tokens": 804004676.0, "step": 21172 }, { "epoch": 2.6934232285968704, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 4.436887264251709, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8816131949424744, "num_tokens": 804047875.0, "step": 21173 }, { "epoch": 2.693550438875461, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.543212652206421, "learning_rate": 1e-05, "loss": 0.3768, "mean_token_accuracy": 0.8983870148658752, "num_tokens": 804083467.0, "step": 21174 }, { "epoch": 2.6936776491540515, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 7.2479248046875e-05, "grad_norm": 4.474346160888672, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8879883289337158, "num_tokens": 804124653.0, "step": 21175 }, { "epoch": 2.693804859432642, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.8494696617126465, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8863740563392639, "num_tokens": 804159081.0, "step": 21176 }, { "epoch": 2.6939320697112326, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.25352144241333, "learning_rate": 1e-05, "loss": 0.4627, "mean_token_accuracy": 0.8697617053985596, "num_tokens": 804200687.0, "step": 21177 }, { "epoch": 2.694059279989823, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.8754873275756836, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8754911422729492, "num_tokens": 804237192.0, "step": 21178 }, { "epoch": 2.6941864902684136, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.04429817199707, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8842417597770691, "num_tokens": 804275078.0, "step": 21179 }, { "epoch": 2.694313700547004, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.782590389251709, "learning_rate": 1e-05, "loss": 0.465, "mean_token_accuracy": 0.8734992146492004, "num_tokens": 804308551.0, "step": 21180 }, { "epoch": 2.6944409108255947, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8813703060150146, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8779894709587097, "num_tokens": 804347530.0, "step": 21181 }, { "epoch": 2.694568121104185, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.667398452758789, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8720442056655884, "num_tokens": 804388663.0, "step": 21182 }, { "epoch": 2.6946953313827757, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7823195457458496, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8847825527191162, "num_tokens": 804423956.0, "step": 21183 }, { "epoch": 2.6948225416613663, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5531325340270996, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8757106065750122, "num_tokens": 804466736.0, "step": 21184 }, { "epoch": 2.694949751939957, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7694127559661865, "learning_rate": 1e-05, "loss": 0.5249, "mean_token_accuracy": 0.8533518314361572, "num_tokens": 804505101.0, "step": 21185 }, { "epoch": 2.6950769622185473, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4446969032287598, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8866031765937805, "num_tokens": 804542520.0, "step": 21186 }, { "epoch": 2.695204172497138, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8165836334228516, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8865844011306763, "num_tokens": 804577749.0, "step": 21187 }, { "epoch": 2.6953313827757284, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3646247386932373, "learning_rate": 1e-05, "loss": 0.3765, "mean_token_accuracy": 0.8959063291549683, "num_tokens": 804611370.0, "step": 21188 }, { "epoch": 2.695458593054319, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.931652069091797, "learning_rate": 1e-05, "loss": 0.3976, "mean_token_accuracy": 0.8898981213569641, "num_tokens": 804643696.0, "step": 21189 }, { "epoch": 2.6955858033329094, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3221263885498047, "learning_rate": 1e-05, "loss": 0.3806, "mean_token_accuracy": 0.8959431052207947, "num_tokens": 804678343.0, "step": 21190 }, { "epoch": 2.6957130136115, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.083254337310791, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8930486440658569, "num_tokens": 804711303.0, "step": 21191 }, { "epoch": 2.6958402238900905, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.375507116317749, "learning_rate": 1e-05, "loss": 0.4984, "mean_token_accuracy": 0.8554420471191406, "num_tokens": 804749016.0, "step": 21192 }, { "epoch": 2.695967434168681, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.193721771240234, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8722198009490967, "num_tokens": 804789600.0, "step": 21193 }, { "epoch": 2.696094644447271, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.527229070663452, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8843904733657837, "num_tokens": 804829033.0, "step": 21194 }, { "epoch": 2.696221854725862, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.171410083770752, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8697733283042908, "num_tokens": 804865114.0, "step": 21195 }, { "epoch": 2.696349065004452, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.6665444374084473, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8646823167800903, "num_tokens": 804907604.0, "step": 21196 }, { "epoch": 2.696476275283043, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.000814914703369, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8724270462989807, "num_tokens": 804942361.0, "step": 21197 }, { "epoch": 2.696603485561633, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.7038938999176025, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8801109790802002, "num_tokens": 804980452.0, "step": 21198 }, { "epoch": 2.6967306958402237, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.783147096633911, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8837980628013611, "num_tokens": 805017969.0, "step": 21199 }, { "epoch": 2.6968579061188143, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6864113807678223, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8855588436126709, "num_tokens": 805057053.0, "step": 21200 }, { "epoch": 2.696985116397405, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.599337577819824, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8813093900680542, "num_tokens": 805098928.0, "step": 21201 }, { "epoch": 2.6971123266759953, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6817853450775146, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.881148099899292, "num_tokens": 805140729.0, "step": 21202 }, { "epoch": 2.697239536954586, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4732327461242676, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8833605051040649, "num_tokens": 805177150.0, "step": 21203 }, { "epoch": 2.6973667472331764, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6921029090881348, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8793889284133911, "num_tokens": 805221233.0, "step": 21204 }, { "epoch": 2.697493957511767, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.380373954772949, "learning_rate": 1e-05, "loss": 0.3923, "mean_token_accuracy": 0.8916257619857788, "num_tokens": 805255075.0, "step": 21205 }, { "epoch": 2.6976211677903574, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.739946126937866, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8742567300796509, "num_tokens": 805292468.0, "step": 21206 }, { "epoch": 2.697748378068948, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.300802707672119, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8809017539024353, "num_tokens": 805330648.0, "step": 21207 }, { "epoch": 2.6978755883475385, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8465094566345215, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8800176382064819, "num_tokens": 805368137.0, "step": 21208 }, { "epoch": 2.698002798626129, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.262115716934204, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.8806288242340088, "num_tokens": 805399898.0, "step": 21209 }, { "epoch": 2.6981300089047195, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.995680332183838, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8857731819152832, "num_tokens": 805439292.0, "step": 21210 }, { "epoch": 2.69825721918331, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2429606914520264, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8650000095367432, "num_tokens": 805479464.0, "step": 21211 }, { "epoch": 2.6983844294619006, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.190544128417969, "learning_rate": 1e-05, "loss": 0.3691, "mean_token_accuracy": 0.9027807116508484, "num_tokens": 805518844.0, "step": 21212 }, { "epoch": 2.698511639740491, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.363473892211914, "learning_rate": 1e-05, "loss": 0.3991, "mean_token_accuracy": 0.8893908858299255, "num_tokens": 805552192.0, "step": 21213 }, { "epoch": 2.6986388500190817, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.327857494354248, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8838821649551392, "num_tokens": 805587343.0, "step": 21214 }, { "epoch": 2.698766060297672, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.5864486694335938, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8824506998062134, "num_tokens": 805623256.0, "step": 21215 }, { "epoch": 2.6988932705762627, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.269083023071289, "learning_rate": 1e-05, "loss": 0.4557, "mean_token_accuracy": 0.8787116408348083, "num_tokens": 805664627.0, "step": 21216 }, { "epoch": 2.699020480854853, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7512898445129395, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8721902966499329, "num_tokens": 805698808.0, "step": 21217 }, { "epoch": 2.6991476911334438, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.069304466247559, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.881198525428772, "num_tokens": 805738468.0, "step": 21218 }, { "epoch": 2.699274901412034, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.753769636154175, "learning_rate": 1e-05, "loss": 0.4781, "mean_token_accuracy": 0.8663909435272217, "num_tokens": 805776233.0, "step": 21219 }, { "epoch": 2.699402111690625, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.859095335006714, "learning_rate": 1e-05, "loss": 0.3942, "mean_token_accuracy": 0.8957844972610474, "num_tokens": 805818459.0, "step": 21220 }, { "epoch": 2.699529321969215, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.6912307739257812, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8875644207000732, "num_tokens": 805858340.0, "step": 21221 }, { "epoch": 2.699656532247806, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7084224224090576, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8764094114303589, "num_tokens": 805895920.0, "step": 21222 }, { "epoch": 2.699783742526396, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.630197525024414, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8826001882553101, "num_tokens": 805937692.0, "step": 21223 }, { "epoch": 2.6999109528049865, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6143124103546143, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8743764162063599, "num_tokens": 805974573.0, "step": 21224 }, { "epoch": 2.700038163083577, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.568551540374756, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8809516429901123, "num_tokens": 806013643.0, "step": 21225 }, { "epoch": 2.7001653733621676, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.568551778793335, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8817114233970642, "num_tokens": 806048027.0, "step": 21226 }, { "epoch": 2.700292583640758, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5207948684692383, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8697507977485657, "num_tokens": 806087269.0, "step": 21227 }, { "epoch": 2.7004197939193486, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.540578603744507, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8722681403160095, "num_tokens": 806126924.0, "step": 21228 }, { "epoch": 2.700547004197939, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4794557094573975, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8795227408409119, "num_tokens": 806162510.0, "step": 21229 }, { "epoch": 2.7006742144765297, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5357754230499268, "learning_rate": 1e-05, "loss": 0.4001, "mean_token_accuracy": 0.8886328339576721, "num_tokens": 806196519.0, "step": 21230 }, { "epoch": 2.70080142475512, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.445706367492676, "learning_rate": 1e-05, "loss": 0.4691, "mean_token_accuracy": 0.8657658696174622, "num_tokens": 806233995.0, "step": 21231 }, { "epoch": 2.7009286350337107, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5495049953460693, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8808824419975281, "num_tokens": 806268660.0, "step": 21232 }, { "epoch": 2.7010558453123013, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4007058143615723, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8702841997146606, "num_tokens": 806304187.0, "step": 21233 }, { "epoch": 2.701183055590892, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.58760404586792, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8769997954368591, "num_tokens": 806341524.0, "step": 21234 }, { "epoch": 2.7013102658694823, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3430702686309814, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8691800832748413, "num_tokens": 806384983.0, "step": 21235 }, { "epoch": 2.701437476148073, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6537013053894043, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8820294141769409, "num_tokens": 806429923.0, "step": 21236 }, { "epoch": 2.7015646864266634, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.27545166015625, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8870593309402466, "num_tokens": 806467562.0, "step": 21237 }, { "epoch": 2.701691896705254, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.786046028137207, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8782293200492859, "num_tokens": 806508634.0, "step": 21238 }, { "epoch": 2.7018191069838444, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.206674337387085, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8705713748931885, "num_tokens": 806547972.0, "step": 21239 }, { "epoch": 2.701946317262435, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9956085681915283, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8717181086540222, "num_tokens": 806591461.0, "step": 21240 }, { "epoch": 2.7020735275410255, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.177586793899536, "learning_rate": 1e-05, "loss": 0.389, "mean_token_accuracy": 0.8923256397247314, "num_tokens": 806623276.0, "step": 21241 }, { "epoch": 2.7022007378196156, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2818217277526855, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8840566873550415, "num_tokens": 806657732.0, "step": 21242 }, { "epoch": 2.7023279480982065, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.310746908187866, "learning_rate": 1e-05, "loss": 0.4001, "mean_token_accuracy": 0.886957049369812, "num_tokens": 806695018.0, "step": 21243 }, { "epoch": 2.7024551583767966, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.5037617683410645, "learning_rate": 1e-05, "loss": 0.3978, "mean_token_accuracy": 0.8973000049591064, "num_tokens": 806735344.0, "step": 21244 }, { "epoch": 2.7025823686553876, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.6695163249969482, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8780271410942078, "num_tokens": 806776854.0, "step": 21245 }, { "epoch": 2.7027095789339777, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.453132629394531, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8853409886360168, "num_tokens": 806811594.0, "step": 21246 }, { "epoch": 2.7028367892125686, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.9147238731384277, "learning_rate": 1e-05, "loss": 0.4773, "mean_token_accuracy": 0.8719128370285034, "num_tokens": 806845249.0, "step": 21247 }, { "epoch": 2.7029639994911587, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.199756622314453, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8737428188323975, "num_tokens": 806885110.0, "step": 21248 }, { "epoch": 2.7030912097697493, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.872281074523926, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8799816966056824, "num_tokens": 806924809.0, "step": 21249 }, { "epoch": 2.70321842004834, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.9881961345672607, "learning_rate": 1e-05, "loss": 0.3916, "mean_token_accuracy": 0.8953708410263062, "num_tokens": 806962386.0, "step": 21250 }, { "epoch": 2.7033456303269303, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7486143112182617, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.874732255935669, "num_tokens": 806998754.0, "step": 21251 }, { "epoch": 2.703472840605521, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8683695793151855, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8811807036399841, "num_tokens": 807036140.0, "step": 21252 }, { "epoch": 2.7036000508841114, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6349127292633057, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8751460909843445, "num_tokens": 807077144.0, "step": 21253 }, { "epoch": 2.703727261162702, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7886202335357666, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8767175674438477, "num_tokens": 807114928.0, "step": 21254 }, { "epoch": 2.7038544714412924, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.518005132675171, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8838505744934082, "num_tokens": 807151067.0, "step": 21255 }, { "epoch": 2.703981681719883, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.800960063934326, "learning_rate": 1e-05, "loss": 0.4644, "mean_token_accuracy": 0.8700785040855408, "num_tokens": 807185992.0, "step": 21256 }, { "epoch": 2.7041088919984735, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.402604818344116, "learning_rate": 1e-05, "loss": 0.3956, "mean_token_accuracy": 0.8891780376434326, "num_tokens": 807229122.0, "step": 21257 }, { "epoch": 2.704236102277064, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.883876323699951, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8730950355529785, "num_tokens": 807264118.0, "step": 21258 }, { "epoch": 2.7043633125556545, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3457274436950684, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8833721280097961, "num_tokens": 807298087.0, "step": 21259 }, { "epoch": 2.704490522834245, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.019705772399902, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8810619115829468, "num_tokens": 807338367.0, "step": 21260 }, { "epoch": 2.7046177331128356, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.3340046405792236, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8888506889343262, "num_tokens": 807376858.0, "step": 21261 }, { "epoch": 2.704744943391426, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.185012340545654, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8688691854476929, "num_tokens": 807410840.0, "step": 21262 }, { "epoch": 2.7048721536700167, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.466149091720581, "learning_rate": 1e-05, "loss": 0.385, "mean_token_accuracy": 0.894172191619873, "num_tokens": 807445128.0, "step": 21263 }, { "epoch": 2.704999363948607, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.227407455444336, "learning_rate": 1e-05, "loss": 0.4635, "mean_token_accuracy": 0.8721920251846313, "num_tokens": 807478884.0, "step": 21264 }, { "epoch": 2.7051265742271977, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6444759368896484, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8886822462081909, "num_tokens": 807517020.0, "step": 21265 }, { "epoch": 2.7052537845057882, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.085525989532471, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8656508922576904, "num_tokens": 807557037.0, "step": 21266 }, { "epoch": 2.7053809947843783, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.721717119216919, "learning_rate": 1e-05, "loss": 0.3977, "mean_token_accuracy": 0.8928191065788269, "num_tokens": 807596876.0, "step": 21267 }, { "epoch": 2.7055082050629693, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.86234450340271, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8709534406661987, "num_tokens": 807638107.0, "step": 21268 }, { "epoch": 2.7056354153415594, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.735563039779663, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8657023310661316, "num_tokens": 807678751.0, "step": 21269 }, { "epoch": 2.7057626256201504, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6646358966827393, "learning_rate": 1e-05, "loss": 0.4988, "mean_token_accuracy": 0.8578788638114929, "num_tokens": 807711447.0, "step": 21270 }, { "epoch": 2.7058898358987404, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7050955295562744, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8833003640174866, "num_tokens": 807747157.0, "step": 21271 }, { "epoch": 2.706017046177331, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4948062896728516, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8722724914550781, "num_tokens": 807782695.0, "step": 21272 }, { "epoch": 2.7061442564559215, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6995766162872314, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.88167405128479, "num_tokens": 807816850.0, "step": 21273 }, { "epoch": 2.706271466734512, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.389676809310913, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.8689444065093994, "num_tokens": 807856246.0, "step": 21274 }, { "epoch": 2.7063986770131025, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7477543354034424, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.8854866027832031, "num_tokens": 807891417.0, "step": 21275 }, { "epoch": 2.706525887291693, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.319572687149048, "learning_rate": 1e-05, "loss": 0.3994, "mean_token_accuracy": 0.8866081237792969, "num_tokens": 807924831.0, "step": 21276 }, { "epoch": 2.7066530975702836, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8340303897857666, "learning_rate": 1e-05, "loss": 0.3905, "mean_token_accuracy": 0.8945637941360474, "num_tokens": 807955427.0, "step": 21277 }, { "epoch": 2.706780307848874, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2665011882781982, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8799750208854675, "num_tokens": 807990799.0, "step": 21278 }, { "epoch": 2.7069075181274647, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9760234355926514, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8665093183517456, "num_tokens": 808027028.0, "step": 21279 }, { "epoch": 2.707034728406055, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2772185802459717, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8789167404174805, "num_tokens": 808056750.0, "step": 21280 }, { "epoch": 2.7071619386846457, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.123950004577637, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8761640787124634, "num_tokens": 808096372.0, "step": 21281 }, { "epoch": 2.7072891489632362, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3415040969848633, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8790929317474365, "num_tokens": 808136783.0, "step": 21282 }, { "epoch": 2.7074163592418268, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.249208450317383, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8786652088165283, "num_tokens": 808172310.0, "step": 21283 }, { "epoch": 2.7075435695204173, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.526097297668457, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8861209154129028, "num_tokens": 808207745.0, "step": 21284 }, { "epoch": 2.707670779799008, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.229866027832031, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.8774073123931885, "num_tokens": 808243135.0, "step": 21285 }, { "epoch": 2.7077979900775984, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6778030395507812, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8779359459877014, "num_tokens": 808280601.0, "step": 21286 }, { "epoch": 2.707925200356189, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.078343391418457, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8860302567481995, "num_tokens": 808315541.0, "step": 21287 }, { "epoch": 2.7080524106347794, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7142629623413086, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8728079795837402, "num_tokens": 808355670.0, "step": 21288 }, { "epoch": 2.70817962091337, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.888295888900757, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8853883743286133, "num_tokens": 808394189.0, "step": 21289 }, { "epoch": 2.7083068311919605, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6699888706207275, "learning_rate": 1e-05, "loss": 0.4029, "mean_token_accuracy": 0.8912690281867981, "num_tokens": 808432861.0, "step": 21290 }, { "epoch": 2.708434041470551, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7583515644073486, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8749452233314514, "num_tokens": 808462423.0, "step": 21291 }, { "epoch": 2.708561251749141, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.612841844558716, "learning_rate": 1e-05, "loss": 0.3893, "mean_token_accuracy": 0.8935196399688721, "num_tokens": 808494112.0, "step": 21292 }, { "epoch": 2.708688462027732, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6445469856262207, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8833628296852112, "num_tokens": 808538399.0, "step": 21293 }, { "epoch": 2.708815672306322, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5478322505950928, "learning_rate": 1e-05, "loss": 0.4282, "mean_token_accuracy": 0.8802292346954346, "num_tokens": 808575857.0, "step": 21294 }, { "epoch": 2.708942882584913, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.598389148712158, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8755403757095337, "num_tokens": 808615529.0, "step": 21295 }, { "epoch": 2.709070092863503, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4889976978302, "learning_rate": 1e-05, "loss": 0.3959, "mean_token_accuracy": 0.8916736841201782, "num_tokens": 808653899.0, "step": 21296 }, { "epoch": 2.7091973031420937, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.581540822982788, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8699405193328857, "num_tokens": 808697372.0, "step": 21297 }, { "epoch": 2.7093245134206843, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.440387725830078, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8763180375099182, "num_tokens": 808733854.0, "step": 21298 }, { "epoch": 2.709451723699275, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6014139652252197, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8799793720245361, "num_tokens": 808777026.0, "step": 21299 }, { "epoch": 2.7095789339778653, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.381187677383423, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8908224105834961, "num_tokens": 808814142.0, "step": 21300 }, { "epoch": 2.709706144256456, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.664214611053467, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8727434873580933, "num_tokens": 808853355.0, "step": 21301 }, { "epoch": 2.7098333545350464, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3075785636901855, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8797755241394043, "num_tokens": 808892782.0, "step": 21302 }, { "epoch": 2.709960564813637, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.770874500274658, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8715450763702393, "num_tokens": 808931935.0, "step": 21303 }, { "epoch": 2.7100877750922274, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2372987270355225, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.877836287021637, "num_tokens": 808973525.0, "step": 21304 }, { "epoch": 2.710214985370818, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.963219165802002, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8692191243171692, "num_tokens": 809008198.0, "step": 21305 }, { "epoch": 2.7103421956494085, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2209341526031494, "learning_rate": 1e-05, "loss": 0.3651, "mean_token_accuracy": 0.8974718451499939, "num_tokens": 809038267.0, "step": 21306 }, { "epoch": 2.710469405927999, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.179786205291748, "learning_rate": 1e-05, "loss": 0.4529, "mean_token_accuracy": 0.8723812103271484, "num_tokens": 809073380.0, "step": 21307 }, { "epoch": 2.7105966162065895, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.314833879470825, "learning_rate": 1e-05, "loss": 0.4044, "mean_token_accuracy": 0.8875607252120972, "num_tokens": 809114635.0, "step": 21308 }, { "epoch": 2.71072382648518, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.342212200164795, "learning_rate": 1e-05, "loss": 0.3841, "mean_token_accuracy": 0.8999108076095581, "num_tokens": 809149441.0, "step": 21309 }, { "epoch": 2.7108510367637706, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5792272090911865, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8850013017654419, "num_tokens": 809190902.0, "step": 21310 }, { "epoch": 2.710978247042361, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.287047863006592, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8837079405784607, "num_tokens": 809239297.0, "step": 21311 }, { "epoch": 2.7111054573209517, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7416625022888184, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8833256959915161, "num_tokens": 809277304.0, "step": 21312 }, { "epoch": 2.711232667599542, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1215291023254395, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8667068481445312, "num_tokens": 809313975.0, "step": 21313 }, { "epoch": 2.7113598778781327, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.75136661529541, "learning_rate": 1e-05, "loss": 0.389, "mean_token_accuracy": 0.8949411511421204, "num_tokens": 809354214.0, "step": 21314 }, { "epoch": 2.711487088156723, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9258642196655273, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.8844307065010071, "num_tokens": 809388248.0, "step": 21315 }, { "epoch": 2.7116142984353138, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.694286346435547, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8779363632202148, "num_tokens": 809428256.0, "step": 21316 }, { "epoch": 2.711741508713904, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7713205814361572, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8843472003936768, "num_tokens": 809463706.0, "step": 21317 }, { "epoch": 2.711868718992495, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6292638778686523, "learning_rate": 1e-05, "loss": 0.4492, "mean_token_accuracy": 0.8794358968734741, "num_tokens": 809503172.0, "step": 21318 }, { "epoch": 2.711995929271085, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6689188480377197, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8760985136032104, "num_tokens": 809539894.0, "step": 21319 }, { "epoch": 2.712123139549676, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5549612045288086, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8629593253135681, "num_tokens": 809582781.0, "step": 21320 }, { "epoch": 2.712250349828266, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.618025541305542, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8815523386001587, "num_tokens": 809627077.0, "step": 21321 }, { "epoch": 2.7123775601068565, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.499819278717041, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8740136027336121, "num_tokens": 809663063.0, "step": 21322 }, { "epoch": 2.712504770385447, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6077139377593994, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8653922080993652, "num_tokens": 809702386.0, "step": 21323 }, { "epoch": 2.7126319806640375, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4248111248016357, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8740665912628174, "num_tokens": 809745927.0, "step": 21324 }, { "epoch": 2.712759190942628, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6437559127807617, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8845940232276917, "num_tokens": 809777796.0, "step": 21325 }, { "epoch": 2.7128864012212186, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3573293685913086, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8869858384132385, "num_tokens": 809813450.0, "step": 21326 }, { "epoch": 2.713013611499809, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7271950244903564, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8661937713623047, "num_tokens": 809850740.0, "step": 21327 }, { "epoch": 2.7131408217783997, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.286057472229004, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8850197196006775, "num_tokens": 809886170.0, "step": 21328 }, { "epoch": 2.71326803205699, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.853911876678467, "learning_rate": 1e-05, "loss": 0.4197, "mean_token_accuracy": 0.8846110105514526, "num_tokens": 809922078.0, "step": 21329 }, { "epoch": 2.7133952423355807, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.225677490234375, "learning_rate": 1e-05, "loss": 0.4302, "mean_token_accuracy": 0.8761017918586731, "num_tokens": 809962495.0, "step": 21330 }, { "epoch": 2.7135224526141712, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.050869464874268, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8752871751785278, "num_tokens": 810002278.0, "step": 21331 }, { "epoch": 2.7136496628927618, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2575337886810303, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8807386159896851, "num_tokens": 810041236.0, "step": 21332 }, { "epoch": 2.7137768731713523, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.256066799163818, "learning_rate": 1e-05, "loss": 0.4029, "mean_token_accuracy": 0.8949693441390991, "num_tokens": 810078649.0, "step": 21333 }, { "epoch": 2.713904083449943, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.438084363937378, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8692023754119873, "num_tokens": 810118246.0, "step": 21334 }, { "epoch": 2.7140312937285334, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.309052467346191, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8923406600952148, "num_tokens": 810153407.0, "step": 21335 }, { "epoch": 2.714158504007124, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6611223220825195, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8790860772132874, "num_tokens": 810188110.0, "step": 21336 }, { "epoch": 2.7142857142857144, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.185057163238525, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8726578950881958, "num_tokens": 810225290.0, "step": 21337 }, { "epoch": 2.714412924564305, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.746131181716919, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8851534128189087, "num_tokens": 810259801.0, "step": 21338 }, { "epoch": 2.7145401348428955, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.974416494369507, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8873149752616882, "num_tokens": 810296556.0, "step": 21339 }, { "epoch": 2.7146673451214856, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7329671382904053, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.881529688835144, "num_tokens": 810332282.0, "step": 21340 }, { "epoch": 2.7147945554000765, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.785715341567993, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8810608386993408, "num_tokens": 810368017.0, "step": 21341 }, { "epoch": 2.7149217656786666, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6647210121154785, "learning_rate": 1e-05, "loss": 0.4889, "mean_token_accuracy": 0.8614152669906616, "num_tokens": 810412239.0, "step": 21342 }, { "epoch": 2.7150489759572576, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.65659499168396, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8727328181266785, "num_tokens": 810452319.0, "step": 21343 }, { "epoch": 2.7151761862358477, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6092236042022705, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8741028904914856, "num_tokens": 810490936.0, "step": 21344 }, { "epoch": 2.7153033965144386, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5658035278320312, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8799269795417786, "num_tokens": 810533565.0, "step": 21345 }, { "epoch": 2.7154306067930287, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5550312995910645, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8759430646896362, "num_tokens": 810574623.0, "step": 21346 }, { "epoch": 2.7155578170716193, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.523188829421997, "learning_rate": 1e-05, "loss": 0.4029, "mean_token_accuracy": 0.8862112760543823, "num_tokens": 810609423.0, "step": 21347 }, { "epoch": 2.71568502735021, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5329456329345703, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8843599557876587, "num_tokens": 810643814.0, "step": 21348 }, { "epoch": 2.7158122376288003, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.500962972640991, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8797205090522766, "num_tokens": 810682393.0, "step": 21349 }, { "epoch": 2.715939447907391, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4949355125427246, "learning_rate": 1e-05, "loss": 0.3898, "mean_token_accuracy": 0.891373872756958, "num_tokens": 810718604.0, "step": 21350 }, { "epoch": 2.7160666581859814, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5001819133758545, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8836469054222107, "num_tokens": 810751259.0, "step": 21351 }, { "epoch": 2.716193868464572, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4895224571228027, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8792389035224915, "num_tokens": 810781741.0, "step": 21352 }, { "epoch": 2.7163210787431624, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4828286170959473, "learning_rate": 1e-05, "loss": 0.5063, "mean_token_accuracy": 0.8530195951461792, "num_tokens": 810820258.0, "step": 21353 }, { "epoch": 2.716448289021753, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4576175212860107, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8856898546218872, "num_tokens": 810861341.0, "step": 21354 }, { "epoch": 2.7165754993003435, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4811244010925293, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8746053576469421, "num_tokens": 810897963.0, "step": 21355 }, { "epoch": 2.716702709578934, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4399001598358154, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.880994439125061, "num_tokens": 810938042.0, "step": 21356 }, { "epoch": 2.7168299198575245, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.49096417427063, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8801667094230652, "num_tokens": 810969905.0, "step": 21357 }, { "epoch": 2.716957130136115, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.421856164932251, "learning_rate": 1e-05, "loss": 0.3931, "mean_token_accuracy": 0.8871176838874817, "num_tokens": 811006818.0, "step": 21358 }, { "epoch": 2.7170843404147056, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5049030780792236, "learning_rate": 1e-05, "loss": 0.431, "mean_token_accuracy": 0.8787598609924316, "num_tokens": 811045867.0, "step": 21359 }, { "epoch": 2.717211550693296, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4021337032318115, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8846272230148315, "num_tokens": 811084207.0, "step": 21360 }, { "epoch": 2.7173387609718866, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5392794609069824, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8683193922042847, "num_tokens": 811123093.0, "step": 21361 }, { "epoch": 2.717465971250477, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3590686321258545, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8815457820892334, "num_tokens": 811157951.0, "step": 21362 }, { "epoch": 2.7175931815290677, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.597975730895996, "learning_rate": 1e-05, "loss": 0.3714, "mean_token_accuracy": 0.8992049098014832, "num_tokens": 811192743.0, "step": 21363 }, { "epoch": 2.7177203918076582, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2919225692749023, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8801746964454651, "num_tokens": 811230042.0, "step": 21364 }, { "epoch": 2.7178476020862483, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.706530809402466, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8779683113098145, "num_tokens": 811270175.0, "step": 21365 }, { "epoch": 2.7179748123648393, "ewc_loss": 0.0810546875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.219221591949463, "learning_rate": 1e-05, "loss": 0.4124, "mean_token_accuracy": 0.8828681111335754, "num_tokens": 811302926.0, "step": 21366 }, { "epoch": 2.7181020226434294, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8920977115631104, "learning_rate": 1e-05, "loss": 0.3724, "mean_token_accuracy": 0.9004868268966675, "num_tokens": 811334503.0, "step": 21367 }, { "epoch": 2.7182292329220203, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.1258561611175537, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8820756673812866, "num_tokens": 811371480.0, "step": 21368 }, { "epoch": 2.7183564432006104, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.208811283111572, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8778995275497437, "num_tokens": 811410044.0, "step": 21369 }, { "epoch": 2.718483653479201, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.1920688152313232, "learning_rate": 1e-05, "loss": 0.509, "mean_token_accuracy": 0.8523797988891602, "num_tokens": 811447815.0, "step": 21370 }, { "epoch": 2.7186108637577915, "ewc_loss": 0.1083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 7.390975952148438e-05, "grad_norm": 4.572969913482666, "learning_rate": 1e-05, "loss": 0.4779, "mean_token_accuracy": 0.8704944252967834, "num_tokens": 811484530.0, "step": 21371 }, { "epoch": 2.718738074036382, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.5649144649505615, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8842264413833618, "num_tokens": 811523010.0, "step": 21372 }, { "epoch": 2.7188652843149725, "ewc_loss": 0.109375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 7.534027099609375e-05, "grad_norm": 4.654955863952637, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8811619281768799, "num_tokens": 811559031.0, "step": 21373 }, { "epoch": 2.718992494593563, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 3.9965853691101074, "learning_rate": 1e-05, "loss": 0.5031, "mean_token_accuracy": 0.8613458871841431, "num_tokens": 811596501.0, "step": 21374 }, { "epoch": 2.7191197048721536, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.374605178833008, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8598726987838745, "num_tokens": 811634619.0, "step": 21375 }, { "epoch": 2.719246915150744, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.013609409332275, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8673363924026489, "num_tokens": 811678142.0, "step": 21376 }, { "epoch": 2.7193741254293347, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.147023677825928, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8732258081436157, "num_tokens": 811711072.0, "step": 21377 }, { "epoch": 2.719501335707925, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.866842269897461, "learning_rate": 1e-05, "loss": 0.3969, "mean_token_accuracy": 0.8944339752197266, "num_tokens": 811746372.0, "step": 21378 }, { "epoch": 2.7196285459865157, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9804956912994385, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8666218519210815, "num_tokens": 811788763.0, "step": 21379 }, { "epoch": 2.7197557562651062, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.7252612113952637, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8863123059272766, "num_tokens": 811820866.0, "step": 21380 }, { "epoch": 2.7198829665436968, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.873450517654419, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8811963796615601, "num_tokens": 811866684.0, "step": 21381 }, { "epoch": 2.7200101768222873, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5852882862091064, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8809415102005005, "num_tokens": 811902728.0, "step": 21382 }, { "epoch": 2.720137387100878, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8502025604248047, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8822334408760071, "num_tokens": 811946051.0, "step": 21383 }, { "epoch": 2.7202645973794684, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4774839878082275, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8770387172698975, "num_tokens": 811984148.0, "step": 21384 }, { "epoch": 2.720391807658059, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.908071994781494, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.8656127452850342, "num_tokens": 812020173.0, "step": 21385 }, { "epoch": 2.7205190179366494, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3844645023345947, "learning_rate": 1e-05, "loss": 0.3943, "mean_token_accuracy": 0.8883124589920044, "num_tokens": 812054509.0, "step": 21386 }, { "epoch": 2.72064622821524, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.048929214477539, "learning_rate": 1e-05, "loss": 0.4234, "mean_token_accuracy": 0.8840087652206421, "num_tokens": 812088717.0, "step": 21387 }, { "epoch": 2.7207734384938305, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.36289381980896, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8752289414405823, "num_tokens": 812126230.0, "step": 21388 }, { "epoch": 2.720900648772421, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.222224235534668, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8887953758239746, "num_tokens": 812157385.0, "step": 21389 }, { "epoch": 2.721027859051011, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.5367558002471924, "learning_rate": 1e-05, "loss": 0.5004, "mean_token_accuracy": 0.8613985776901245, "num_tokens": 812184653.0, "step": 21390 }, { "epoch": 2.721155069329602, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.217341423034668, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8903298377990723, "num_tokens": 812226784.0, "step": 21391 }, { "epoch": 2.721282279608192, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.679542303085327, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.876192569732666, "num_tokens": 812261088.0, "step": 21392 }, { "epoch": 2.721409489886783, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.073129177093506, "learning_rate": 1e-05, "loss": 0.4042, "mean_token_accuracy": 0.8912916779518127, "num_tokens": 812296859.0, "step": 21393 }, { "epoch": 2.721536700165373, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.7316665649414062, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8855308890342712, "num_tokens": 812334027.0, "step": 21394 }, { "epoch": 2.7216639104439637, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.875061273574829, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.877036452293396, "num_tokens": 812370425.0, "step": 21395 }, { "epoch": 2.7217911207225542, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7340331077575684, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8809746503829956, "num_tokens": 812402067.0, "step": 21396 }, { "epoch": 2.7219183310011448, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.658710479736328, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.8944529891014099, "num_tokens": 812436482.0, "step": 21397 }, { "epoch": 2.7220455412797353, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.689913034439087, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8653807044029236, "num_tokens": 812476059.0, "step": 21398 }, { "epoch": 2.722172751558326, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5071825981140137, "learning_rate": 1e-05, "loss": 0.3875, "mean_token_accuracy": 0.8921739459037781, "num_tokens": 812513846.0, "step": 21399 }, { "epoch": 2.7222999618369164, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6839640140533447, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8851942420005798, "num_tokens": 812550000.0, "step": 21400 }, { "epoch": 2.722427172115507, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4090702533721924, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8793438076972961, "num_tokens": 812589539.0, "step": 21401 }, { "epoch": 2.7225543823940974, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7113709449768066, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8908053636550903, "num_tokens": 812622699.0, "step": 21402 }, { "epoch": 2.722681592672688, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.327547073364258, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.8868023157119751, "num_tokens": 812662945.0, "step": 21403 }, { "epoch": 2.7228088029512785, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7943148612976074, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8835842609405518, "num_tokens": 812700650.0, "step": 21404 }, { "epoch": 2.722936013229869, "ewc_loss": 0.08154296875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.240849733352661, "learning_rate": 1e-05, "loss": 0.3865, "mean_token_accuracy": 0.8942830562591553, "num_tokens": 812742651.0, "step": 21405 }, { "epoch": 2.7230632235084595, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.95410418510437, "learning_rate": 1e-05, "loss": 0.4612, "mean_token_accuracy": 0.8714257478713989, "num_tokens": 812783792.0, "step": 21406 }, { "epoch": 2.72319043378705, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2212235927581787, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8727472424507141, "num_tokens": 812829059.0, "step": 21407 }, { "epoch": 2.7233176440656406, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.190611839294434, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8839999437332153, "num_tokens": 812859088.0, "step": 21408 }, { "epoch": 2.723444854344231, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3375134468078613, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8723627924919128, "num_tokens": 812899057.0, "step": 21409 }, { "epoch": 2.7235720646228216, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.34754753112793, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8709795475006104, "num_tokens": 812940711.0, "step": 21410 }, { "epoch": 2.723699274901412, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5782580375671387, "learning_rate": 1e-05, "loss": 0.4909, "mean_token_accuracy": 0.8598537445068359, "num_tokens": 812986967.0, "step": 21411 }, { "epoch": 2.7238264851800027, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.3086137771606445, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8802987933158875, "num_tokens": 813024396.0, "step": 21412 }, { "epoch": 2.723953695458593, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7712244987487793, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8717770576477051, "num_tokens": 813058204.0, "step": 21413 }, { "epoch": 2.7240809057371838, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.10755729675293, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8701778054237366, "num_tokens": 813098477.0, "step": 21414 }, { "epoch": 2.724208116015774, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.77305269241333, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.87726891040802, "num_tokens": 813138091.0, "step": 21415 }, { "epoch": 2.724335326294365, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9217135906219482, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8726155757904053, "num_tokens": 813170280.0, "step": 21416 }, { "epoch": 2.724462536572955, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.710742712020874, "learning_rate": 1e-05, "loss": 0.4704, "mean_token_accuracy": 0.8686252236366272, "num_tokens": 813209198.0, "step": 21417 }, { "epoch": 2.724589746851546, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7510013580322266, "learning_rate": 1e-05, "loss": 0.3673, "mean_token_accuracy": 0.9007084369659424, "num_tokens": 813248836.0, "step": 21418 }, { "epoch": 2.724716957130136, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6637344360351562, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8734303712844849, "num_tokens": 813279421.0, "step": 21419 }, { "epoch": 2.7248441674087265, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.636521577835083, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.868828296661377, "num_tokens": 813321615.0, "step": 21420 }, { "epoch": 2.724971377687317, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5873053073883057, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.880276620388031, "num_tokens": 813357338.0, "step": 21421 }, { "epoch": 2.7250985879659075, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.585646152496338, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8760814666748047, "num_tokens": 813397774.0, "step": 21422 }, { "epoch": 2.725225798244498, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.524613618850708, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8864443302154541, "num_tokens": 813432373.0, "step": 21423 }, { "epoch": 2.7253530085230886, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.553149700164795, "learning_rate": 1e-05, "loss": 0.3792, "mean_token_accuracy": 0.8968216776847839, "num_tokens": 813472354.0, "step": 21424 }, { "epoch": 2.725480218801679, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.471231698989868, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8877918124198914, "num_tokens": 813505498.0, "step": 21425 }, { "epoch": 2.7256074290802697, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5496301651000977, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8837037086486816, "num_tokens": 813546886.0, "step": 21426 }, { "epoch": 2.72573463935886, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4265964031219482, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.883881688117981, "num_tokens": 813583753.0, "step": 21427 }, { "epoch": 2.7258618496374507, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.58648943901062, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8776289224624634, "num_tokens": 813617561.0, "step": 21428 }, { "epoch": 2.7259890599160412, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003452301025390625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.384495973587036, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8647769093513489, "num_tokens": 813652440.0, "step": 21429 }, { "epoch": 2.7261162701946318, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.624262571334839, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8854137659072876, "num_tokens": 813691118.0, "step": 21430 }, { "epoch": 2.7262434804732223, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.304321765899658, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8706322312355042, "num_tokens": 813733238.0, "step": 21431 }, { "epoch": 2.726370690751813, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7376351356506348, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8849008083343506, "num_tokens": 813770002.0, "step": 21432 }, { "epoch": 2.7264979010304033, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2295501232147217, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8768910765647888, "num_tokens": 813810778.0, "step": 21433 }, { "epoch": 2.726625111308994, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9196925163269043, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8780592679977417, "num_tokens": 813844004.0, "step": 21434 }, { "epoch": 2.7267523215875844, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.1781129837036133, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8841942548751831, "num_tokens": 813884399.0, "step": 21435 }, { "epoch": 2.726879531866175, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.184086799621582, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8893254995346069, "num_tokens": 813920454.0, "step": 21436 }, { "epoch": 2.7270067421447655, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.27030611038208, "learning_rate": 1e-05, "loss": 0.4014, "mean_token_accuracy": 0.8870909214019775, "num_tokens": 813954629.0, "step": 21437 }, { "epoch": 2.7271339524233555, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.428279876708984, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8830482363700867, "num_tokens": 813992661.0, "step": 21438 }, { "epoch": 2.7272611627019465, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.5760252475738525, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8813755512237549, "num_tokens": 814024401.0, "step": 21439 }, { "epoch": 2.7273883729805366, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.414856910705566, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8800212740898132, "num_tokens": 814064950.0, "step": 21440 }, { "epoch": 2.7275155832591276, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.8089118003845215, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8833897113800049, "num_tokens": 814109503.0, "step": 21441 }, { "epoch": 2.7276427935377177, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.221515655517578, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8751882314682007, "num_tokens": 814143981.0, "step": 21442 }, { "epoch": 2.7277700038163086, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.8386106491088867, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8886460065841675, "num_tokens": 814180781.0, "step": 21443 }, { "epoch": 2.7278972140948987, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.005086421966553, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8772215247154236, "num_tokens": 814224841.0, "step": 21444 }, { "epoch": 2.7280244243734892, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7622783184051514, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8808079957962036, "num_tokens": 814258618.0, "step": 21445 }, { "epoch": 2.7281516346520798, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8421614170074463, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8831698894500732, "num_tokens": 814297286.0, "step": 21446 }, { "epoch": 2.7282788449306703, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6592764854431152, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8797980546951294, "num_tokens": 814331540.0, "step": 21447 }, { "epoch": 2.728406055209261, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7434298992156982, "learning_rate": 1e-05, "loss": 0.4744, "mean_token_accuracy": 0.8657346963882446, "num_tokens": 814371870.0, "step": 21448 }, { "epoch": 2.7285332654878514, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.555513381958008, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8843127489089966, "num_tokens": 814406561.0, "step": 21449 }, { "epoch": 2.728660475766442, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.709130048751831, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.883022129535675, "num_tokens": 814444176.0, "step": 21450 }, { "epoch": 2.7287876860450324, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4621362686157227, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.875948429107666, "num_tokens": 814485224.0, "step": 21451 }, { "epoch": 2.728914896323623, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.737910032272339, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8854933381080627, "num_tokens": 814518006.0, "step": 21452 }, { "epoch": 2.7290421066022135, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3672168254852295, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8770017623901367, "num_tokens": 814562883.0, "step": 21453 }, { "epoch": 2.729169316880804, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8231589794158936, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8847897052764893, "num_tokens": 814604619.0, "step": 21454 }, { "epoch": 2.7292965271593945, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3031129837036133, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.8745399713516235, "num_tokens": 814641112.0, "step": 21455 }, { "epoch": 2.729423737437985, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.977766513824463, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8785644769668579, "num_tokens": 814674300.0, "step": 21456 }, { "epoch": 2.7295509477165756, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.286165475845337, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8841444253921509, "num_tokens": 814711013.0, "step": 21457 }, { "epoch": 2.729678157995166, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.149889945983887, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8857061862945557, "num_tokens": 814750789.0, "step": 21458 }, { "epoch": 2.7298053682737566, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3944971561431885, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8792978525161743, "num_tokens": 814791498.0, "step": 21459 }, { "epoch": 2.729932578552347, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.246522903442383, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8814717531204224, "num_tokens": 814829811.0, "step": 21460 }, { "epoch": 2.7300597888309377, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.598931312561035, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8787179589271545, "num_tokens": 814864878.0, "step": 21461 }, { "epoch": 2.7301869991095282, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.1472320556640625, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8845336437225342, "num_tokens": 814903843.0, "step": 21462 }, { "epoch": 2.7303142093881183, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.712556838989258, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8844238519668579, "num_tokens": 814943076.0, "step": 21463 }, { "epoch": 2.7304414196667093, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9392712116241455, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.887222170829773, "num_tokens": 814981257.0, "step": 21464 }, { "epoch": 2.7305686299452994, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.714456796646118, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8868668675422668, "num_tokens": 815019763.0, "step": 21465 }, { "epoch": 2.7306958402238903, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7324090003967285, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8807491064071655, "num_tokens": 815061444.0, "step": 21466 }, { "epoch": 2.7308230505024804, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6830966472625732, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8762655258178711, "num_tokens": 815105141.0, "step": 21467 }, { "epoch": 2.730950260781071, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5777432918548584, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8770701289176941, "num_tokens": 815142839.0, "step": 21468 }, { "epoch": 2.7310774710596615, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6689059734344482, "learning_rate": 1e-05, "loss": 0.4038, "mean_token_accuracy": 0.8883253335952759, "num_tokens": 815183077.0, "step": 21469 }, { "epoch": 2.731204681338252, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4594361782073975, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8806205987930298, "num_tokens": 815222601.0, "step": 21470 }, { "epoch": 2.7313318916168425, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.677105188369751, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8751634359359741, "num_tokens": 815264587.0, "step": 21471 }, { "epoch": 2.731459101895433, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3803532123565674, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8780301213264465, "num_tokens": 815302575.0, "step": 21472 }, { "epoch": 2.7315863121740236, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.723276138305664, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8797508478164673, "num_tokens": 815342230.0, "step": 21473 }, { "epoch": 2.731713522452614, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3232321739196777, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8869763612747192, "num_tokens": 815375906.0, "step": 21474 }, { "epoch": 2.7318407327312046, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8084070682525635, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8757719397544861, "num_tokens": 815415766.0, "step": 21475 }, { "epoch": 2.731967943009795, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2639830112457275, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.886070728302002, "num_tokens": 815449979.0, "step": 21476 }, { "epoch": 2.7320951532883857, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9650962352752686, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8746320009231567, "num_tokens": 815489736.0, "step": 21477 }, { "epoch": 2.7322223635669762, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.23363995552063, "learning_rate": 1e-05, "loss": 0.394, "mean_token_accuracy": 0.8879146575927734, "num_tokens": 815525713.0, "step": 21478 }, { "epoch": 2.7323495738455668, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.168998718261719, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8746744394302368, "num_tokens": 815569231.0, "step": 21479 }, { "epoch": 2.7324767841241573, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3318707942962646, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8732656240463257, "num_tokens": 815607958.0, "step": 21480 }, { "epoch": 2.732603994402748, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.3264241218566895, "learning_rate": 1e-05, "loss": 0.4655, "mean_token_accuracy": 0.8704694509506226, "num_tokens": 815645919.0, "step": 21481 }, { "epoch": 2.7327312046813383, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.549536943435669, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8736908435821533, "num_tokens": 815688568.0, "step": 21482 }, { "epoch": 2.732858414959929, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.322005748748779, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8725165724754333, "num_tokens": 815723130.0, "step": 21483 }, { "epoch": 2.7329856252385194, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7376060485839844, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8771041035652161, "num_tokens": 815764806.0, "step": 21484 }, { "epoch": 2.73311283551711, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.143172264099121, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8706819415092468, "num_tokens": 815800226.0, "step": 21485 }, { "epoch": 2.7332400457957005, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.775425910949707, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8897566795349121, "num_tokens": 815833625.0, "step": 21486 }, { "epoch": 2.733367256074291, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9366188049316406, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8865694999694824, "num_tokens": 815870617.0, "step": 21487 }, { "epoch": 2.733494466352881, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.7232041358947754, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8745667934417725, "num_tokens": 815910854.0, "step": 21488 }, { "epoch": 2.733621676631472, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7733068466186523, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8844010233879089, "num_tokens": 815946467.0, "step": 21489 }, { "epoch": 2.733748886910062, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.661803722381592, "learning_rate": 1e-05, "loss": 0.4911, "mean_token_accuracy": 0.8628913164138794, "num_tokens": 815986924.0, "step": 21490 }, { "epoch": 2.733876097188653, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6551551818847656, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.886502206325531, "num_tokens": 816023960.0, "step": 21491 }, { "epoch": 2.734003307467243, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5781795978546143, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.875826895236969, "num_tokens": 816063911.0, "step": 21492 }, { "epoch": 2.7341305177458337, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6114416122436523, "learning_rate": 1e-05, "loss": 0.5001, "mean_token_accuracy": 0.8585945963859558, "num_tokens": 816099979.0, "step": 21493 }, { "epoch": 2.7342577280244242, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5173392295837402, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.867933452129364, "num_tokens": 816135226.0, "step": 21494 }, { "epoch": 2.7343849383030148, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.58380389213562, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8717799782752991, "num_tokens": 816174031.0, "step": 21495 }, { "epoch": 2.7345121485816053, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4466848373413086, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8852092027664185, "num_tokens": 816211194.0, "step": 21496 }, { "epoch": 2.734639358860196, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5980746746063232, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8809993267059326, "num_tokens": 816253849.0, "step": 21497 }, { "epoch": 2.7347665691387864, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4043257236480713, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.878520131111145, "num_tokens": 816290081.0, "step": 21498 }, { "epoch": 2.734893779417377, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6354258060455322, "learning_rate": 1e-05, "loss": 0.3671, "mean_token_accuracy": 0.898023247718811, "num_tokens": 816324818.0, "step": 21499 }, { "epoch": 2.7350209896959674, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3208401203155518, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8773480653762817, "num_tokens": 816369946.0, "step": 21500 }, { "epoch": 2.735148199974558, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7550346851348877, "learning_rate": 1e-05, "loss": 0.481, "mean_token_accuracy": 0.8660449981689453, "num_tokens": 816405541.0, "step": 21501 }, { "epoch": 2.7352754102531485, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.252434492111206, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8746967315673828, "num_tokens": 816441232.0, "step": 21502 }, { "epoch": 2.735402620531739, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.925576686859131, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8753882646560669, "num_tokens": 816474538.0, "step": 21503 }, { "epoch": 2.7355298308103295, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2097249031066895, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8744330406188965, "num_tokens": 816512706.0, "step": 21504 }, { "epoch": 2.73565704108892, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.160923957824707, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8880960941314697, "num_tokens": 816548638.0, "step": 21505 }, { "epoch": 2.7357842513675106, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2897801399230957, "learning_rate": 1e-05, "loss": 0.4523, "mean_token_accuracy": 0.8724299073219299, "num_tokens": 816583721.0, "step": 21506 }, { "epoch": 2.735911461646101, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.361246109008789, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8812921643257141, "num_tokens": 816621873.0, "step": 21507 }, { "epoch": 2.7360386719246916, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.5501530170440674, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8679218292236328, "num_tokens": 816659372.0, "step": 21508 }, { "epoch": 2.736165882203282, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.353647232055664, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8930981159210205, "num_tokens": 816696426.0, "step": 21509 }, { "epoch": 2.7362930924818727, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.758502960205078, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8829033374786377, "num_tokens": 816735344.0, "step": 21510 }, { "epoch": 2.7364203027604628, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.173161029815674, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8902696967124939, "num_tokens": 816773026.0, "step": 21511 }, { "epoch": 2.7365475130390537, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7968506813049316, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8836250305175781, "num_tokens": 816809777.0, "step": 21512 }, { "epoch": 2.736674723317644, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.974184513092041, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.866134524345398, "num_tokens": 816849018.0, "step": 21513 }, { "epoch": 2.736801933596235, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7358431816101074, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8741084933280945, "num_tokens": 816888867.0, "step": 21514 }, { "epoch": 2.736929143874825, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8050825595855713, "learning_rate": 1e-05, "loss": 0.407, "mean_token_accuracy": 0.886201024055481, "num_tokens": 816929221.0, "step": 21515 }, { "epoch": 2.737056354153416, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.646702766418457, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.885211706161499, "num_tokens": 816965628.0, "step": 21516 }, { "epoch": 2.737183564432006, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.711557149887085, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8757402300834656, "num_tokens": 816997908.0, "step": 21517 }, { "epoch": 2.7373107747105965, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.567063808441162, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8809652924537659, "num_tokens": 817034339.0, "step": 21518 }, { "epoch": 2.737437984989187, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.669802665710449, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8717379570007324, "num_tokens": 817073006.0, "step": 21519 }, { "epoch": 2.7375651952677775, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4814846515655518, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8747804760932922, "num_tokens": 817114517.0, "step": 21520 }, { "epoch": 2.737692405546368, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6655642986297607, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8830296993255615, "num_tokens": 817158249.0, "step": 21521 }, { "epoch": 2.7378196158249586, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.405410051345825, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8825654983520508, "num_tokens": 817197178.0, "step": 21522 }, { "epoch": 2.737946826103549, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.720235586166382, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8860939741134644, "num_tokens": 817231406.0, "step": 21523 }, { "epoch": 2.7380740363821396, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3480825424194336, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8814185261726379, "num_tokens": 817269984.0, "step": 21524 }, { "epoch": 2.73820124666073, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8070356845855713, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8834399580955505, "num_tokens": 817308997.0, "step": 21525 }, { "epoch": 2.7383284569393207, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.276918649673462, "learning_rate": 1e-05, "loss": 0.462, "mean_token_accuracy": 0.8700677156448364, "num_tokens": 817349329.0, "step": 21526 }, { "epoch": 2.7384556672179112, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9592196941375732, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8785485029220581, "num_tokens": 817384683.0, "step": 21527 }, { "epoch": 2.7385828774965018, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.2598061561584473, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8789265155792236, "num_tokens": 817423202.0, "step": 21528 }, { "epoch": 2.7387100877750923, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.145400524139404, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.889337420463562, "num_tokens": 817462048.0, "step": 21529 }, { "epoch": 2.738837298053683, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3568196296691895, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8745719194412231, "num_tokens": 817508820.0, "step": 21530 }, { "epoch": 2.7389645083322733, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2653937339782715, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8749231100082397, "num_tokens": 817554140.0, "step": 21531 }, { "epoch": 2.739091718610864, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.570186138153076, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8659629821777344, "num_tokens": 817601629.0, "step": 21532 }, { "epoch": 2.7392189288894544, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.202169895172119, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8733118772506714, "num_tokens": 817641900.0, "step": 21533 }, { "epoch": 2.739346139168045, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.702655553817749, "learning_rate": 1e-05, "loss": 0.4082, "mean_token_accuracy": 0.8906506299972534, "num_tokens": 817681843.0, "step": 21534 }, { "epoch": 2.7394733494466355, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.0115966796875, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8620401620864868, "num_tokens": 817728548.0, "step": 21535 }, { "epoch": 2.7396005597252255, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7395663261413574, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8842762112617493, "num_tokens": 817759087.0, "step": 21536 }, { "epoch": 2.7397277700038165, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.788555383682251, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8721511363983154, "num_tokens": 817799056.0, "step": 21537 }, { "epoch": 2.7398549802824066, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.704890012741089, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8798742890357971, "num_tokens": 817831859.0, "step": 21538 }, { "epoch": 2.7399821905609976, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.61499285697937, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8806352615356445, "num_tokens": 817869699.0, "step": 21539 }, { "epoch": 2.7401094008395877, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6774847507476807, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8700786828994751, "num_tokens": 817909218.0, "step": 21540 }, { "epoch": 2.7402366111181786, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4914474487304688, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8850370049476624, "num_tokens": 817948398.0, "step": 21541 }, { "epoch": 2.7403638213967687, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6675829887390137, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.86985182762146, "num_tokens": 817987351.0, "step": 21542 }, { "epoch": 2.7404910316753592, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4072022438049316, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.884695291519165, "num_tokens": 818031306.0, "step": 21543 }, { "epoch": 2.7406182419539498, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.698193311691284, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8820532560348511, "num_tokens": 818067658.0, "step": 21544 }, { "epoch": 2.7407454522325403, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.341515302658081, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8669894337654114, "num_tokens": 818109766.0, "step": 21545 }, { "epoch": 2.740872662511131, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.788956880569458, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8688960075378418, "num_tokens": 818144793.0, "step": 21546 }, { "epoch": 2.7409998727897213, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2764477729797363, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8727955222129822, "num_tokens": 818184140.0, "step": 21547 }, { "epoch": 2.741127083068312, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.909372091293335, "learning_rate": 1e-05, "loss": 0.4515, "mean_token_accuracy": 0.8744120597839355, "num_tokens": 818222216.0, "step": 21548 }, { "epoch": 2.7412542933469024, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2422711849212646, "learning_rate": 1e-05, "loss": 0.4028, "mean_token_accuracy": 0.8874908089637756, "num_tokens": 818256356.0, "step": 21549 }, { "epoch": 2.741381503625493, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.100740432739258, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8711678981781006, "num_tokens": 818300012.0, "step": 21550 }, { "epoch": 2.7415087139040835, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.284459114074707, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8759342432022095, "num_tokens": 818336090.0, "step": 21551 }, { "epoch": 2.741635924182674, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.278862476348877, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8796902298927307, "num_tokens": 818376475.0, "step": 21552 }, { "epoch": 2.7417631344612645, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.4626667499542236, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8873296976089478, "num_tokens": 818417483.0, "step": 21553 }, { "epoch": 2.741890344739855, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.330247402191162, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8780602812767029, "num_tokens": 818460323.0, "step": 21554 }, { "epoch": 2.7420175550184456, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6852869987487793, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.877628743648529, "num_tokens": 818496989.0, "step": 21555 }, { "epoch": 2.742144765297036, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.191215515136719, "learning_rate": 1e-05, "loss": 0.4044, "mean_token_accuracy": 0.8913078904151917, "num_tokens": 818532755.0, "step": 21556 }, { "epoch": 2.7422719755756266, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7549149990081787, "learning_rate": 1e-05, "loss": 0.4934, "mean_token_accuracy": 0.8615511655807495, "num_tokens": 818575237.0, "step": 21557 }, { "epoch": 2.742399185854217, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.998100996017456, "learning_rate": 1e-05, "loss": 0.4182, "mean_token_accuracy": 0.8858842253684998, "num_tokens": 818607865.0, "step": 21558 }, { "epoch": 2.7425263961328077, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.726283073425293, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8882473707199097, "num_tokens": 818647807.0, "step": 21559 }, { "epoch": 2.742653606411398, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.823552131652832, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8788552284240723, "num_tokens": 818686951.0, "step": 21560 }, { "epoch": 2.7427808166899883, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.667440891265869, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.875091552734375, "num_tokens": 818724813.0, "step": 21561 }, { "epoch": 2.7429080269685793, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.704165458679199, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8818050026893616, "num_tokens": 818757865.0, "step": 21562 }, { "epoch": 2.7430352372471694, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5890355110168457, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.8692470788955688, "num_tokens": 818796994.0, "step": 21563 }, { "epoch": 2.7431624475257603, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6223654747009277, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8853309154510498, "num_tokens": 818836408.0, "step": 21564 }, { "epoch": 2.7432896578043504, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.522141218185425, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8801419734954834, "num_tokens": 818876134.0, "step": 21565 }, { "epoch": 2.743416868082941, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.607999086380005, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8800333738327026, "num_tokens": 818911357.0, "step": 21566 }, { "epoch": 2.7435440783615315, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.481215000152588, "learning_rate": 1e-05, "loss": 0.4287, "mean_token_accuracy": 0.8792665004730225, "num_tokens": 818942743.0, "step": 21567 }, { "epoch": 2.743671288640122, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.599431037902832, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8774752616882324, "num_tokens": 818979217.0, "step": 21568 }, { "epoch": 2.7437984989187125, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.412262201309204, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8691579103469849, "num_tokens": 819019991.0, "step": 21569 }, { "epoch": 2.743925709197303, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.63413143157959, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8742967844009399, "num_tokens": 819059264.0, "step": 21570 }, { "epoch": 2.7440529194758936, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3483381271362305, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.8791344165802002, "num_tokens": 819099649.0, "step": 21571 }, { "epoch": 2.744180129754484, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7236454486846924, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8749986886978149, "num_tokens": 819142434.0, "step": 21572 }, { "epoch": 2.7443073400330746, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.310055732727051, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8590666055679321, "num_tokens": 819183908.0, "step": 21573 }, { "epoch": 2.744434550311665, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8374783992767334, "learning_rate": 1e-05, "loss": 0.3913, "mean_token_accuracy": 0.8946981430053711, "num_tokens": 819218975.0, "step": 21574 }, { "epoch": 2.7445617605902557, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2176196575164795, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8816922903060913, "num_tokens": 819264427.0, "step": 21575 }, { "epoch": 2.7446889708688462, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.038114547729492, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8705464601516724, "num_tokens": 819304947.0, "step": 21576 }, { "epoch": 2.7448161811474368, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.228085994720459, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8866651058197021, "num_tokens": 819345540.0, "step": 21577 }, { "epoch": 2.7449433914260273, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.290583610534668, "learning_rate": 1e-05, "loss": 0.4461, "mean_token_accuracy": 0.8792486786842346, "num_tokens": 819383007.0, "step": 21578 }, { "epoch": 2.745070601704618, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.4331214427948, "learning_rate": 1e-05, "loss": 0.4742, "mean_token_accuracy": 0.8726106882095337, "num_tokens": 819415640.0, "step": 21579 }, { "epoch": 2.7451978119832083, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.373855113983154, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8682776689529419, "num_tokens": 819459497.0, "step": 21580 }, { "epoch": 2.745325022261799, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6880459785461426, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8676849007606506, "num_tokens": 819497010.0, "step": 21581 }, { "epoch": 2.7454522325403894, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.238814830780029, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8799086213111877, "num_tokens": 819535659.0, "step": 21582 }, { "epoch": 2.74557944281898, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7824580669403076, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8784137964248657, "num_tokens": 819572774.0, "step": 21583 }, { "epoch": 2.7457066530975704, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.030119895935059, "learning_rate": 1e-05, "loss": 0.3969, "mean_token_accuracy": 0.8923732042312622, "num_tokens": 819607461.0, "step": 21584 }, { "epoch": 2.745833863376161, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.749286413192749, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8848471641540527, "num_tokens": 819645213.0, "step": 21585 }, { "epoch": 2.745961073654751, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.861605167388916, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8818187117576599, "num_tokens": 819679537.0, "step": 21586 }, { "epoch": 2.746088283933342, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.672159433364868, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8838125467300415, "num_tokens": 819717414.0, "step": 21587 }, { "epoch": 2.746215494211932, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7233669757843018, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8862645626068115, "num_tokens": 819757875.0, "step": 21588 }, { "epoch": 2.746342704490523, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6034281253814697, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.875714898109436, "num_tokens": 819796002.0, "step": 21589 }, { "epoch": 2.746469914769113, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6542916297912598, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8644806146621704, "num_tokens": 819834736.0, "step": 21590 }, { "epoch": 2.7465971250477037, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.504971981048584, "learning_rate": 1e-05, "loss": 0.3917, "mean_token_accuracy": 0.8908095359802246, "num_tokens": 819877566.0, "step": 21591 }, { "epoch": 2.7467243353262942, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.632387161254883, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8766095638275146, "num_tokens": 819914843.0, "step": 21592 }, { "epoch": 2.7468515456048848, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4389662742614746, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8773595690727234, "num_tokens": 819957181.0, "step": 21593 }, { "epoch": 2.7469787558834753, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.672367572784424, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8825675249099731, "num_tokens": 819991291.0, "step": 21594 }, { "epoch": 2.747105966162066, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3771142959594727, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8746441006660461, "num_tokens": 820033096.0, "step": 21595 }, { "epoch": 2.7472331764406563, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.729240894317627, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8865136504173279, "num_tokens": 820070200.0, "step": 21596 }, { "epoch": 2.747360386719247, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.311034917831421, "learning_rate": 1e-05, "loss": 0.3901, "mean_token_accuracy": 0.8918789029121399, "num_tokens": 820100989.0, "step": 21597 }, { "epoch": 2.7474875969978374, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8282721042633057, "learning_rate": 1e-05, "loss": 0.4495, "mean_token_accuracy": 0.8752651214599609, "num_tokens": 820145052.0, "step": 21598 }, { "epoch": 2.747614807276428, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.250922918319702, "learning_rate": 1e-05, "loss": 0.389, "mean_token_accuracy": 0.8899586200714111, "num_tokens": 820181463.0, "step": 21599 }, { "epoch": 2.7477420175550185, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.011340141296387, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.8762869834899902, "num_tokens": 820219226.0, "step": 21600 }, { "epoch": 2.747869227833609, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.284170627593994, "learning_rate": 1e-05, "loss": 0.4455, "mean_token_accuracy": 0.8742197155952454, "num_tokens": 820255953.0, "step": 21601 }, { "epoch": 2.7479964381121995, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.175087928771973, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.875577986240387, "num_tokens": 820296587.0, "step": 21602 }, { "epoch": 2.74812364839079, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.3999643325805664, "learning_rate": 1e-05, "loss": 0.43, "mean_token_accuracy": 0.8826660513877869, "num_tokens": 820333815.0, "step": 21603 }, { "epoch": 2.7482508586693806, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2497148513793945, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.881157636642456, "num_tokens": 820376401.0, "step": 21604 }, { "epoch": 2.748378068947971, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.5778377056121826, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8821648359298706, "num_tokens": 820417355.0, "step": 21605 }, { "epoch": 2.7485052792265616, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.165517330169678, "learning_rate": 1e-05, "loss": 0.3635, "mean_token_accuracy": 0.9048783183097839, "num_tokens": 820456788.0, "step": 21606 }, { "epoch": 2.748632489505152, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6857407093048096, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.8930311799049377, "num_tokens": 820499767.0, "step": 21607 }, { "epoch": 2.7487596997837427, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.986144781112671, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8862868547439575, "num_tokens": 820543052.0, "step": 21608 }, { "epoch": 2.7488869100623328, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6916842460632324, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8912229537963867, "num_tokens": 820586162.0, "step": 21609 }, { "epoch": 2.7490141203409237, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.823909044265747, "learning_rate": 1e-05, "loss": 0.4732, "mean_token_accuracy": 0.8710775375366211, "num_tokens": 820626703.0, "step": 21610 }, { "epoch": 2.749141330619514, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6659293174743652, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8685693740844727, "num_tokens": 820663936.0, "step": 21611 }, { "epoch": 2.749268540898105, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.67209529876709, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8840787410736084, "num_tokens": 820703761.0, "step": 21612 }, { "epoch": 2.749395751176695, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.605233907699585, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8709224462509155, "num_tokens": 820742505.0, "step": 21613 }, { "epoch": 2.749522961455286, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.57159161567688, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.890233039855957, "num_tokens": 820784185.0, "step": 21614 }, { "epoch": 2.749650171733876, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.56829571723938, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8615759015083313, "num_tokens": 820826163.0, "step": 21615 }, { "epoch": 2.7497773820124665, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5260355472564697, "learning_rate": 1e-05, "loss": 0.4389, "mean_token_accuracy": 0.8741315603256226, "num_tokens": 820861598.0, "step": 21616 }, { "epoch": 2.749904592291057, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.535581588745117, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8749216794967651, "num_tokens": 820901201.0, "step": 21617 }, { "epoch": 2.7500318025696475, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4899232387542725, "learning_rate": 1e-05, "loss": 0.4074, "mean_token_accuracy": 0.8876779675483704, "num_tokens": 820941773.0, "step": 21618 }, { "epoch": 2.750159012848238, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5307986736297607, "learning_rate": 1e-05, "loss": 0.5005, "mean_token_accuracy": 0.8569337725639343, "num_tokens": 820984610.0, "step": 21619 }, { "epoch": 2.7502862231268286, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.456059455871582, "learning_rate": 1e-05, "loss": 0.3943, "mean_token_accuracy": 0.8894275426864624, "num_tokens": 821020814.0, "step": 21620 }, { "epoch": 2.750413433405419, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5188093185424805, "learning_rate": 1e-05, "loss": 0.3984, "mean_token_accuracy": 0.8877136707305908, "num_tokens": 821053073.0, "step": 21621 }, { "epoch": 2.7505406436840096, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.432755708694458, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8747307062149048, "num_tokens": 821088872.0, "step": 21622 }, { "epoch": 2.7506678539626, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.533902168273926, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8697478771209717, "num_tokens": 821124314.0, "step": 21623 }, { "epoch": 2.7507950642411907, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.4223666191101074, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8786309957504272, "num_tokens": 821152025.0, "step": 21624 }, { "epoch": 2.750922274519781, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.53706955909729, "learning_rate": 1e-05, "loss": 0.463, "mean_token_accuracy": 0.8675167560577393, "num_tokens": 821198138.0, "step": 21625 }, { "epoch": 2.7510494847983717, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3613812923431396, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8785470128059387, "num_tokens": 821241256.0, "step": 21626 }, { "epoch": 2.7511766950769623, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.600170850753784, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8854703903198242, "num_tokens": 821274932.0, "step": 21627 }, { "epoch": 2.751303905355553, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3099658489227295, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8802038431167603, "num_tokens": 821310769.0, "step": 21628 }, { "epoch": 2.7514311156341433, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.696010112762451, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8791633248329163, "num_tokens": 821346474.0, "step": 21629 }, { "epoch": 2.751558325912734, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2341861724853516, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8795121908187866, "num_tokens": 821385293.0, "step": 21630 }, { "epoch": 2.7516855361913244, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8760533332824707, "learning_rate": 1e-05, "loss": 0.4046, "mean_token_accuracy": 0.8893659114837646, "num_tokens": 821421161.0, "step": 21631 }, { "epoch": 2.751812746469915, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.140273094177246, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8891395330429077, "num_tokens": 821459937.0, "step": 21632 }, { "epoch": 2.7519399567485054, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.160586833953857, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8811399340629578, "num_tokens": 821503892.0, "step": 21633 }, { "epoch": 2.7520671670270955, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.1812877655029297, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8849741220474243, "num_tokens": 821537781.0, "step": 21634 }, { "epoch": 2.7521943773056865, "ewc_loss": 0.107421875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.518772125244141, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.8670051097869873, "num_tokens": 821576114.0, "step": 21635 }, { "epoch": 2.7523215875842766, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.5227978229522705, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8850029706954956, "num_tokens": 821613587.0, "step": 21636 }, { "epoch": 2.7524487978628676, "ewc_loss": 0.109375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 7.486343383789062e-05, "grad_norm": 4.610069751739502, "learning_rate": 1e-05, "loss": 0.3775, "mean_token_accuracy": 0.9032437205314636, "num_tokens": 821650956.0, "step": 21637 }, { "epoch": 2.7525760081414576, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 3.9269614219665527, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8707923889160156, "num_tokens": 821692943.0, "step": 21638 }, { "epoch": 2.7527032184200486, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.371984004974365, "learning_rate": 1e-05, "loss": 0.4948, "mean_token_accuracy": 0.8639322519302368, "num_tokens": 821735344.0, "step": 21639 }, { "epoch": 2.7528304286986387, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.997861623764038, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8809245824813843, "num_tokens": 821772910.0, "step": 21640 }, { "epoch": 2.7529576389772292, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.108545780181885, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8858903646469116, "num_tokens": 821810211.0, "step": 21641 }, { "epoch": 2.7530848492558198, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.8764586448669434, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8784493207931519, "num_tokens": 821840782.0, "step": 21642 }, { "epoch": 2.7532120595344103, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 4.048798084259033, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8761438131332397, "num_tokens": 821879160.0, "step": 21643 }, { "epoch": 2.753339269813001, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.720428943634033, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8882532715797424, "num_tokens": 821915344.0, "step": 21644 }, { "epoch": 2.7534664800915913, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.000347137451171875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.876835346221924, "learning_rate": 1e-05, "loss": 0.4901, "mean_token_accuracy": 0.8583253622055054, "num_tokens": 821955628.0, "step": 21645 }, { "epoch": 2.753593690370182, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.579310655593872, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8737616539001465, "num_tokens": 821996917.0, "step": 21646 }, { "epoch": 2.7537209006487724, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8663687705993652, "learning_rate": 1e-05, "loss": 0.4261, "mean_token_accuracy": 0.8814446926116943, "num_tokens": 822034222.0, "step": 21647 }, { "epoch": 2.753848110927363, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.485454559326172, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8758196234703064, "num_tokens": 822072350.0, "step": 21648 }, { "epoch": 2.7539753212059535, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.907106637954712, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8754070997238159, "num_tokens": 822111495.0, "step": 21649 }, { "epoch": 2.754102531484544, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3964202404022217, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8771281242370605, "num_tokens": 822148132.0, "step": 21650 }, { "epoch": 2.7542297417631345, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.024177551269531, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8804590106010437, "num_tokens": 822187440.0, "step": 21651 }, { "epoch": 2.754356952041725, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.377146005630493, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.879284143447876, "num_tokens": 822225936.0, "step": 21652 }, { "epoch": 2.7544841623203156, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.169314861297607, "learning_rate": 1e-05, "loss": 0.4573, "mean_token_accuracy": 0.8765550851821899, "num_tokens": 822266187.0, "step": 21653 }, { "epoch": 2.754611372598906, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.486609697341919, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8869681358337402, "num_tokens": 822302963.0, "step": 21654 }, { "epoch": 2.7547385828774966, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.221081733703613, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8894505500793457, "num_tokens": 822336969.0, "step": 21655 }, { "epoch": 2.754865793156087, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.648385524749756, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.878539502620697, "num_tokens": 822376847.0, "step": 21656 }, { "epoch": 2.7549930034346777, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.090099334716797, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8728615045547485, "num_tokens": 822417858.0, "step": 21657 }, { "epoch": 2.755120213713268, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7384016513824463, "learning_rate": 1e-05, "loss": 0.4772, "mean_token_accuracy": 0.8694244027137756, "num_tokens": 822459946.0, "step": 21658 }, { "epoch": 2.7552474239918583, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.873999834060669, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8728325366973877, "num_tokens": 822498093.0, "step": 21659 }, { "epoch": 2.7553746342704493, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7224631309509277, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8852808475494385, "num_tokens": 822535918.0, "step": 21660 }, { "epoch": 2.7555018445490393, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.676051616668701, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8781733512878418, "num_tokens": 822574589.0, "step": 21661 }, { "epoch": 2.7556290548276303, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.686161994934082, "learning_rate": 1e-05, "loss": 0.4068, "mean_token_accuracy": 0.8848082423210144, "num_tokens": 822612437.0, "step": 21662 }, { "epoch": 2.7557562651062204, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.52166748046875, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.880995512008667, "num_tokens": 822654834.0, "step": 21663 }, { "epoch": 2.755883475384811, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6859869956970215, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8876550197601318, "num_tokens": 822689555.0, "step": 21664 }, { "epoch": 2.7560106856634015, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4151699542999268, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8636698722839355, "num_tokens": 822731609.0, "step": 21665 }, { "epoch": 2.756137895941992, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.721907377243042, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8710799217224121, "num_tokens": 822769881.0, "step": 21666 }, { "epoch": 2.7562651062205825, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.326984167098999, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.866936445236206, "num_tokens": 822810369.0, "step": 21667 }, { "epoch": 2.756392316499173, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.798405885696411, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8758609294891357, "num_tokens": 822852476.0, "step": 21668 }, { "epoch": 2.7565195267777636, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2675509452819824, "learning_rate": 1e-05, "loss": 0.4794, "mean_token_accuracy": 0.8644556999206543, "num_tokens": 822890745.0, "step": 21669 }, { "epoch": 2.756646737056354, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9298315048217773, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8780969381332397, "num_tokens": 822931243.0, "step": 21670 }, { "epoch": 2.7567739473349446, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.240919351577759, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8776544332504272, "num_tokens": 822967560.0, "step": 21671 }, { "epoch": 2.756901157613535, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.126436233520508, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8817709684371948, "num_tokens": 823007129.0, "step": 21672 }, { "epoch": 2.7570283678921257, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3226208686828613, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8742440938949585, "num_tokens": 823043722.0, "step": 21673 }, { "epoch": 2.757155578170716, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.29153299331665, "learning_rate": 1e-05, "loss": 0.5095, "mean_token_accuracy": 0.860926628112793, "num_tokens": 823079937.0, "step": 21674 }, { "epoch": 2.7572827884493067, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.53627610206604, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8753218650817871, "num_tokens": 823120455.0, "step": 21675 }, { "epoch": 2.7574099987278973, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.281569480895996, "learning_rate": 1e-05, "loss": 0.4654, "mean_token_accuracy": 0.8717678785324097, "num_tokens": 823151415.0, "step": 21676 }, { "epoch": 2.757537209006488, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.697998285293579, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8792582750320435, "num_tokens": 823188281.0, "step": 21677 }, { "epoch": 2.7576644192850783, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.124299049377441, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8834637403488159, "num_tokens": 823226040.0, "step": 21678 }, { "epoch": 2.757791629563669, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.742061138153076, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8755002021789551, "num_tokens": 823261246.0, "step": 21679 }, { "epoch": 2.7579188398422594, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9262804985046387, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8808600306510925, "num_tokens": 823300034.0, "step": 21680 }, { "epoch": 2.75804605012085, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7063369750976562, "learning_rate": 1e-05, "loss": 0.3851, "mean_token_accuracy": 0.8930398225784302, "num_tokens": 823332152.0, "step": 21681 }, { "epoch": 2.7581732603994404, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7551982402801514, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.878227710723877, "num_tokens": 823373319.0, "step": 21682 }, { "epoch": 2.758300470678031, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.638490915298462, "learning_rate": 1e-05, "loss": 0.3928, "mean_token_accuracy": 0.8939713835716248, "num_tokens": 823409875.0, "step": 21683 }, { "epoch": 2.758427680956621, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.635035276412964, "learning_rate": 1e-05, "loss": 0.3945, "mean_token_accuracy": 0.8936578631401062, "num_tokens": 823450941.0, "step": 21684 }, { "epoch": 2.758554891235212, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5757877826690674, "learning_rate": 1e-05, "loss": 0.4083, "mean_token_accuracy": 0.886058509349823, "num_tokens": 823493483.0, "step": 21685 }, { "epoch": 2.758682101513802, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.581298589706421, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8827129006385803, "num_tokens": 823531682.0, "step": 21686 }, { "epoch": 2.758809311792393, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.53440523147583, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8887234926223755, "num_tokens": 823564089.0, "step": 21687 }, { "epoch": 2.758936522070983, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5535202026367188, "learning_rate": 1e-05, "loss": 0.3737, "mean_token_accuracy": 0.8958452939987183, "num_tokens": 823598700.0, "step": 21688 }, { "epoch": 2.7590637323495737, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4766337871551514, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8743409514427185, "num_tokens": 823640248.0, "step": 21689 }, { "epoch": 2.7591909426281642, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5512921810150146, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8747673034667969, "num_tokens": 823676303.0, "step": 21690 }, { "epoch": 2.7593181529067548, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.435150623321533, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.872829020023346, "num_tokens": 823717736.0, "step": 21691 }, { "epoch": 2.7594453631853453, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5706915855407715, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8747366070747375, "num_tokens": 823755602.0, "step": 21692 }, { "epoch": 2.759572573463936, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3924667835235596, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.864343523979187, "num_tokens": 823792257.0, "step": 21693 }, { "epoch": 2.7596997837425263, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.601433753967285, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8793926239013672, "num_tokens": 823835655.0, "step": 21694 }, { "epoch": 2.759826994021117, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3357627391815186, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8726757764816284, "num_tokens": 823874840.0, "step": 21695 }, { "epoch": 2.7599542042997074, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.701181650161743, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.870173454284668, "num_tokens": 823910844.0, "step": 21696 }, { "epoch": 2.760081414578298, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2489657402038574, "learning_rate": 1e-05, "loss": 0.3874, "mean_token_accuracy": 0.8918026685714722, "num_tokens": 823948472.0, "step": 21697 }, { "epoch": 2.7602086248568884, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.855746269226074, "learning_rate": 1e-05, "loss": 0.4083, "mean_token_accuracy": 0.8898971676826477, "num_tokens": 823997532.0, "step": 21698 }, { "epoch": 2.760335835135479, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.192063570022583, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8708221316337585, "num_tokens": 824037426.0, "step": 21699 }, { "epoch": 2.7604630454140695, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.101970672607422, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8733868598937988, "num_tokens": 824076731.0, "step": 21700 }, { "epoch": 2.76059025569266, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2196428775787354, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8631324768066406, "num_tokens": 824119445.0, "step": 21701 }, { "epoch": 2.7607174659712506, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.374795436859131, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8686476349830627, "num_tokens": 824157372.0, "step": 21702 }, { "epoch": 2.760844676249841, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4356372356414795, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8777366876602173, "num_tokens": 824204312.0, "step": 21703 }, { "epoch": 2.7609718865284316, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.496893405914307, "learning_rate": 1e-05, "loss": 0.4317, "mean_token_accuracy": 0.8844780921936035, "num_tokens": 824245951.0, "step": 21704 }, { "epoch": 2.761099096807022, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.7868189811706543, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8783429861068726, "num_tokens": 824288742.0, "step": 21705 }, { "epoch": 2.7612263070856127, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.326173782348633, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.873442530632019, "num_tokens": 824326435.0, "step": 21706 }, { "epoch": 2.7613535173642028, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.903561592102051, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.878198504447937, "num_tokens": 824359442.0, "step": 21707 }, { "epoch": 2.7614807276427937, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.094759941101074, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8681319355964661, "num_tokens": 824392387.0, "step": 21708 }, { "epoch": 2.761607937921384, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.8246171474456787, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8737611770629883, "num_tokens": 824427860.0, "step": 21709 }, { "epoch": 2.761735148199975, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8943030834198, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8765757083892822, "num_tokens": 824466011.0, "step": 21710 }, { "epoch": 2.761862358478565, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.698158025741577, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8898286819458008, "num_tokens": 824507235.0, "step": 21711 }, { "epoch": 2.761989568757156, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.799467086791992, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8838123679161072, "num_tokens": 824538807.0, "step": 21712 }, { "epoch": 2.762116779035746, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.5920846462249756, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8781893253326416, "num_tokens": 824575047.0, "step": 21713 }, { "epoch": 2.7622439893143365, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7485952377319336, "learning_rate": 1e-05, "loss": 0.3764, "mean_token_accuracy": 0.8972159028053284, "num_tokens": 824604113.0, "step": 21714 }, { "epoch": 2.762371199592927, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.583167552947998, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8661004304885864, "num_tokens": 824640912.0, "step": 21715 }, { "epoch": 2.7624984098715175, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.687298059463501, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.884355366230011, "num_tokens": 824682739.0, "step": 21716 }, { "epoch": 2.762625620150108, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4506945610046387, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8609496355056763, "num_tokens": 824719973.0, "step": 21717 }, { "epoch": 2.7627528304286986, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.696568250656128, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8829482793807983, "num_tokens": 824761214.0, "step": 21718 }, { "epoch": 2.762880040707289, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3585097789764404, "learning_rate": 1e-05, "loss": 0.4541, "mean_token_accuracy": 0.8681535720825195, "num_tokens": 824801114.0, "step": 21719 }, { "epoch": 2.7630072509858796, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.81573486328125, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8731326460838318, "num_tokens": 824832734.0, "step": 21720 }, { "epoch": 2.76313446126447, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2957444190979004, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8728366494178772, "num_tokens": 824867089.0, "step": 21721 }, { "epoch": 2.7632616715430607, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9468748569488525, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8896384239196777, "num_tokens": 824909173.0, "step": 21722 }, { "epoch": 2.763388881821651, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.27510929107666, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8872942924499512, "num_tokens": 824943815.0, "step": 21723 }, { "epoch": 2.7635160921002417, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.143026351928711, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.895393431186676, "num_tokens": 824978708.0, "step": 21724 }, { "epoch": 2.7636433023788323, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.378434896469116, "learning_rate": 1e-05, "loss": 0.4992, "mean_token_accuracy": 0.8636789321899414, "num_tokens": 825015226.0, "step": 21725 }, { "epoch": 2.763770512657423, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.255967617034912, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.868873119354248, "num_tokens": 825054417.0, "step": 21726 }, { "epoch": 2.7638977229360133, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.552654266357422, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8925626277923584, "num_tokens": 825093757.0, "step": 21727 }, { "epoch": 2.764024933214604, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.221444129943848, "learning_rate": 1e-05, "loss": 0.5059, "mean_token_accuracy": 0.8626293540000916, "num_tokens": 825129955.0, "step": 21728 }, { "epoch": 2.7641521434931944, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7021093368530273, "learning_rate": 1e-05, "loss": 0.4084, "mean_token_accuracy": 0.8890905976295471, "num_tokens": 825168426.0, "step": 21729 }, { "epoch": 2.764279353771785, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.028077602386475, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8653637170791626, "num_tokens": 825213475.0, "step": 21730 }, { "epoch": 2.7644065640503754, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.726489543914795, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.889029860496521, "num_tokens": 825249146.0, "step": 21731 }, { "epoch": 2.7645337743289655, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8319473266601562, "learning_rate": 1e-05, "loss": 0.388, "mean_token_accuracy": 0.8963654637336731, "num_tokens": 825286504.0, "step": 21732 }, { "epoch": 2.7646609846075565, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6847035884857178, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8757232427597046, "num_tokens": 825327820.0, "step": 21733 }, { "epoch": 2.7647881948861466, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.667067289352417, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8898600339889526, "num_tokens": 825365967.0, "step": 21734 }, { "epoch": 2.7649154051647375, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6459596157073975, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8770526051521301, "num_tokens": 825406959.0, "step": 21735 }, { "epoch": 2.7650426154433276, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5626325607299805, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8676736950874329, "num_tokens": 825447355.0, "step": 21736 }, { "epoch": 2.7651698257219186, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6151270866394043, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8679910898208618, "num_tokens": 825485618.0, "step": 21737 }, { "epoch": 2.7652970360005087, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4869563579559326, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8826322555541992, "num_tokens": 825522287.0, "step": 21738 }, { "epoch": 2.765424246279099, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5882813930511475, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8791351318359375, "num_tokens": 825566214.0, "step": 21739 }, { "epoch": 2.7655514565576897, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4355814456939697, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8788942098617554, "num_tokens": 825608169.0, "step": 21740 }, { "epoch": 2.7656786668362803, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.605088233947754, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8845049738883972, "num_tokens": 825643076.0, "step": 21741 }, { "epoch": 2.765805877114871, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.382045269012451, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8785010576248169, "num_tokens": 825679868.0, "step": 21742 }, { "epoch": 2.7659330873934613, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6485884189605713, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8769997954368591, "num_tokens": 825719391.0, "step": 21743 }, { "epoch": 2.766060297672052, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3199429512023926, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8779692053794861, "num_tokens": 825761972.0, "step": 21744 }, { "epoch": 2.7661875079506424, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.731156349182129, "learning_rate": 1e-05, "loss": 0.4404, "mean_token_accuracy": 0.8775718808174133, "num_tokens": 825802672.0, "step": 21745 }, { "epoch": 2.766314718229233, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2606117725372314, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8751033544540405, "num_tokens": 825840219.0, "step": 21746 }, { "epoch": 2.7664419285078234, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8804056644439697, "learning_rate": 1e-05, "loss": 0.4361, "mean_token_accuracy": 0.8793811202049255, "num_tokens": 825876793.0, "step": 21747 }, { "epoch": 2.766569138786414, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.208343505859375, "learning_rate": 1e-05, "loss": 0.4499, "mean_token_accuracy": 0.8726797103881836, "num_tokens": 825913231.0, "step": 21748 }, { "epoch": 2.7666963490650045, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.099794387817383, "learning_rate": 1e-05, "loss": 0.4794, "mean_token_accuracy": 0.8666737675666809, "num_tokens": 825952600.0, "step": 21749 }, { "epoch": 2.766823559343595, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.222069263458252, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8769270777702332, "num_tokens": 825997239.0, "step": 21750 }, { "epoch": 2.7669507696221856, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.366589546203613, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8840959668159485, "num_tokens": 826037492.0, "step": 21751 }, { "epoch": 2.767077979900776, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.450392484664917, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8804084062576294, "num_tokens": 826077607.0, "step": 21752 }, { "epoch": 2.7672051901793666, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.483016014099121, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8741250038146973, "num_tokens": 826115503.0, "step": 21753 }, { "epoch": 2.767332400457957, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.7780487537384033, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8739949464797974, "num_tokens": 826154526.0, "step": 21754 }, { "epoch": 2.7674596107365477, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.3080525398254395, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8693740963935852, "num_tokens": 826194945.0, "step": 21755 }, { "epoch": 2.767586821015138, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.8769543170928955, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8827691674232483, "num_tokens": 826231000.0, "step": 21756 }, { "epoch": 2.7677140312937283, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.072714328765869, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8886756896972656, "num_tokens": 826266453.0, "step": 21757 }, { "epoch": 2.7678412415723193, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.796290636062622, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8817746639251709, "num_tokens": 826310442.0, "step": 21758 }, { "epoch": 2.7679684518509093, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.910919427871704, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8781643509864807, "num_tokens": 826344595.0, "step": 21759 }, { "epoch": 2.7680956621295003, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.705251932144165, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8793512582778931, "num_tokens": 826384556.0, "step": 21760 }, { "epoch": 2.7682228724080904, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.787156581878662, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8565284013748169, "num_tokens": 826418344.0, "step": 21761 }, { "epoch": 2.768350082686681, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.598534107208252, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8783208131790161, "num_tokens": 826460027.0, "step": 21762 }, { "epoch": 2.7684772929652715, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7076730728149414, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8773622512817383, "num_tokens": 826497422.0, "step": 21763 }, { "epoch": 2.768604503243862, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.509063720703125, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8706759810447693, "num_tokens": 826536624.0, "step": 21764 }, { "epoch": 2.7687317135224525, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.708843469619751, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8729526996612549, "num_tokens": 826572303.0, "step": 21765 }, { "epoch": 2.768858923801043, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4135072231292725, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8599294424057007, "num_tokens": 826613326.0, "step": 21766 }, { "epoch": 2.7689861340796336, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7636451721191406, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8763464689254761, "num_tokens": 826649059.0, "step": 21767 }, { "epoch": 2.769113344358224, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.339892625808716, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8823323249816895, "num_tokens": 826682983.0, "step": 21768 }, { "epoch": 2.7692405546368146, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.880195140838623, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8766215443611145, "num_tokens": 826715179.0, "step": 21769 }, { "epoch": 2.769367764915405, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3023900985717773, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8671890497207642, "num_tokens": 826748442.0, "step": 21770 }, { "epoch": 2.7694949751939957, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.019157886505127, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.886249840259552, "num_tokens": 826791014.0, "step": 21771 }, { "epoch": 2.769622185472586, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.309926986694336, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8724274635314941, "num_tokens": 826827987.0, "step": 21772 }, { "epoch": 2.7697493957511767, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.171952247619629, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8826375603675842, "num_tokens": 826869315.0, "step": 21773 }, { "epoch": 2.7698766060297673, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4374516010284424, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8760610222816467, "num_tokens": 826912679.0, "step": 21774 }, { "epoch": 2.770003816308358, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.232611179351807, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8817588090896606, "num_tokens": 826949276.0, "step": 21775 }, { "epoch": 2.7701310265869483, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.626208543777466, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8850734233856201, "num_tokens": 826989441.0, "step": 21776 }, { "epoch": 2.770258236865539, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.109570503234863, "learning_rate": 1e-05, "loss": 0.4695, "mean_token_accuracy": 0.8715449571609497, "num_tokens": 827025030.0, "step": 21777 }, { "epoch": 2.7703854471441294, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7319705486297607, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8766868114471436, "num_tokens": 827060252.0, "step": 21778 }, { "epoch": 2.77051265742272, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8848981857299805, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8791269659996033, "num_tokens": 827100063.0, "step": 21779 }, { "epoch": 2.77063986770131, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.717703104019165, "learning_rate": 1e-05, "loss": 0.4011, "mean_token_accuracy": 0.8915151357650757, "num_tokens": 827131772.0, "step": 21780 }, { "epoch": 2.770767077979901, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.687441110610962, "learning_rate": 1e-05, "loss": 0.4753, "mean_token_accuracy": 0.8670693039894104, "num_tokens": 827167659.0, "step": 21781 }, { "epoch": 2.770894288258491, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.69050669670105, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8709098100662231, "num_tokens": 827211215.0, "step": 21782 }, { "epoch": 2.771021498537082, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5379059314727783, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8698593974113464, "num_tokens": 827248075.0, "step": 21783 }, { "epoch": 2.771148708815672, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6843624114990234, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.8831503391265869, "num_tokens": 827283270.0, "step": 21784 }, { "epoch": 2.771275919094263, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.428386688232422, "learning_rate": 1e-05, "loss": 0.415, "mean_token_accuracy": 0.883586585521698, "num_tokens": 827317520.0, "step": 21785 }, { "epoch": 2.771403129372853, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7055771350860596, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.8710888624191284, "num_tokens": 827355899.0, "step": 21786 }, { "epoch": 2.7715303396514437, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.364109754562378, "learning_rate": 1e-05, "loss": 0.498, "mean_token_accuracy": 0.8589757680892944, "num_tokens": 827390038.0, "step": 21787 }, { "epoch": 2.771657549930034, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7621536254882812, "learning_rate": 1e-05, "loss": 0.4574, "mean_token_accuracy": 0.873499870300293, "num_tokens": 827428563.0, "step": 21788 }, { "epoch": 2.7717847602086247, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2842116355895996, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8857345581054688, "num_tokens": 827469658.0, "step": 21789 }, { "epoch": 2.7719119704872153, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8783552646636963, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8823121190071106, "num_tokens": 827504277.0, "step": 21790 }, { "epoch": 2.772039180765806, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.465611696243286, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8727802038192749, "num_tokens": 827544840.0, "step": 21791 }, { "epoch": 2.7721663910443963, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.836571455001831, "learning_rate": 1e-05, "loss": 0.4852, "mean_token_accuracy": 0.86516934633255, "num_tokens": 827590607.0, "step": 21792 }, { "epoch": 2.772293601322987, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2864556312561035, "learning_rate": 1e-05, "loss": 0.3717, "mean_token_accuracy": 0.8948758244514465, "num_tokens": 827622039.0, "step": 21793 }, { "epoch": 2.7724208116015774, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.912496566772461, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8853809833526611, "num_tokens": 827659234.0, "step": 21794 }, { "epoch": 2.772548021880168, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.273801326751709, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8812789916992188, "num_tokens": 827702044.0, "step": 21795 }, { "epoch": 2.7726752321587584, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.061487197875977, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8840435743331909, "num_tokens": 827738189.0, "step": 21796 }, { "epoch": 2.772802442437349, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3280866146087646, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8680068254470825, "num_tokens": 827780483.0, "step": 21797 }, { "epoch": 2.7729296527159395, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.180635452270508, "learning_rate": 1e-05, "loss": 0.4374, "mean_token_accuracy": 0.882081925868988, "num_tokens": 827817156.0, "step": 21798 }, { "epoch": 2.77305686299453, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.46539568901062, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8801783323287964, "num_tokens": 827851591.0, "step": 21799 }, { "epoch": 2.7731840732731206, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.197753429412842, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8853275775909424, "num_tokens": 827891289.0, "step": 21800 }, { "epoch": 2.773311283551711, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.591841459274292, "learning_rate": 1e-05, "loss": 0.376, "mean_token_accuracy": 0.8994658589363098, "num_tokens": 827926466.0, "step": 21801 }, { "epoch": 2.7734384938303016, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.105411052703857, "learning_rate": 1e-05, "loss": 0.512, "mean_token_accuracy": 0.8511037230491638, "num_tokens": 827964885.0, "step": 21802 }, { "epoch": 2.773565704108892, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.670579671859741, "learning_rate": 1e-05, "loss": 0.4682, "mean_token_accuracy": 0.8686211705207825, "num_tokens": 828006063.0, "step": 21803 }, { "epoch": 2.7736929143874827, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.934526205062866, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.885678768157959, "num_tokens": 828037533.0, "step": 21804 }, { "epoch": 2.7738201246660728, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.668043851852417, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8864842653274536, "num_tokens": 828072298.0, "step": 21805 }, { "epoch": 2.7739473349446637, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.761859178543091, "learning_rate": 1e-05, "loss": 0.4097, "mean_token_accuracy": 0.8857015371322632, "num_tokens": 828107900.0, "step": 21806 }, { "epoch": 2.774074545223254, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.621978759765625, "learning_rate": 1e-05, "loss": 0.4221, "mean_token_accuracy": 0.8831275105476379, "num_tokens": 828143323.0, "step": 21807 }, { "epoch": 2.774201755501845, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6339802742004395, "learning_rate": 1e-05, "loss": 0.3971, "mean_token_accuracy": 0.8938581943511963, "num_tokens": 828183439.0, "step": 21808 }, { "epoch": 2.774328965780435, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.607623815536499, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8833130598068237, "num_tokens": 828218360.0, "step": 21809 }, { "epoch": 2.774456176059026, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5525057315826416, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.872374415397644, "num_tokens": 828254042.0, "step": 21810 }, { "epoch": 2.774583386337616, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6100354194641113, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8765350580215454, "num_tokens": 828290992.0, "step": 21811 }, { "epoch": 2.7747105966162064, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.476860284805298, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8830987811088562, "num_tokens": 828323351.0, "step": 21812 }, { "epoch": 2.774837806894797, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5820553302764893, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.883382260799408, "num_tokens": 828354555.0, "step": 21813 }, { "epoch": 2.7749650171733875, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4167098999023438, "learning_rate": 1e-05, "loss": 0.4201, "mean_token_accuracy": 0.8833354711532593, "num_tokens": 828395641.0, "step": 21814 }, { "epoch": 2.775092227451978, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5928540229797363, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8814912438392639, "num_tokens": 828437871.0, "step": 21815 }, { "epoch": 2.7752194377305686, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.372663974761963, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8645919561386108, "num_tokens": 828480545.0, "step": 21816 }, { "epoch": 2.775346648009159, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6436004638671875, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8810601234436035, "num_tokens": 828513519.0, "step": 21817 }, { "epoch": 2.7754738582877496, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3191988468170166, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8733025789260864, "num_tokens": 828549329.0, "step": 21818 }, { "epoch": 2.77560106856634, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7212846279144287, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8661904335021973, "num_tokens": 828592357.0, "step": 21819 }, { "epoch": 2.7757282788449307, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.261296033859253, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.87785404920578, "num_tokens": 828627622.0, "step": 21820 }, { "epoch": 2.775855489123521, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.847250461578369, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8895821571350098, "num_tokens": 828668680.0, "step": 21821 }, { "epoch": 2.7759826994021117, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.1941304206848145, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8774412870407104, "num_tokens": 828705233.0, "step": 21822 }, { "epoch": 2.7761099096807023, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.0813446044921875, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8696404695510864, "num_tokens": 828745183.0, "step": 21823 }, { "epoch": 2.776237119959293, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.196303367614746, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8736610412597656, "num_tokens": 828787769.0, "step": 21824 }, { "epoch": 2.7763643302378833, "ewc_loss": 0.10595703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.375580310821533, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.875218391418457, "num_tokens": 828826511.0, "step": 21825 }, { "epoch": 2.776491540516474, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.424499988555908, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8786116242408752, "num_tokens": 828868632.0, "step": 21826 }, { "epoch": 2.7766187507950644, "ewc_loss": 0.10791015625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.5148091316223145, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8697258830070496, "num_tokens": 828904342.0, "step": 21827 }, { "epoch": 2.776745961073655, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.8767807483673096, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8749757409095764, "num_tokens": 828947074.0, "step": 21828 }, { "epoch": 2.7768731713522454, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.260756015777588, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.88228440284729, "num_tokens": 828989583.0, "step": 21829 }, { "epoch": 2.7770003816308355, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.840562343597412, "learning_rate": 1e-05, "loss": 0.5256, "mean_token_accuracy": 0.8503909111022949, "num_tokens": 829028848.0, "step": 21830 }, { "epoch": 2.7771275919094265, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.068341255187988, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8783993124961853, "num_tokens": 829063904.0, "step": 21831 }, { "epoch": 2.7772548021880166, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.76986026763916, "learning_rate": 1e-05, "loss": 0.4844, "mean_token_accuracy": 0.8650836944580078, "num_tokens": 829102416.0, "step": 21832 }, { "epoch": 2.7773820124666075, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.911106824874878, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8730480670928955, "num_tokens": 829138870.0, "step": 21833 }, { "epoch": 2.7775092227451976, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.6734848022460938, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8710640072822571, "num_tokens": 829176704.0, "step": 21834 }, { "epoch": 2.7776364330237886, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8084731101989746, "learning_rate": 1e-05, "loss": 0.4165, "mean_token_accuracy": 0.8846179246902466, "num_tokens": 829210170.0, "step": 21835 }, { "epoch": 2.7777636433023787, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.558645009994507, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8920173645019531, "num_tokens": 829249692.0, "step": 21836 }, { "epoch": 2.777890853580969, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7792766094207764, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.8870854377746582, "num_tokens": 829285342.0, "step": 21837 }, { "epoch": 2.7780180638595597, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.476283311843872, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.879325270652771, "num_tokens": 829320281.0, "step": 21838 }, { "epoch": 2.7781452741381503, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.800187826156616, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8729194402694702, "num_tokens": 829359990.0, "step": 21839 }, { "epoch": 2.778272484416741, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.397164821624756, "learning_rate": 1e-05, "loss": 0.4008, "mean_token_accuracy": 0.8882631063461304, "num_tokens": 829395330.0, "step": 21840 }, { "epoch": 2.7783996946953313, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8780431747436523, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8762360215187073, "num_tokens": 829432127.0, "step": 21841 }, { "epoch": 2.778526904973922, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3466310501098633, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8749080896377563, "num_tokens": 829469735.0, "step": 21842 }, { "epoch": 2.7786541152525124, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.992954730987549, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8771257400512695, "num_tokens": 829502893.0, "step": 21843 }, { "epoch": 2.778781325531103, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3455379009246826, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8844833374023438, "num_tokens": 829538307.0, "step": 21844 }, { "epoch": 2.7789085358096934, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.1229753494262695, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8873904943466187, "num_tokens": 829575992.0, "step": 21845 }, { "epoch": 2.779035746088284, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.4372665882110596, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8712337017059326, "num_tokens": 829610167.0, "step": 21846 }, { "epoch": 2.7791629563668745, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1720452308654785, "learning_rate": 1e-05, "loss": 0.4752, "mean_token_accuracy": 0.8697929382324219, "num_tokens": 829646828.0, "step": 21847 }, { "epoch": 2.779290166645465, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.5905306339263916, "learning_rate": 1e-05, "loss": 0.4697, "mean_token_accuracy": 0.8700277805328369, "num_tokens": 829683593.0, "step": 21848 }, { "epoch": 2.7794173769240555, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.055841445922852, "learning_rate": 1e-05, "loss": 0.445, "mean_token_accuracy": 0.8752634525299072, "num_tokens": 829721080.0, "step": 21849 }, { "epoch": 2.779544587202646, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.665678024291992, "learning_rate": 1e-05, "loss": 0.3929, "mean_token_accuracy": 0.892979621887207, "num_tokens": 829755871.0, "step": 21850 }, { "epoch": 2.7796717974812366, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.871225118637085, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8798351287841797, "num_tokens": 829793377.0, "step": 21851 }, { "epoch": 2.779799007759827, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6704657077789307, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8758458495140076, "num_tokens": 829833165.0, "step": 21852 }, { "epoch": 2.7799262180384177, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.691598892211914, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.882070004940033, "num_tokens": 829869558.0, "step": 21853 }, { "epoch": 2.780053428317008, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6619770526885986, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8770102262496948, "num_tokens": 829906385.0, "step": 21854 }, { "epoch": 2.7801806385955983, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5399930477142334, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.868755042552948, "num_tokens": 829944216.0, "step": 21855 }, { "epoch": 2.7803078488741892, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.67246150970459, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8570377826690674, "num_tokens": 829976817.0, "step": 21856 }, { "epoch": 2.7804350591527793, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4188945293426514, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8705881237983704, "num_tokens": 830020434.0, "step": 21857 }, { "epoch": 2.7805622694313703, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.700714588165283, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8799487948417664, "num_tokens": 830054469.0, "step": 21858 }, { "epoch": 2.7806894797099604, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3339290618896484, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8789487481117249, "num_tokens": 830091536.0, "step": 21859 }, { "epoch": 2.780816689988551, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7779338359832764, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8704866170883179, "num_tokens": 830129425.0, "step": 21860 }, { "epoch": 2.7809439002671414, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2675654888153076, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8712862133979797, "num_tokens": 830171137.0, "step": 21861 }, { "epoch": 2.781071110545732, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9111108779907227, "learning_rate": 1e-05, "loss": 0.4084, "mean_token_accuracy": 0.8904343843460083, "num_tokens": 830211434.0, "step": 21862 }, { "epoch": 2.7811983208243225, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2229511737823486, "learning_rate": 1e-05, "loss": 0.3861, "mean_token_accuracy": 0.8919494152069092, "num_tokens": 830250843.0, "step": 21863 }, { "epoch": 2.781325531102913, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.112762451171875, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.868837833404541, "num_tokens": 830292131.0, "step": 21864 }, { "epoch": 2.7814527413815036, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.292052745819092, "learning_rate": 1e-05, "loss": 0.3975, "mean_token_accuracy": 0.8869757652282715, "num_tokens": 830329678.0, "step": 21865 }, { "epoch": 2.781579951660094, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.290255546569824, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8865898847579956, "num_tokens": 830367822.0, "step": 21866 }, { "epoch": 2.7817071619386846, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5005276203155518, "learning_rate": 1e-05, "loss": 0.4576, "mean_token_accuracy": 0.8748695850372314, "num_tokens": 830408170.0, "step": 21867 }, { "epoch": 2.781834372217275, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.302780628204346, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8834514617919922, "num_tokens": 830447402.0, "step": 21868 }, { "epoch": 2.7819615824958657, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6970698833465576, "learning_rate": 1e-05, "loss": 0.513, "mean_token_accuracy": 0.8572199940681458, "num_tokens": 830485411.0, "step": 21869 }, { "epoch": 2.782088792774456, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.148716926574707, "learning_rate": 1e-05, "loss": 0.4464, "mean_token_accuracy": 0.8789342641830444, "num_tokens": 830522562.0, "step": 21870 }, { "epoch": 2.7822160030530467, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.738858938217163, "learning_rate": 1e-05, "loss": 0.4786, "mean_token_accuracy": 0.8665745258331299, "num_tokens": 830565979.0, "step": 21871 }, { "epoch": 2.7823432133316373, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9680919647216797, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8749386072158813, "num_tokens": 830602771.0, "step": 21872 }, { "epoch": 2.782470423610228, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.7009897232055664, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.868580162525177, "num_tokens": 830639891.0, "step": 21873 }, { "epoch": 2.7825976338888183, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.798428773880005, "learning_rate": 1e-05, "loss": 0.389, "mean_token_accuracy": 0.8891254663467407, "num_tokens": 830676044.0, "step": 21874 }, { "epoch": 2.782724844167409, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6322996616363525, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8764606714248657, "num_tokens": 830718252.0, "step": 21875 }, { "epoch": 2.7828520544459994, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6920456886291504, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.8895276188850403, "num_tokens": 830757563.0, "step": 21876 }, { "epoch": 2.78297926472459, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.558429718017578, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8731256723403931, "num_tokens": 830797808.0, "step": 21877 }, { "epoch": 2.78310647500318, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.641042709350586, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8753147721290588, "num_tokens": 830838691.0, "step": 21878 }, { "epoch": 2.783233685281771, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4925825595855713, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8906745910644531, "num_tokens": 830877016.0, "step": 21879 }, { "epoch": 2.783360895560361, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6469130516052246, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8812070488929749, "num_tokens": 830908841.0, "step": 21880 }, { "epoch": 2.783488105838952, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4352312088012695, "learning_rate": 1e-05, "loss": 0.4542, "mean_token_accuracy": 0.8738881945610046, "num_tokens": 830947580.0, "step": 21881 }, { "epoch": 2.783615316117542, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.683328628540039, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.877890944480896, "num_tokens": 830984385.0, "step": 21882 }, { "epoch": 2.783742526396133, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3446216583251953, "learning_rate": 1e-05, "loss": 0.4065, "mean_token_accuracy": 0.889889121055603, "num_tokens": 831022280.0, "step": 21883 }, { "epoch": 2.783869736674723, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7774581909179688, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8808627724647522, "num_tokens": 831055764.0, "step": 21884 }, { "epoch": 2.7839969469533137, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2961649894714355, "learning_rate": 1e-05, "loss": 0.4904, "mean_token_accuracy": 0.8668478727340698, "num_tokens": 831093397.0, "step": 21885 }, { "epoch": 2.784124157231904, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.9088428020477295, "learning_rate": 1e-05, "loss": 0.4408, "mean_token_accuracy": 0.8784584403038025, "num_tokens": 831130720.0, "step": 21886 }, { "epoch": 2.7842513675104947, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.264737844467163, "learning_rate": 1e-05, "loss": 0.4372, "mean_token_accuracy": 0.8786637783050537, "num_tokens": 831164092.0, "step": 21887 }, { "epoch": 2.7843785777890853, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.081711292266846, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8734444379806519, "num_tokens": 831202570.0, "step": 21888 }, { "epoch": 2.784505788067676, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.3162076473236084, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8650234341621399, "num_tokens": 831242596.0, "step": 21889 }, { "epoch": 2.7846329983462663, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.227474212646484, "learning_rate": 1e-05, "loss": 0.4095, "mean_token_accuracy": 0.8909753561019897, "num_tokens": 831283305.0, "step": 21890 }, { "epoch": 2.784760208624857, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.484109401702881, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8912732601165771, "num_tokens": 831322355.0, "step": 21891 }, { "epoch": 2.7848874189034474, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.242718696594238, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8758901953697205, "num_tokens": 831361410.0, "step": 21892 }, { "epoch": 2.785014629182038, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6647651195526123, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8717018365859985, "num_tokens": 831402493.0, "step": 21893 }, { "epoch": 2.7851418394606284, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.102931499481201, "learning_rate": 1e-05, "loss": 0.517, "mean_token_accuracy": 0.853786826133728, "num_tokens": 831443231.0, "step": 21894 }, { "epoch": 2.785269049739219, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.728187084197998, "learning_rate": 1e-05, "loss": 0.3779, "mean_token_accuracy": 0.9017413258552551, "num_tokens": 831474716.0, "step": 21895 }, { "epoch": 2.7853962600178095, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.8830060958862305, "learning_rate": 1e-05, "loss": 0.4905, "mean_token_accuracy": 0.8618985414505005, "num_tokens": 831519410.0, "step": 21896 }, { "epoch": 2.7855234702964, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6859278678894043, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8834110498428345, "num_tokens": 831562040.0, "step": 21897 }, { "epoch": 2.7856506805749905, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.70965576171875, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8699864149093628, "num_tokens": 831604956.0, "step": 21898 }, { "epoch": 2.785777890853581, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.661990165710449, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8815208673477173, "num_tokens": 831637270.0, "step": 21899 }, { "epoch": 2.7859051011321716, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5880093574523926, "learning_rate": 1e-05, "loss": 0.4091, "mean_token_accuracy": 0.8838448524475098, "num_tokens": 831671342.0, "step": 21900 }, { "epoch": 2.786032311410762, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.618665933609009, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8758875131607056, "num_tokens": 831709222.0, "step": 21901 }, { "epoch": 2.7861595216893527, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5033469200134277, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8860220313072205, "num_tokens": 831749690.0, "step": 21902 }, { "epoch": 2.7862867319679427, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.608309268951416, "learning_rate": 1e-05, "loss": 0.4898, "mean_token_accuracy": 0.8665837049484253, "num_tokens": 831787963.0, "step": 21903 }, { "epoch": 2.7864139422465337, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.452242612838745, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8855831623077393, "num_tokens": 831820925.0, "step": 21904 }, { "epoch": 2.786541152525124, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6051104068756104, "learning_rate": 1e-05, "loss": 0.4565, "mean_token_accuracy": 0.8732120394706726, "num_tokens": 831856190.0, "step": 21905 }, { "epoch": 2.7866683628037148, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.391569137573242, "learning_rate": 1e-05, "loss": 0.4698, "mean_token_accuracy": 0.8656275272369385, "num_tokens": 831900339.0, "step": 21906 }, { "epoch": 2.786795573082305, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.650728225708008, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8717122077941895, "num_tokens": 831932166.0, "step": 21907 }, { "epoch": 2.786922783360896, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.344857931137085, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8816249370574951, "num_tokens": 831964796.0, "step": 21908 }, { "epoch": 2.787049993639486, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.709552526473999, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8753124475479126, "num_tokens": 832004603.0, "step": 21909 }, { "epoch": 2.7871772039180764, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2628684043884277, "learning_rate": 1e-05, "loss": 0.3978, "mean_token_accuracy": 0.8891078233718872, "num_tokens": 832042847.0, "step": 21910 }, { "epoch": 2.787304414196667, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8465354442596436, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8797190189361572, "num_tokens": 832079746.0, "step": 21911 }, { "epoch": 2.7874316244752575, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2192015647888184, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8706697821617126, "num_tokens": 832116300.0, "step": 21912 }, { "epoch": 2.787558834753848, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.047337532043457, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8894163370132446, "num_tokens": 832151985.0, "step": 21913 }, { "epoch": 2.7876860450324386, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2319347858428955, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8775234222412109, "num_tokens": 832181327.0, "step": 21914 }, { "epoch": 2.787813255311029, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2923431396484375, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8835713267326355, "num_tokens": 832217710.0, "step": 21915 }, { "epoch": 2.7879404655896196, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.373776435852051, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.874719500541687, "num_tokens": 832255628.0, "step": 21916 }, { "epoch": 2.78806767586821, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 4.446784973144531, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8780847787857056, "num_tokens": 832294239.0, "step": 21917 }, { "epoch": 2.7881948861468007, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.6833138465881348, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8826407194137573, "num_tokens": 832331270.0, "step": 21918 }, { "epoch": 2.788322096425391, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.358290672302246, "learning_rate": 1e-05, "loss": 0.483, "mean_token_accuracy": 0.8689358234405518, "num_tokens": 832372276.0, "step": 21919 }, { "epoch": 2.7884493067039817, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.8450276851654053, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8865588903427124, "num_tokens": 832410352.0, "step": 21920 }, { "epoch": 2.7885765169825723, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.139608860015869, "learning_rate": 1e-05, "loss": 0.4845, "mean_token_accuracy": 0.8705940842628479, "num_tokens": 832452633.0, "step": 21921 }, { "epoch": 2.788703727261163, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.810666799545288, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.883642315864563, "num_tokens": 832492100.0, "step": 21922 }, { "epoch": 2.7888309375397533, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.949683666229248, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8738750219345093, "num_tokens": 832529972.0, "step": 21923 }, { "epoch": 2.788958147818344, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.7174487113952637, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8874924182891846, "num_tokens": 832571042.0, "step": 21924 }, { "epoch": 2.7890853580969344, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.813560724258423, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8703552484512329, "num_tokens": 832610575.0, "step": 21925 }, { "epoch": 2.789212568375525, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6224207878112793, "learning_rate": 1e-05, "loss": 0.3946, "mean_token_accuracy": 0.8904355764389038, "num_tokens": 832649477.0, "step": 21926 }, { "epoch": 2.7893397786541154, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7320330142974854, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.88414067029953, "num_tokens": 832688958.0, "step": 21927 }, { "epoch": 2.7894669889327055, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5315394401550293, "learning_rate": 1e-05, "loss": 0.3989, "mean_token_accuracy": 0.8905854821205139, "num_tokens": 832724368.0, "step": 21928 }, { "epoch": 2.7895941992112965, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7171528339385986, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8742080926895142, "num_tokens": 832765950.0, "step": 21929 }, { "epoch": 2.7897214094898866, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.444817304611206, "learning_rate": 1e-05, "loss": 0.3768, "mean_token_accuracy": 0.8935717940330505, "num_tokens": 832800666.0, "step": 21930 }, { "epoch": 2.7898486197684775, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.772221326828003, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8680385947227478, "num_tokens": 832835830.0, "step": 21931 }, { "epoch": 2.7899758300470676, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.370457649230957, "learning_rate": 1e-05, "loss": 0.3978, "mean_token_accuracy": 0.888867974281311, "num_tokens": 832866370.0, "step": 21932 }, { "epoch": 2.7901030403256586, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8494744300842285, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8778836131095886, "num_tokens": 832914174.0, "step": 21933 }, { "epoch": 2.7902302506042487, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3045601844787598, "learning_rate": 1e-05, "loss": 0.3919, "mean_token_accuracy": 0.8900203704833984, "num_tokens": 832951397.0, "step": 21934 }, { "epoch": 2.790357460882839, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.016385078430176, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8836662769317627, "num_tokens": 832986282.0, "step": 21935 }, { "epoch": 2.7904846711614297, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.316356897354126, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8866089582443237, "num_tokens": 833023288.0, "step": 21936 }, { "epoch": 2.7906118814400203, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.189084053039551, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8750110268592834, "num_tokens": 833053744.0, "step": 21937 }, { "epoch": 2.790739091718611, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.4381144046783447, "learning_rate": 1e-05, "loss": 0.4088, "mean_token_accuracy": 0.8879847526550293, "num_tokens": 833096499.0, "step": 21938 }, { "epoch": 2.7908663019972013, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.23439359664917, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8874163627624512, "num_tokens": 833132405.0, "step": 21939 }, { "epoch": 2.790993512275792, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.630890130996704, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8793064951896667, "num_tokens": 833175581.0, "step": 21940 }, { "epoch": 2.7911207225543824, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1005377769470215, "learning_rate": 1e-05, "loss": 0.477, "mean_token_accuracy": 0.8689057230949402, "num_tokens": 833213466.0, "step": 21941 }, { "epoch": 2.791247932832973, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.726673126220703, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8732879757881165, "num_tokens": 833250344.0, "step": 21942 }, { "epoch": 2.7913751431115634, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8759706020355225, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8783278465270996, "num_tokens": 833291032.0, "step": 21943 }, { "epoch": 2.791502353390154, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.7174856662750244, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8823673129081726, "num_tokens": 833331213.0, "step": 21944 }, { "epoch": 2.7916295636687445, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.667952299118042, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8863970041275024, "num_tokens": 833373906.0, "step": 21945 }, { "epoch": 2.791756773947335, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.713799476623535, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8717408180236816, "num_tokens": 833411724.0, "step": 21946 }, { "epoch": 2.7918839842259255, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.514173984527588, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8862288594245911, "num_tokens": 833446337.0, "step": 21947 }, { "epoch": 2.792011194504516, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.71565580368042, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8658339977264404, "num_tokens": 833484444.0, "step": 21948 }, { "epoch": 2.7921384047831066, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4038522243499756, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8811842203140259, "num_tokens": 833521375.0, "step": 21949 }, { "epoch": 2.792265615061697, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7599828243255615, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8872745633125305, "num_tokens": 833557784.0, "step": 21950 }, { "epoch": 2.7923928253402877, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3081412315368652, "learning_rate": 1e-05, "loss": 0.382, "mean_token_accuracy": 0.8974263668060303, "num_tokens": 833598872.0, "step": 21951 }, { "epoch": 2.792520035618878, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.865358591079712, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8870401382446289, "num_tokens": 833635799.0, "step": 21952 }, { "epoch": 2.7926472458974683, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2676703929901123, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.8867186903953552, "num_tokens": 833671814.0, "step": 21953 }, { "epoch": 2.7927744561760592, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0115742683410645, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8668222427368164, "num_tokens": 833718081.0, "step": 21954 }, { "epoch": 2.7929016664546493, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2706668376922607, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8760630488395691, "num_tokens": 833759197.0, "step": 21955 }, { "epoch": 2.7930288767332403, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.208412170410156, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8854986429214478, "num_tokens": 833785582.0, "step": 21956 }, { "epoch": 2.7931560870118304, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3963873386383057, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8864217400550842, "num_tokens": 833826611.0, "step": 21957 }, { "epoch": 2.793283297290421, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.2946295738220215, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.8752961158752441, "num_tokens": 833870657.0, "step": 21958 }, { "epoch": 2.7934105075690114, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.59818696975708, "learning_rate": 1e-05, "loss": 0.4412, "mean_token_accuracy": 0.8788719177246094, "num_tokens": 833907621.0, "step": 21959 }, { "epoch": 2.793537717847602, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2335381507873535, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8842808604240417, "num_tokens": 833944364.0, "step": 21960 }, { "epoch": 2.7936649281261925, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7143421173095703, "learning_rate": 1e-05, "loss": 0.4093, "mean_token_accuracy": 0.8897843956947327, "num_tokens": 833983840.0, "step": 21961 }, { "epoch": 2.793792138404783, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.068411827087402, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.8913227319717407, "num_tokens": 834016367.0, "step": 21962 }, { "epoch": 2.7939193486833735, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7296712398529053, "learning_rate": 1e-05, "loss": 0.4113, "mean_token_accuracy": 0.887258768081665, "num_tokens": 834050190.0, "step": 21963 }, { "epoch": 2.794046558961964, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.896925687789917, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8712743520736694, "num_tokens": 834087013.0, "step": 21964 }, { "epoch": 2.7941737692405546, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.669323444366455, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.876653790473938, "num_tokens": 834131771.0, "step": 21965 }, { "epoch": 2.794300979519145, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7533578872680664, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8843480348587036, "num_tokens": 834169055.0, "step": 21966 }, { "epoch": 2.7944281897977357, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.609180450439453, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8851783275604248, "num_tokens": 834206081.0, "step": 21967 }, { "epoch": 2.794555400076326, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6705005168914795, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8872817158699036, "num_tokens": 834240846.0, "step": 21968 }, { "epoch": 2.7946826103549167, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5385024547576904, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8778985142707825, "num_tokens": 834275037.0, "step": 21969 }, { "epoch": 2.7948098206335072, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.623886823654175, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8795828223228455, "num_tokens": 834316408.0, "step": 21970 }, { "epoch": 2.7949370309120978, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4698891639709473, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8876062035560608, "num_tokens": 834358473.0, "step": 21971 }, { "epoch": 2.7950642411906883, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6292686462402344, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8785188794136047, "num_tokens": 834399982.0, "step": 21972 }, { "epoch": 2.795191451469279, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.419713258743286, "learning_rate": 1e-05, "loss": 0.3977, "mean_token_accuracy": 0.8877594470977783, "num_tokens": 834429731.0, "step": 21973 }, { "epoch": 2.7953186617478694, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.663344144821167, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8773033618927002, "num_tokens": 834476958.0, "step": 21974 }, { "epoch": 2.79544587202646, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.358268976211548, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.884790301322937, "num_tokens": 834516820.0, "step": 21975 }, { "epoch": 2.79557308230505, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.741386890411377, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8789579272270203, "num_tokens": 834552455.0, "step": 21976 }, { "epoch": 2.795700292583641, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2832281589508057, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8762847781181335, "num_tokens": 834593059.0, "step": 21977 }, { "epoch": 2.795827502862231, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8806655406951904, "learning_rate": 1e-05, "loss": 0.4696, "mean_token_accuracy": 0.8698486685752869, "num_tokens": 834640134.0, "step": 21978 }, { "epoch": 2.795954713140822, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.24202299118042, "learning_rate": 1e-05, "loss": 0.437, "mean_token_accuracy": 0.8766106367111206, "num_tokens": 834679172.0, "step": 21979 }, { "epoch": 2.796081923419412, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.095613956451416, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8685618042945862, "num_tokens": 834713463.0, "step": 21980 }, { "epoch": 2.796209133698003, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.2914230823516846, "learning_rate": 1e-05, "loss": 0.4916, "mean_token_accuracy": 0.8570931553840637, "num_tokens": 834757782.0, "step": 21981 }, { "epoch": 2.796336343976593, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.2931671142578125, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8718783855438232, "num_tokens": 834791598.0, "step": 21982 }, { "epoch": 2.7964635542551837, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.478637933731079, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8698188662528992, "num_tokens": 834835205.0, "step": 21983 }, { "epoch": 2.796590764533774, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.31564998626709, "learning_rate": 1e-05, "loss": 0.4154, "mean_token_accuracy": 0.8887423276901245, "num_tokens": 834867450.0, "step": 21984 }, { "epoch": 2.7967179748123647, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6877827644348145, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8769155740737915, "num_tokens": 834911067.0, "step": 21985 }, { "epoch": 2.7968451850909553, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.160306453704834, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8792707920074463, "num_tokens": 834949117.0, "step": 21986 }, { "epoch": 2.796972395369546, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.76188063621521, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8762532472610474, "num_tokens": 834988909.0, "step": 21987 }, { "epoch": 2.7970996056481363, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9561264514923096, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.884260892868042, "num_tokens": 835031168.0, "step": 21988 }, { "epoch": 2.797226815926727, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.714362144470215, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8855804800987244, "num_tokens": 835072110.0, "step": 21989 }, { "epoch": 2.7973540262053174, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.793567657470703, "learning_rate": 1e-05, "loss": 0.4067, "mean_token_accuracy": 0.8915196657180786, "num_tokens": 835104605.0, "step": 21990 }, { "epoch": 2.797481236483908, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6724185943603516, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8641357421875, "num_tokens": 835141970.0, "step": 21991 }, { "epoch": 2.7976084467624984, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6705482006073, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8617318272590637, "num_tokens": 835177469.0, "step": 21992 }, { "epoch": 2.797735657041089, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5958569049835205, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8756020069122314, "num_tokens": 835220286.0, "step": 21993 }, { "epoch": 2.7978628673196795, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5963737964630127, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.881893515586853, "num_tokens": 835254290.0, "step": 21994 }, { "epoch": 2.79799007759827, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5492496490478516, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.876010537147522, "num_tokens": 835286375.0, "step": 21995 }, { "epoch": 2.7981172878768605, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.566570997238159, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8799349665641785, "num_tokens": 835315472.0, "step": 21996 }, { "epoch": 2.798244498155451, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4936907291412354, "learning_rate": 1e-05, "loss": 0.4055, "mean_token_accuracy": 0.8868459463119507, "num_tokens": 835355865.0, "step": 21997 }, { "epoch": 2.7983717084340416, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5582783222198486, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8785256147384644, "num_tokens": 835391251.0, "step": 21998 }, { "epoch": 2.798498918712632, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4481914043426514, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.8903257250785828, "num_tokens": 835424869.0, "step": 21999 }, { "epoch": 2.7986261289912227, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5829296112060547, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8806425333023071, "num_tokens": 835464916.0, "step": 22000 }, { "epoch": 2.7987533392698127, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3925998210906982, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8710067868232727, "num_tokens": 835503114.0, "step": 22001 }, { "epoch": 2.7988805495484037, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.643674612045288, "learning_rate": 1e-05, "loss": 0.4767, "mean_token_accuracy": 0.867462158203125, "num_tokens": 835541446.0, "step": 22002 }, { "epoch": 2.799007759826994, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.317648410797119, "learning_rate": 1e-05, "loss": 0.3881, "mean_token_accuracy": 0.8939462900161743, "num_tokens": 835579261.0, "step": 22003 }, { "epoch": 2.7991349701055848, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.752154588699341, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8758219480514526, "num_tokens": 835615847.0, "step": 22004 }, { "epoch": 2.799262180384175, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.238381862640381, "learning_rate": 1e-05, "loss": 0.4913, "mean_token_accuracy": 0.858556866645813, "num_tokens": 835658461.0, "step": 22005 }, { "epoch": 2.799389390662766, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9243853092193604, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8856123685836792, "num_tokens": 835693489.0, "step": 22006 }, { "epoch": 2.799516600941356, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.170630931854248, "learning_rate": 1e-05, "loss": 0.3841, "mean_token_accuracy": 0.8934638500213623, "num_tokens": 835733925.0, "step": 22007 }, { "epoch": 2.7996438112199464, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.195164680480957, "learning_rate": 1e-05, "loss": 0.4739, "mean_token_accuracy": 0.8727449774742126, "num_tokens": 835775071.0, "step": 22008 }, { "epoch": 2.799771021498537, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.2691473960876465, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8764623999595642, "num_tokens": 835813940.0, "step": 22009 }, { "epoch": 2.7998982317771275, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.451422691345215, "learning_rate": 1e-05, "loss": 0.4978, "mean_token_accuracy": 0.8643175959587097, "num_tokens": 835849667.0, "step": 22010 }, { "epoch": 2.800025442055718, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.5804014205932617, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8841224908828735, "num_tokens": 835880928.0, "step": 22011 }, { "epoch": 2.8001526523343085, "ewc_loss": 0.10498046875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.44094181060791, "learning_rate": 1e-05, "loss": 0.3948, "mean_token_accuracy": 0.8974018096923828, "num_tokens": 835917502.0, "step": 22012 }, { "epoch": 2.800279862612899, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.8282227516174316, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.869369387626648, "num_tokens": 835958547.0, "step": 22013 }, { "epoch": 2.8004070728914896, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.243053913116455, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8758623600006104, "num_tokens": 835997427.0, "step": 22014 }, { "epoch": 2.80053428317008, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.848564863204956, "learning_rate": 1e-05, "loss": 0.4213, "mean_token_accuracy": 0.8842520713806152, "num_tokens": 836039096.0, "step": 22015 }, { "epoch": 2.8006614934486707, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.039821147918701, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.8774940371513367, "num_tokens": 836079613.0, "step": 22016 }, { "epoch": 2.800788703727261, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.766775369644165, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8837471008300781, "num_tokens": 836117500.0, "step": 22017 }, { "epoch": 2.8009159140058517, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.895869731903076, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8791711926460266, "num_tokens": 836149680.0, "step": 22018 }, { "epoch": 2.8010431242844422, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6691746711730957, "learning_rate": 1e-05, "loss": 0.3781, "mean_token_accuracy": 0.8966012001037598, "num_tokens": 836179869.0, "step": 22019 }, { "epoch": 2.8011703345630328, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7876462936401367, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8748773336410522, "num_tokens": 836222305.0, "step": 22020 }, { "epoch": 2.8012975448416233, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5699546337127686, "learning_rate": 1e-05, "loss": 0.4509, "mean_token_accuracy": 0.8749376535415649, "num_tokens": 836257687.0, "step": 22021 }, { "epoch": 2.801424755120214, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7539055347442627, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8775895833969116, "num_tokens": 836299105.0, "step": 22022 }, { "epoch": 2.8015519653988044, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.470688819885254, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8811782598495483, "num_tokens": 836337661.0, "step": 22023 }, { "epoch": 2.801679175677395, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7689132690429688, "learning_rate": 1e-05, "loss": 0.4707, "mean_token_accuracy": 0.8703885078430176, "num_tokens": 836378792.0, "step": 22024 }, { "epoch": 2.8018063859559854, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.404078960418701, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.8749956488609314, "num_tokens": 836412754.0, "step": 22025 }, { "epoch": 2.8019335962345755, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8428285121917725, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8836758136749268, "num_tokens": 836452173.0, "step": 22026 }, { "epoch": 2.8020608065131665, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.316709041595459, "learning_rate": 1e-05, "loss": 0.404, "mean_token_accuracy": 0.8878889679908752, "num_tokens": 836491398.0, "step": 22027 }, { "epoch": 2.8021880167917566, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 4.011746406555176, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.869717538356781, "num_tokens": 836525130.0, "step": 22028 }, { "epoch": 2.8023152270703475, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.3340256214141846, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8757233023643494, "num_tokens": 836561288.0, "step": 22029 }, { "epoch": 2.8024424373489376, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.152714252471924, "learning_rate": 1e-05, "loss": 0.4986, "mean_token_accuracy": 0.8638644814491272, "num_tokens": 836603595.0, "step": 22030 }, { "epoch": 2.8025696476275286, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.4434001445770264, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8715258836746216, "num_tokens": 836640460.0, "step": 22031 }, { "epoch": 2.8026968579061187, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.211243152618408, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.869412362575531, "num_tokens": 836678024.0, "step": 22032 }, { "epoch": 2.802824068184709, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5896432399749756, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.8936147689819336, "num_tokens": 836718383.0, "step": 22033 }, { "epoch": 2.8029512784632997, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.11812686920166, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8666170835494995, "num_tokens": 836755564.0, "step": 22034 }, { "epoch": 2.8030784887418903, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.706414222717285, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8766225576400757, "num_tokens": 836792891.0, "step": 22035 }, { "epoch": 2.803205699020481, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9112415313720703, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.874024510383606, "num_tokens": 836830024.0, "step": 22036 }, { "epoch": 2.8033329092990713, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.7043251991271973, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.870173454284668, "num_tokens": 836871600.0, "step": 22037 }, { "epoch": 2.803460119577662, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.703181266784668, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8862237334251404, "num_tokens": 836911495.0, "step": 22038 }, { "epoch": 2.8035873298562524, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.680051803588867, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8803001642227173, "num_tokens": 836952840.0, "step": 22039 }, { "epoch": 2.803714540134843, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.553375244140625, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8749604225158691, "num_tokens": 836988355.0, "step": 22040 }, { "epoch": 2.8038417504134334, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6858043670654297, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8790171146392822, "num_tokens": 837021559.0, "step": 22041 }, { "epoch": 2.803968960692024, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.432753801345825, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8745502829551697, "num_tokens": 837060401.0, "step": 22042 }, { "epoch": 2.8040961709706145, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.70985746383667, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8890563249588013, "num_tokens": 837094769.0, "step": 22043 }, { "epoch": 2.804223381249205, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3638086318969727, "learning_rate": 1e-05, "loss": 0.3852, "mean_token_accuracy": 0.8932972550392151, "num_tokens": 837125820.0, "step": 22044 }, { "epoch": 2.8043505915277955, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.760211229324341, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8865004181861877, "num_tokens": 837167392.0, "step": 22045 }, { "epoch": 2.804477801806386, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.289259433746338, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8770836591720581, "num_tokens": 837207439.0, "step": 22046 }, { "epoch": 2.8046050120849766, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8871779441833496, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.883516788482666, "num_tokens": 837240038.0, "step": 22047 }, { "epoch": 2.804732222363567, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2602646350860596, "learning_rate": 1e-05, "loss": 0.4572, "mean_token_accuracy": 0.8716111183166504, "num_tokens": 837278842.0, "step": 22048 }, { "epoch": 2.8048594326421576, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.030595302581787, "learning_rate": 1e-05, "loss": 0.3939, "mean_token_accuracy": 0.8947023153305054, "num_tokens": 837317519.0, "step": 22049 }, { "epoch": 2.804986642920748, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.277313232421875, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.887825608253479, "num_tokens": 837349053.0, "step": 22050 }, { "epoch": 2.8051138531993383, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.198398113250732, "learning_rate": 1e-05, "loss": 0.4512, "mean_token_accuracy": 0.8759593963623047, "num_tokens": 837386018.0, "step": 22051 }, { "epoch": 2.8052410634779292, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4070379734039307, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8729478120803833, "num_tokens": 837423640.0, "step": 22052 }, { "epoch": 2.8053682737565193, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.281297206878662, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8787747621536255, "num_tokens": 837462198.0, "step": 22053 }, { "epoch": 2.8054954840351103, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6003708839416504, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8765853047370911, "num_tokens": 837492722.0, "step": 22054 }, { "epoch": 2.8056226943137004, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.197129726409912, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.884141743183136, "num_tokens": 837531598.0, "step": 22055 }, { "epoch": 2.805749904592291, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7026548385620117, "learning_rate": 1e-05, "loss": 0.4632, "mean_token_accuracy": 0.8722567558288574, "num_tokens": 837572580.0, "step": 22056 }, { "epoch": 2.8058771148708814, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.028009414672852, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8862025737762451, "num_tokens": 837610398.0, "step": 22057 }, { "epoch": 2.806004325149472, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.6984355449676514, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8755730986595154, "num_tokens": 837653518.0, "step": 22058 }, { "epoch": 2.8061315354280625, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.855018377304077, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8764699101448059, "num_tokens": 837697397.0, "step": 22059 }, { "epoch": 2.806258745706653, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.647519826889038, "learning_rate": 1e-05, "loss": 0.39, "mean_token_accuracy": 0.8943493962287903, "num_tokens": 837736565.0, "step": 22060 }, { "epoch": 2.8063859559852435, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7379696369171143, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8795101642608643, "num_tokens": 837772514.0, "step": 22061 }, { "epoch": 2.806513166263834, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.5893704891204834, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8860486745834351, "num_tokens": 837805147.0, "step": 22062 }, { "epoch": 2.8066403765424246, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6618764400482178, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8655970692634583, "num_tokens": 837846795.0, "step": 22063 }, { "epoch": 2.806767586821015, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.531642436981201, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8828154802322388, "num_tokens": 837882703.0, "step": 22064 }, { "epoch": 2.8068947970996057, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6287660598754883, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.8642664551734924, "num_tokens": 837924698.0, "step": 22065 }, { "epoch": 2.807022007378196, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4779088497161865, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8815279603004456, "num_tokens": 837958553.0, "step": 22066 }, { "epoch": 2.8071492176567867, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6211185455322266, "learning_rate": 1e-05, "loss": 0.3644, "mean_token_accuracy": 0.9020438194274902, "num_tokens": 837998674.0, "step": 22067 }, { "epoch": 2.8072764279353772, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4117908477783203, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8797277808189392, "num_tokens": 838035111.0, "step": 22068 }, { "epoch": 2.8074036382139678, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.664550542831421, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8747355937957764, "num_tokens": 838075444.0, "step": 22069 }, { "epoch": 2.8075308484925583, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3545289039611816, "learning_rate": 1e-05, "loss": 0.5038, "mean_token_accuracy": 0.859380841255188, "num_tokens": 838117972.0, "step": 22070 }, { "epoch": 2.807658058771149, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7500345706939697, "learning_rate": 1e-05, "loss": 0.4667, "mean_token_accuracy": 0.8683610558509827, "num_tokens": 838158797.0, "step": 22071 }, { "epoch": 2.8077852690497394, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2784574031829834, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8736777305603027, "num_tokens": 838198958.0, "step": 22072 }, { "epoch": 2.80791247932833, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8990349769592285, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8757156729698181, "num_tokens": 838233936.0, "step": 22073 }, { "epoch": 2.80803968960692, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2356693744659424, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8743327856063843, "num_tokens": 838274333.0, "step": 22074 }, { "epoch": 2.808166899885511, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.115185260772705, "learning_rate": 1e-05, "loss": 0.4335, "mean_token_accuracy": 0.8817427158355713, "num_tokens": 838304932.0, "step": 22075 }, { "epoch": 2.808294110164101, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.299800395965576, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8681286573410034, "num_tokens": 838345377.0, "step": 22076 }, { "epoch": 2.808421320442692, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.281878471374512, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8773666024208069, "num_tokens": 838383105.0, "step": 22077 }, { "epoch": 2.808548530721282, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.496199131011963, "learning_rate": 1e-05, "loss": 0.4801, "mean_token_accuracy": 0.8651440143585205, "num_tokens": 838421857.0, "step": 22078 }, { "epoch": 2.808675740999873, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.2896318435668945, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.867599606513977, "num_tokens": 838459439.0, "step": 22079 }, { "epoch": 2.808802951278463, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6842129230499268, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8747997283935547, "num_tokens": 838501176.0, "step": 22080 }, { "epoch": 2.8089301615570537, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.133342266082764, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8654172420501709, "num_tokens": 838538725.0, "step": 22081 }, { "epoch": 2.809057371835644, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.739957809448242, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.871810793876648, "num_tokens": 838577365.0, "step": 22082 }, { "epoch": 2.8091845821142347, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9238603115081787, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8839163780212402, "num_tokens": 838617872.0, "step": 22083 }, { "epoch": 2.8093117923928252, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.705874443054199, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8725947737693787, "num_tokens": 838656599.0, "step": 22084 }, { "epoch": 2.8094390026714158, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7566957473754883, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8766346573829651, "num_tokens": 838693270.0, "step": 22085 }, { "epoch": 2.8095662129500063, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6584994792938232, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8723519444465637, "num_tokens": 838728493.0, "step": 22086 }, { "epoch": 2.809693423228597, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6251063346862793, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8787381052970886, "num_tokens": 838767136.0, "step": 22087 }, { "epoch": 2.8098206335071874, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.613232135772705, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8766484260559082, "num_tokens": 838801894.0, "step": 22088 }, { "epoch": 2.809947843785778, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5391340255737305, "learning_rate": 1e-05, "loss": 0.3735, "mean_token_accuracy": 0.8987594246864319, "num_tokens": 838838732.0, "step": 22089 }, { "epoch": 2.8100750540643684, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.563741445541382, "learning_rate": 1e-05, "loss": 0.3646, "mean_token_accuracy": 0.9019320011138916, "num_tokens": 838876331.0, "step": 22090 }, { "epoch": 2.810202264342959, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.509828567504883, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.8810740113258362, "num_tokens": 838909613.0, "step": 22091 }, { "epoch": 2.8103294746215495, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5471932888031006, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.869144082069397, "num_tokens": 838944694.0, "step": 22092 }, { "epoch": 2.81045668490014, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4704997539520264, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8839008808135986, "num_tokens": 838984595.0, "step": 22093 }, { "epoch": 2.8105838951787305, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5482125282287598, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.875050961971283, "num_tokens": 839018528.0, "step": 22094 }, { "epoch": 2.810711105457321, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.442591667175293, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8770487308502197, "num_tokens": 839056447.0, "step": 22095 }, { "epoch": 2.8108383157359116, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5464956760406494, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8852936029434204, "num_tokens": 839093612.0, "step": 22096 }, { "epoch": 2.810965526014502, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.415609836578369, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8774346709251404, "num_tokens": 839130192.0, "step": 22097 }, { "epoch": 2.8110927362930926, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.569838523864746, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8742673397064209, "num_tokens": 839167821.0, "step": 22098 }, { "epoch": 2.8112199465716827, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.383512258529663, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.868999183177948, "num_tokens": 839200124.0, "step": 22099 }, { "epoch": 2.8113471568502737, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6074094772338867, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8751580119132996, "num_tokens": 839242787.0, "step": 22100 }, { "epoch": 2.811474367128864, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.322911500930786, "learning_rate": 1e-05, "loss": 0.388, "mean_token_accuracy": 0.8937485218048096, "num_tokens": 839277517.0, "step": 22101 }, { "epoch": 2.8116015774074548, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003509521484375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6723921298980713, "learning_rate": 1e-05, "loss": 0.3994, "mean_token_accuracy": 0.8911055326461792, "num_tokens": 839318609.0, "step": 22102 }, { "epoch": 2.811728787686045, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.264967918395996, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8729252219200134, "num_tokens": 839353964.0, "step": 22103 }, { "epoch": 2.811855997964636, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.803015947341919, "learning_rate": 1e-05, "loss": 0.3933, "mean_token_accuracy": 0.8948800563812256, "num_tokens": 839393926.0, "step": 22104 }, { "epoch": 2.811983208243226, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.178798198699951, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8775771856307983, "num_tokens": 839437857.0, "step": 22105 }, { "epoch": 2.8121104185218164, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.056048393249512, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8683444261550903, "num_tokens": 839478444.0, "step": 22106 }, { "epoch": 2.812237628800407, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.158496379852295, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8826769590377808, "num_tokens": 839512752.0, "step": 22107 }, { "epoch": 2.8123648390789975, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.393150329589844, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8651882410049438, "num_tokens": 839550024.0, "step": 22108 }, { "epoch": 2.812492049357588, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.375743865966797, "learning_rate": 1e-05, "loss": 0.4127, "mean_token_accuracy": 0.8859379291534424, "num_tokens": 839582612.0, "step": 22109 }, { "epoch": 2.8126192596361785, "ewc_loss": 0.1083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.603593826293945, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8807374238967896, "num_tokens": 839621791.0, "step": 22110 }, { "epoch": 2.812746469914769, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.797926187515259, "learning_rate": 1e-05, "loss": 0.4936, "mean_token_accuracy": 0.8645505905151367, "num_tokens": 839660829.0, "step": 22111 }, { "epoch": 2.8128736801933596, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.461945533752441, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8795440793037415, "num_tokens": 839701675.0, "step": 22112 }, { "epoch": 2.81300089047195, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 3.9711427688598633, "learning_rate": 1e-05, "loss": 0.4821, "mean_token_accuracy": 0.8701043128967285, "num_tokens": 839742958.0, "step": 22113 }, { "epoch": 2.8131281007505406, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.199782848358154, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8794158697128296, "num_tokens": 839787552.0, "step": 22114 }, { "epoch": 2.813255311029131, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.8937904834747314, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8761860132217407, "num_tokens": 839824630.0, "step": 22115 }, { "epoch": 2.8133825213077217, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.024507522583008, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8881146907806396, "num_tokens": 839857671.0, "step": 22116 }, { "epoch": 2.8135097315863122, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.7632908821105957, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8754596710205078, "num_tokens": 839896395.0, "step": 22117 }, { "epoch": 2.8136369418649028, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8957955837249756, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8776805400848389, "num_tokens": 839942188.0, "step": 22118 }, { "epoch": 2.8137641521434933, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6437346935272217, "learning_rate": 1e-05, "loss": 0.378, "mean_token_accuracy": 0.8942945599555969, "num_tokens": 839970510.0, "step": 22119 }, { "epoch": 2.813891362422084, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.834076166152954, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8798022866249084, "num_tokens": 840007458.0, "step": 22120 }, { "epoch": 2.8140185727006743, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5115933418273926, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8763563632965088, "num_tokens": 840049729.0, "step": 22121 }, { "epoch": 2.814145782979265, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8443150520324707, "learning_rate": 1e-05, "loss": 0.4397, "mean_token_accuracy": 0.8801626563072205, "num_tokens": 840089883.0, "step": 22122 }, { "epoch": 2.8142729932578554, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.431023597717285, "learning_rate": 1e-05, "loss": 0.3847, "mean_token_accuracy": 0.8960001468658447, "num_tokens": 840121237.0, "step": 22123 }, { "epoch": 2.8144002035364455, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.9272618293762207, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8687636852264404, "num_tokens": 840157213.0, "step": 22124 }, { "epoch": 2.8145274138150365, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3624372482299805, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8837459087371826, "num_tokens": 840191404.0, "step": 22125 }, { "epoch": 2.8146546240936265, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.086111068725586, "learning_rate": 1e-05, "loss": 0.478, "mean_token_accuracy": 0.8711165189743042, "num_tokens": 840227803.0, "step": 22126 }, { "epoch": 2.8147818343722175, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.394115924835205, "learning_rate": 1e-05, "loss": 0.502, "mean_token_accuracy": 0.8581790924072266, "num_tokens": 840266686.0, "step": 22127 }, { "epoch": 2.8149090446508076, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.190008640289307, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8855012655258179, "num_tokens": 840310437.0, "step": 22128 }, { "epoch": 2.815036254929398, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.5143768787384033, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8778432607650757, "num_tokens": 840349599.0, "step": 22129 }, { "epoch": 2.8151634652079887, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.213365077972412, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8829582333564758, "num_tokens": 840389143.0, "step": 22130 }, { "epoch": 2.815290675486579, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6782376766204834, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8848090767860413, "num_tokens": 840429440.0, "step": 22131 }, { "epoch": 2.8154178857651697, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.051087379455566, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8789912462234497, "num_tokens": 840466056.0, "step": 22132 }, { "epoch": 2.8155450960437602, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7360877990722656, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8769431114196777, "num_tokens": 840503039.0, "step": 22133 }, { "epoch": 2.8156723063223508, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8181569576263428, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8826496601104736, "num_tokens": 840546809.0, "step": 22134 }, { "epoch": 2.8157995166009413, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.715553045272827, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.8749275803565979, "num_tokens": 840582617.0, "step": 22135 }, { "epoch": 2.815926726879532, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.631582498550415, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8805218935012817, "num_tokens": 840620279.0, "step": 22136 }, { "epoch": 2.8160539371581224, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7069034576416016, "learning_rate": 1e-05, "loss": 0.4305, "mean_token_accuracy": 0.881839394569397, "num_tokens": 840655925.0, "step": 22137 }, { "epoch": 2.816181147436713, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4847569465637207, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.880186140537262, "num_tokens": 840692328.0, "step": 22138 }, { "epoch": 2.8163083577153034, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.709984540939331, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.879538357257843, "num_tokens": 840730674.0, "step": 22139 }, { "epoch": 2.816435567993894, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3902437686920166, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.872299313545227, "num_tokens": 840764791.0, "step": 22140 }, { "epoch": 2.8165627782724845, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7502388954162598, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8793538808822632, "num_tokens": 840803527.0, "step": 22141 }, { "epoch": 2.816689988551075, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.311532974243164, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.87132728099823, "num_tokens": 840843823.0, "step": 22142 }, { "epoch": 2.8168171988296655, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8567276000976562, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8691117763519287, "num_tokens": 840879648.0, "step": 22143 }, { "epoch": 2.816944409108256, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2546370029449463, "learning_rate": 1e-05, "loss": 0.3821, "mean_token_accuracy": 0.8927356600761414, "num_tokens": 840914979.0, "step": 22144 }, { "epoch": 2.8170716193868466, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.02094030380249, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.881508469581604, "num_tokens": 840951534.0, "step": 22145 }, { "epoch": 2.817198829665437, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.264897346496582, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8715412020683289, "num_tokens": 840987416.0, "step": 22146 }, { "epoch": 2.8173260399440276, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.217260360717773, "learning_rate": 1e-05, "loss": 0.4985, "mean_token_accuracy": 0.8618617653846741, "num_tokens": 841026734.0, "step": 22147 }, { "epoch": 2.817453250222618, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.3862967491149902, "learning_rate": 1e-05, "loss": 0.3915, "mean_token_accuracy": 0.8916773796081543, "num_tokens": 841065333.0, "step": 22148 }, { "epoch": 2.8175804605012083, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.334277629852295, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.8861631751060486, "num_tokens": 841102257.0, "step": 22149 }, { "epoch": 2.8177076707797992, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.619962692260742, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8805959224700928, "num_tokens": 841140253.0, "step": 22150 }, { "epoch": 2.8178348810583893, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.244851589202881, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8822283744812012, "num_tokens": 841181675.0, "step": 22151 }, { "epoch": 2.8179620913369803, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.749173641204834, "learning_rate": 1e-05, "loss": 0.4297, "mean_token_accuracy": 0.8846713304519653, "num_tokens": 841219922.0, "step": 22152 }, { "epoch": 2.8180893016155704, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.058810710906982, "learning_rate": 1e-05, "loss": 0.4265, "mean_token_accuracy": 0.8861038088798523, "num_tokens": 841256296.0, "step": 22153 }, { "epoch": 2.818216511894161, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7436015605926514, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8853515386581421, "num_tokens": 841296737.0, "step": 22154 }, { "epoch": 2.8183437221727514, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.886798620223999, "learning_rate": 1e-05, "loss": 0.4917, "mean_token_accuracy": 0.8651492595672607, "num_tokens": 841329863.0, "step": 22155 }, { "epoch": 2.818470932451342, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6956095695495605, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8745307922363281, "num_tokens": 841365057.0, "step": 22156 }, { "epoch": 2.8185981427299325, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.722771644592285, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8826900720596313, "num_tokens": 841404885.0, "step": 22157 }, { "epoch": 2.818725353008523, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.619713068008423, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.874443769454956, "num_tokens": 841445067.0, "step": 22158 }, { "epoch": 2.8188525632871135, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6288251876831055, "learning_rate": 1e-05, "loss": 0.3781, "mean_token_accuracy": 0.8951417803764343, "num_tokens": 841480569.0, "step": 22159 }, { "epoch": 2.818979773565704, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.557579278945923, "learning_rate": 1e-05, "loss": 0.4077, "mean_token_accuracy": 0.885956883430481, "num_tokens": 841520540.0, "step": 22160 }, { "epoch": 2.8191069838442946, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5859787464141846, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8868309259414673, "num_tokens": 841557765.0, "step": 22161 }, { "epoch": 2.819234194122885, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5105700492858887, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.876093864440918, "num_tokens": 841593345.0, "step": 22162 }, { "epoch": 2.8193614044014756, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.577148914337158, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8716762065887451, "num_tokens": 841629977.0, "step": 22163 }, { "epoch": 2.819488614680066, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.446068525314331, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8795503377914429, "num_tokens": 841671901.0, "step": 22164 }, { "epoch": 2.8196158249586567, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.605058431625366, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.87944495677948, "num_tokens": 841706072.0, "step": 22165 }, { "epoch": 2.8197430352372472, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3871986865997314, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8690863847732544, "num_tokens": 841745707.0, "step": 22166 }, { "epoch": 2.8198702455158378, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.653303623199463, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8800528049468994, "num_tokens": 841781468.0, "step": 22167 }, { "epoch": 2.8199974557944283, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.316976547241211, "learning_rate": 1e-05, "loss": 0.4026, "mean_token_accuracy": 0.8877612352371216, "num_tokens": 841821890.0, "step": 22168 }, { "epoch": 2.820124666073019, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7665677070617676, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8696014881134033, "num_tokens": 841857668.0, "step": 22169 }, { "epoch": 2.8202518763516093, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.255420207977295, "learning_rate": 1e-05, "loss": 0.4249, "mean_token_accuracy": 0.8781141638755798, "num_tokens": 841892595.0, "step": 22170 }, { "epoch": 2.8203790866302, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9324722290039062, "learning_rate": 1e-05, "loss": 0.4011, "mean_token_accuracy": 0.8925154209136963, "num_tokens": 841928077.0, "step": 22171 }, { "epoch": 2.82050629690879, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.198392391204834, "learning_rate": 1e-05, "loss": 0.4131, "mean_token_accuracy": 0.8805940747261047, "num_tokens": 841967847.0, "step": 22172 }, { "epoch": 2.820633507187381, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.177916526794434, "learning_rate": 1e-05, "loss": 0.5047, "mean_token_accuracy": 0.866482138633728, "num_tokens": 842008931.0, "step": 22173 }, { "epoch": 2.820760717465971, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.306398391723633, "learning_rate": 1e-05, "loss": 0.4329, "mean_token_accuracy": 0.8769635558128357, "num_tokens": 842042870.0, "step": 22174 }, { "epoch": 2.820887927744562, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.382116794586182, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8807519674301147, "num_tokens": 842080193.0, "step": 22175 }, { "epoch": 2.821015138023152, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.5511302947998047, "learning_rate": 1e-05, "loss": 0.3829, "mean_token_accuracy": 0.8961403369903564, "num_tokens": 842120953.0, "step": 22176 }, { "epoch": 2.821142348301743, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.37349796295166, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8888758420944214, "num_tokens": 842158466.0, "step": 22177 }, { "epoch": 2.821269558580333, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.78525972366333, "learning_rate": 1e-05, "loss": 0.4488, "mean_token_accuracy": 0.8764876127243042, "num_tokens": 842194164.0, "step": 22178 }, { "epoch": 2.8213967688589237, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1778244972229, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.876846194267273, "num_tokens": 842232137.0, "step": 22179 }, { "epoch": 2.821523979137514, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.8262717723846436, "learning_rate": 1e-05, "loss": 0.4737, "mean_token_accuracy": 0.8689687252044678, "num_tokens": 842263770.0, "step": 22180 }, { "epoch": 2.8216511894161047, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.956564426422119, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8710101842880249, "num_tokens": 842309420.0, "step": 22181 }, { "epoch": 2.8217783996946952, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.7377545833587646, "learning_rate": 1e-05, "loss": 0.4838, "mean_token_accuracy": 0.8644793033599854, "num_tokens": 842350533.0, "step": 22182 }, { "epoch": 2.8219056099732858, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8241302967071533, "learning_rate": 1e-05, "loss": 0.4708, "mean_token_accuracy": 0.8667952418327332, "num_tokens": 842389734.0, "step": 22183 }, { "epoch": 2.8220328202518763, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6537535190582275, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8694776296615601, "num_tokens": 842424425.0, "step": 22184 }, { "epoch": 2.822160030530467, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.715749502182007, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.878455400466919, "num_tokens": 842460269.0, "step": 22185 }, { "epoch": 2.8222872408090574, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5649235248565674, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8754594326019287, "num_tokens": 842497018.0, "step": 22186 }, { "epoch": 2.822414451087648, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6627895832061768, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8810083270072937, "num_tokens": 842537563.0, "step": 22187 }, { "epoch": 2.8225416613662384, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.481844663619995, "learning_rate": 1e-05, "loss": 0.3774, "mean_token_accuracy": 0.8977624177932739, "num_tokens": 842573842.0, "step": 22188 }, { "epoch": 2.822668871644829, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.67216157913208, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.867746114730835, "num_tokens": 842614216.0, "step": 22189 }, { "epoch": 2.8227960819234195, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.41717529296875, "learning_rate": 1e-05, "loss": 0.5024, "mean_token_accuracy": 0.8604201078414917, "num_tokens": 842651823.0, "step": 22190 }, { "epoch": 2.82292329220201, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7217350006103516, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.8608771562576294, "num_tokens": 842695113.0, "step": 22191 }, { "epoch": 2.8230505024806005, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3349640369415283, "learning_rate": 1e-05, "loss": 0.3785, "mean_token_accuracy": 0.8945446014404297, "num_tokens": 842731341.0, "step": 22192 }, { "epoch": 2.823177712759191, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.82450795173645, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8695588111877441, "num_tokens": 842769611.0, "step": 22193 }, { "epoch": 2.8233049230377816, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.281270980834961, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8692331910133362, "num_tokens": 842807106.0, "step": 22194 }, { "epoch": 2.823432133316372, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.977280855178833, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8817404508590698, "num_tokens": 842847957.0, "step": 22195 }, { "epoch": 2.8235593435949626, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2670023441314697, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8665481805801392, "num_tokens": 842889222.0, "step": 22196 }, { "epoch": 2.8236865538735527, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.161048889160156, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.885665774345398, "num_tokens": 842925164.0, "step": 22197 }, { "epoch": 2.8238137641521437, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.3812942504882812, "learning_rate": 1e-05, "loss": 0.4747, "mean_token_accuracy": 0.8644042015075684, "num_tokens": 842963832.0, "step": 22198 }, { "epoch": 2.8239409744307338, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.271112442016602, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.8720548748970032, "num_tokens": 843004357.0, "step": 22199 }, { "epoch": 2.8240681847093247, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.578765869140625, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8572776317596436, "num_tokens": 843045241.0, "step": 22200 }, { "epoch": 2.824195394987915, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.200509548187256, "learning_rate": 1e-05, "loss": 0.4673, "mean_token_accuracy": 0.8720483779907227, "num_tokens": 843084473.0, "step": 22201 }, { "epoch": 2.824322605266506, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7111587524414062, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8772191405296326, "num_tokens": 843122118.0, "step": 22202 }, { "epoch": 2.824449815545096, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.017149925231934, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8655257225036621, "num_tokens": 843159936.0, "step": 22203 }, { "epoch": 2.8245770258236864, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7631423473358154, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8743075132369995, "num_tokens": 843196212.0, "step": 22204 }, { "epoch": 2.824704236102277, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8022923469543457, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.8821375966072083, "num_tokens": 843236870.0, "step": 22205 }, { "epoch": 2.8248314463808675, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6922824382781982, "learning_rate": 1e-05, "loss": 0.4239, "mean_token_accuracy": 0.8831056356430054, "num_tokens": 843270701.0, "step": 22206 }, { "epoch": 2.824958656659458, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.628345012664795, "learning_rate": 1e-05, "loss": 0.4139, "mean_token_accuracy": 0.8835588693618774, "num_tokens": 843308547.0, "step": 22207 }, { "epoch": 2.8250858669380485, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.63767147064209, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.8892890214920044, "num_tokens": 843354496.0, "step": 22208 }, { "epoch": 2.825213077216639, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.545397996902466, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.869236946105957, "num_tokens": 843389995.0, "step": 22209 }, { "epoch": 2.8253402874952296, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6155173778533936, "learning_rate": 1e-05, "loss": 0.4615, "mean_token_accuracy": 0.8751797676086426, "num_tokens": 843428444.0, "step": 22210 }, { "epoch": 2.82546749777382, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.473903179168701, "learning_rate": 1e-05, "loss": 0.425, "mean_token_accuracy": 0.8815946578979492, "num_tokens": 843466390.0, "step": 22211 }, { "epoch": 2.8255947080524106, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6143765449523926, "learning_rate": 1e-05, "loss": 0.5097, "mean_token_accuracy": 0.8612802624702454, "num_tokens": 843503424.0, "step": 22212 }, { "epoch": 2.825721918331001, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.426164388656616, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8680212497711182, "num_tokens": 843544669.0, "step": 22213 }, { "epoch": 2.8258491286095917, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.625215530395508, "learning_rate": 1e-05, "loss": 0.3893, "mean_token_accuracy": 0.8930521607398987, "num_tokens": 843579777.0, "step": 22214 }, { "epoch": 2.8259763388881822, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.36989688873291, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8786035776138306, "num_tokens": 843613481.0, "step": 22215 }, { "epoch": 2.8261035491667728, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.665396213531494, "learning_rate": 1e-05, "loss": 0.3901, "mean_token_accuracy": 0.892654299736023, "num_tokens": 843656870.0, "step": 22216 }, { "epoch": 2.8262307594453633, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3170886039733887, "learning_rate": 1e-05, "loss": 0.4142, "mean_token_accuracy": 0.8801150918006897, "num_tokens": 843693859.0, "step": 22217 }, { "epoch": 2.826357969723954, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7659149169921875, "learning_rate": 1e-05, "loss": 0.3946, "mean_token_accuracy": 0.8936567902565002, "num_tokens": 843726999.0, "step": 22218 }, { "epoch": 2.8264851800025443, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2410616874694824, "learning_rate": 1e-05, "loss": 0.4507, "mean_token_accuracy": 0.8724026083946228, "num_tokens": 843767244.0, "step": 22219 }, { "epoch": 2.826612390281135, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9454047679901123, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8762361407279968, "num_tokens": 843799828.0, "step": 22220 }, { "epoch": 2.8267396005597254, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.191302537918091, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8778978586196899, "num_tokens": 843837485.0, "step": 22221 }, { "epoch": 2.8268668108383155, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.189698219299316, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8850553035736084, "num_tokens": 843878796.0, "step": 22222 }, { "epoch": 2.8269940211169065, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.273646354675293, "learning_rate": 1e-05, "loss": 0.4562, "mean_token_accuracy": 0.8724232912063599, "num_tokens": 843920200.0, "step": 22223 }, { "epoch": 2.8271212313954965, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.427008152008057, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8865363597869873, "num_tokens": 843961719.0, "step": 22224 }, { "epoch": 2.8272484416740875, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.563063383102417, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8695551753044128, "num_tokens": 844005085.0, "step": 22225 }, { "epoch": 2.8273756519526776, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.445744037628174, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.870090126991272, "num_tokens": 844036964.0, "step": 22226 }, { "epoch": 2.827502862231268, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.833523750305176, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.872031569480896, "num_tokens": 844078297.0, "step": 22227 }, { "epoch": 2.8276300725098586, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.226502895355225, "learning_rate": 1e-05, "loss": 0.4084, "mean_token_accuracy": 0.8905249834060669, "num_tokens": 844113749.0, "step": 22228 }, { "epoch": 2.827757282788449, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.858229398727417, "learning_rate": 1e-05, "loss": 0.4712, "mean_token_accuracy": 0.871001124382019, "num_tokens": 844155993.0, "step": 22229 }, { "epoch": 2.8278844930670397, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.005673885345459, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8778078556060791, "num_tokens": 844204302.0, "step": 22230 }, { "epoch": 2.8280117033456302, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7783734798431396, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8829062581062317, "num_tokens": 844241490.0, "step": 22231 }, { "epoch": 2.8281389136242208, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.851440906524658, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8819590210914612, "num_tokens": 844282891.0, "step": 22232 }, { "epoch": 2.8282661239028113, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6694371700286865, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8749234080314636, "num_tokens": 844324407.0, "step": 22233 }, { "epoch": 2.828393334181402, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.76761531829834, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8810807466506958, "num_tokens": 844356656.0, "step": 22234 }, { "epoch": 2.8285205444599923, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5772290229797363, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8730675578117371, "num_tokens": 844389789.0, "step": 22235 }, { "epoch": 2.828647754738583, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.728285789489746, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.879202127456665, "num_tokens": 844423599.0, "step": 22236 }, { "epoch": 2.8287749650171734, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4669559001922607, "learning_rate": 1e-05, "loss": 0.4274, "mean_token_accuracy": 0.8817960023880005, "num_tokens": 844460117.0, "step": 22237 }, { "epoch": 2.828902175295764, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7429637908935547, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8683706521987915, "num_tokens": 844501263.0, "step": 22238 }, { "epoch": 2.8290293855743545, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3693013191223145, "learning_rate": 1e-05, "loss": 0.3979, "mean_token_accuracy": 0.8812238574028015, "num_tokens": 844541573.0, "step": 22239 }, { "epoch": 2.829156595852945, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.8380680084228516, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8776125907897949, "num_tokens": 844583728.0, "step": 22240 }, { "epoch": 2.8292838061315355, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3227763175964355, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8806323409080505, "num_tokens": 844616611.0, "step": 22241 }, { "epoch": 2.829411016410126, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9956953525543213, "learning_rate": 1e-05, "loss": 0.4949, "mean_token_accuracy": 0.8614140748977661, "num_tokens": 844652067.0, "step": 22242 }, { "epoch": 2.8295382266887166, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.309692144393921, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8718295097351074, "num_tokens": 844688582.0, "step": 22243 }, { "epoch": 2.829665436967307, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.147730350494385, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8787124156951904, "num_tokens": 844727120.0, "step": 22244 }, { "epoch": 2.8297926472458976, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.391692876815796, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8793644905090332, "num_tokens": 844772328.0, "step": 22245 }, { "epoch": 2.829919857524488, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.246482849121094, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.8922892212867737, "num_tokens": 844812528.0, "step": 22246 }, { "epoch": 2.8300470678030782, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.592963695526123, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8718525767326355, "num_tokens": 844857208.0, "step": 22247 }, { "epoch": 2.830174278081669, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.170917987823486, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8701573014259338, "num_tokens": 844890362.0, "step": 22248 }, { "epoch": 2.8303014883602593, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7316222190856934, "learning_rate": 1e-05, "loss": 0.4824, "mean_token_accuracy": 0.8646281957626343, "num_tokens": 844928664.0, "step": 22249 }, { "epoch": 2.8304286986388503, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9484806060791016, "learning_rate": 1e-05, "loss": 0.3893, "mean_token_accuracy": 0.8971948623657227, "num_tokens": 844957043.0, "step": 22250 }, { "epoch": 2.8305559089174404, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.7407803535461426, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8749891519546509, "num_tokens": 844991808.0, "step": 22251 }, { "epoch": 2.830683119196031, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7292864322662354, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8799118995666504, "num_tokens": 845025571.0, "step": 22252 }, { "epoch": 2.8308103294746214, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7133612632751465, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8692032098770142, "num_tokens": 845063154.0, "step": 22253 }, { "epoch": 2.830937539753212, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5452723503112793, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8863710165023804, "num_tokens": 845100256.0, "step": 22254 }, { "epoch": 2.8310647500318025, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7141270637512207, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8778091669082642, "num_tokens": 845138111.0, "step": 22255 }, { "epoch": 2.831191960310393, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4272899627685547, "learning_rate": 1e-05, "loss": 0.4119, "mean_token_accuracy": 0.8828606605529785, "num_tokens": 845175759.0, "step": 22256 }, { "epoch": 2.8313191705889835, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7462856769561768, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8780304193496704, "num_tokens": 845212232.0, "step": 22257 }, { "epoch": 2.831446380867574, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3521065711975098, "learning_rate": 1e-05, "loss": 0.3775, "mean_token_accuracy": 0.8956360816955566, "num_tokens": 845247141.0, "step": 22258 }, { "epoch": 2.8315735911461646, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.807643175125122, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.883272647857666, "num_tokens": 845287908.0, "step": 22259 }, { "epoch": 2.831700801424755, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2824416160583496, "learning_rate": 1e-05, "loss": 0.4157, "mean_token_accuracy": 0.8822526931762695, "num_tokens": 845328134.0, "step": 22260 }, { "epoch": 2.8318280117033456, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9456279277801514, "learning_rate": 1e-05, "loss": 0.4328, "mean_token_accuracy": 0.8829852938652039, "num_tokens": 845366096.0, "step": 22261 }, { "epoch": 2.831955221981936, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.280956745147705, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8798962831497192, "num_tokens": 845397737.0, "step": 22262 }, { "epoch": 2.8320824322605267, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.098907947540283, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8816601634025574, "num_tokens": 845437015.0, "step": 22263 }, { "epoch": 2.8322096425391172, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.3161590099334717, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.8898351192474365, "num_tokens": 845473778.0, "step": 22264 }, { "epoch": 2.8323368528177078, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.250190258026123, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8915989995002747, "num_tokens": 845509254.0, "step": 22265 }, { "epoch": 2.8324640630962983, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.489649534225464, "learning_rate": 1e-05, "loss": 0.4665, "mean_token_accuracy": 0.8715521693229675, "num_tokens": 845547682.0, "step": 22266 }, { "epoch": 2.832591273374889, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.25964879989624, "learning_rate": 1e-05, "loss": 0.4054, "mean_token_accuracy": 0.8895276188850403, "num_tokens": 845585315.0, "step": 22267 }, { "epoch": 2.8327184836534793, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6539056301116943, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8839742541313171, "num_tokens": 845625065.0, "step": 22268 }, { "epoch": 2.83284569393207, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.125181198120117, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.88784259557724, "num_tokens": 845664148.0, "step": 22269 }, { "epoch": 2.83297290421066, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.725188732147217, "learning_rate": 1e-05, "loss": 0.413, "mean_token_accuracy": 0.8830779790878296, "num_tokens": 845695986.0, "step": 22270 }, { "epoch": 2.833100114489251, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.965458631515503, "learning_rate": 1e-05, "loss": 0.4481, "mean_token_accuracy": 0.879187285900116, "num_tokens": 845732807.0, "step": 22271 }, { "epoch": 2.833227324767841, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.698573112487793, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.8858177661895752, "num_tokens": 845763238.0, "step": 22272 }, { "epoch": 2.833354535046432, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7705085277557373, "learning_rate": 1e-05, "loss": 0.4, "mean_token_accuracy": 0.8916230201721191, "num_tokens": 845802218.0, "step": 22273 }, { "epoch": 2.833481745325022, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6389706134796143, "learning_rate": 1e-05, "loss": 0.4975, "mean_token_accuracy": 0.8584941029548645, "num_tokens": 845845056.0, "step": 22274 }, { "epoch": 2.833608955603613, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6641721725463867, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.876652181148529, "num_tokens": 845884909.0, "step": 22275 }, { "epoch": 2.833736165882203, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5777082443237305, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8769106864929199, "num_tokens": 845923894.0, "step": 22276 }, { "epoch": 2.8338633761607936, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.62764573097229, "learning_rate": 1e-05, "loss": 0.4926, "mean_token_accuracy": 0.8599640727043152, "num_tokens": 845962452.0, "step": 22277 }, { "epoch": 2.833990586439384, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5154075622558594, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8782776594161987, "num_tokens": 845998031.0, "step": 22278 }, { "epoch": 2.8341177967179747, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6002211570739746, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8743287324905396, "num_tokens": 846036920.0, "step": 22279 }, { "epoch": 2.8342450069965652, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.507566213607788, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8665773272514343, "num_tokens": 846075978.0, "step": 22280 }, { "epoch": 2.8343722172751558, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5800719261169434, "learning_rate": 1e-05, "loss": 0.4342, "mean_token_accuracy": 0.8780556917190552, "num_tokens": 846111893.0, "step": 22281 }, { "epoch": 2.8344994275537463, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.44622540473938, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8756306767463684, "num_tokens": 846153417.0, "step": 22282 }, { "epoch": 2.834626637832337, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5819008350372314, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.872795581817627, "num_tokens": 846194375.0, "step": 22283 }, { "epoch": 2.8347538481109273, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3896684646606445, "learning_rate": 1e-05, "loss": 0.406, "mean_token_accuracy": 0.8836534023284912, "num_tokens": 846231040.0, "step": 22284 }, { "epoch": 2.834881058389518, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6310160160064697, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8880212306976318, "num_tokens": 846270645.0, "step": 22285 }, { "epoch": 2.8350082686681084, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3368160724639893, "learning_rate": 1e-05, "loss": 0.4056, "mean_token_accuracy": 0.8842728137969971, "num_tokens": 846304038.0, "step": 22286 }, { "epoch": 2.835135478946699, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7502918243408203, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8683182001113892, "num_tokens": 846338136.0, "step": 22287 }, { "epoch": 2.8352626892252895, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2646358013153076, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8770226240158081, "num_tokens": 846372654.0, "step": 22288 }, { "epoch": 2.83538989950388, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.892271041870117, "learning_rate": 1e-05, "loss": 0.4451, "mean_token_accuracy": 0.8760457634925842, "num_tokens": 846409180.0, "step": 22289 }, { "epoch": 2.8355171097824705, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2009010314941406, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8666543364524841, "num_tokens": 846453410.0, "step": 22290 }, { "epoch": 2.835644320061061, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.127624988555908, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8866244554519653, "num_tokens": 846488271.0, "step": 22291 }, { "epoch": 2.8357715303396516, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.2430238723754883, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8853834867477417, "num_tokens": 846528730.0, "step": 22292 }, { "epoch": 2.835898740618242, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.369935512542725, "learning_rate": 1e-05, "loss": 0.4024, "mean_token_accuracy": 0.8911490440368652, "num_tokens": 846567276.0, "step": 22293 }, { "epoch": 2.8360259508968326, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.4857099056243896, "learning_rate": 1e-05, "loss": 0.4248, "mean_token_accuracy": 0.8848615884780884, "num_tokens": 846608482.0, "step": 22294 }, { "epoch": 2.8361531611754227, "ewc_loss": 0.107421875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.453857898712158, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8742899894714355, "num_tokens": 846648582.0, "step": 22295 }, { "epoch": 2.8362803714540137, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.781381845474243, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8873792290687561, "num_tokens": 846687392.0, "step": 22296 }, { "epoch": 2.8364075817326038, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.28236198425293, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8843004703521729, "num_tokens": 846721542.0, "step": 22297 }, { "epoch": 2.8365347920111947, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.8579375743865967, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.8690212965011597, "num_tokens": 846759262.0, "step": 22298 }, { "epoch": 2.836662002289785, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.055604934692383, "learning_rate": 1e-05, "loss": 0.4662, "mean_token_accuracy": 0.8734996318817139, "num_tokens": 846796979.0, "step": 22299 }, { "epoch": 2.836789212568376, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.7915637493133545, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8779242038726807, "num_tokens": 846832417.0, "step": 22300 }, { "epoch": 2.836916422846966, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8708224296569824, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8737045526504517, "num_tokens": 846870289.0, "step": 22301 }, { "epoch": 2.8370436331255564, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.707825183868408, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8793494701385498, "num_tokens": 846909821.0, "step": 22302 }, { "epoch": 2.837170843404147, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7501864433288574, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8802869319915771, "num_tokens": 846945512.0, "step": 22303 }, { "epoch": 2.8372980536827375, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5972259044647217, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8852166533470154, "num_tokens": 846980834.0, "step": 22304 }, { "epoch": 2.837425263961328, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6855032444000244, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.880954384803772, "num_tokens": 847025320.0, "step": 22305 }, { "epoch": 2.8375524742399185, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.502509355545044, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8882397413253784, "num_tokens": 847063637.0, "step": 22306 }, { "epoch": 2.837679684518509, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.687443256378174, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.88688063621521, "num_tokens": 847102314.0, "step": 22307 }, { "epoch": 2.8378068947970996, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4144532680511475, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.8820477724075317, "num_tokens": 847141956.0, "step": 22308 }, { "epoch": 2.83793410507569, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.755195140838623, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8707795143127441, "num_tokens": 847178855.0, "step": 22309 }, { "epoch": 2.8380613153542806, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3387887477874756, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8802391886711121, "num_tokens": 847214228.0, "step": 22310 }, { "epoch": 2.838188525632871, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8602752685546875, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.876248836517334, "num_tokens": 847255942.0, "step": 22311 }, { "epoch": 2.8383157359114617, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.286440134048462, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8657183647155762, "num_tokens": 847294742.0, "step": 22312 }, { "epoch": 2.838442946190052, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.034636974334717, "learning_rate": 1e-05, "loss": 0.4533, "mean_token_accuracy": 0.876460075378418, "num_tokens": 847330105.0, "step": 22313 }, { "epoch": 2.8385701564686427, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.2936322689056396, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8724881410598755, "num_tokens": 847373291.0, "step": 22314 }, { "epoch": 2.8386973667472333, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.211756706237793, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8902735710144043, "num_tokens": 847409059.0, "step": 22315 }, { "epoch": 2.838824577025824, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.4456937313079834, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.8834596276283264, "num_tokens": 847445869.0, "step": 22316 }, { "epoch": 2.8389517873044143, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.269405364990234, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8825551867485046, "num_tokens": 847486816.0, "step": 22317 }, { "epoch": 2.839078997583005, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.6315228939056396, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8858045339584351, "num_tokens": 847527482.0, "step": 22318 }, { "epoch": 2.8392062078615954, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.153316497802734, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8666996359825134, "num_tokens": 847570850.0, "step": 22319 }, { "epoch": 2.8393334181401855, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.741591215133667, "learning_rate": 1e-05, "loss": 0.4607, "mean_token_accuracy": 0.8744257688522339, "num_tokens": 847608872.0, "step": 22320 }, { "epoch": 2.8394606284187764, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.940584421157837, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8723151087760925, "num_tokens": 847647813.0, "step": 22321 }, { "epoch": 2.8395878386973665, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.726034641265869, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8813993334770203, "num_tokens": 847685726.0, "step": 22322 }, { "epoch": 2.8397150489759575, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7466509342193604, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8721784353256226, "num_tokens": 847723029.0, "step": 22323 }, { "epoch": 2.8398422592545476, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6778745651245117, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8847524523735046, "num_tokens": 847761922.0, "step": 22324 }, { "epoch": 2.839969469533138, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5933926105499268, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8739932775497437, "num_tokens": 847806229.0, "step": 22325 }, { "epoch": 2.8400966798117286, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.648756504058838, "learning_rate": 1e-05, "loss": 0.4465, "mean_token_accuracy": 0.8768095970153809, "num_tokens": 847844448.0, "step": 22326 }, { "epoch": 2.840223890090319, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4991674423217773, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8690191507339478, "num_tokens": 847889387.0, "step": 22327 }, { "epoch": 2.8403511003689097, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6310243606567383, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8757990598678589, "num_tokens": 847926993.0, "step": 22328 }, { "epoch": 2.8404783106475002, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4359755516052246, "learning_rate": 1e-05, "loss": 0.403, "mean_token_accuracy": 0.8875864148139954, "num_tokens": 847958884.0, "step": 22329 }, { "epoch": 2.8406055209260908, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6436257362365723, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8759627342224121, "num_tokens": 847997894.0, "step": 22330 }, { "epoch": 2.8407327312046813, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3851444721221924, "learning_rate": 1e-05, "loss": 0.4364, "mean_token_accuracy": 0.8782798647880554, "num_tokens": 848037392.0, "step": 22331 }, { "epoch": 2.840859941483272, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6750526428222656, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8784309029579163, "num_tokens": 848072844.0, "step": 22332 }, { "epoch": 2.8409871517618623, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3362302780151367, "learning_rate": 1e-05, "loss": 0.4678, "mean_token_accuracy": 0.8659286499023438, "num_tokens": 848107477.0, "step": 22333 }, { "epoch": 2.841114362040453, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.752365827560425, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8782262206077576, "num_tokens": 848143124.0, "step": 22334 }, { "epoch": 2.8412415723190434, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2636024951934814, "learning_rate": 1e-05, "loss": 0.4257, "mean_token_accuracy": 0.8813875913619995, "num_tokens": 848181454.0, "step": 22335 }, { "epoch": 2.841368782597634, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.897547960281372, "learning_rate": 1e-05, "loss": 0.4149, "mean_token_accuracy": 0.886081337928772, "num_tokens": 848219300.0, "step": 22336 }, { "epoch": 2.8414959928762245, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.206918954849243, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8823989629745483, "num_tokens": 848256650.0, "step": 22337 }, { "epoch": 2.841623203154815, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.126579761505127, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8708086013793945, "num_tokens": 848294836.0, "step": 22338 }, { "epoch": 2.8417504134334055, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.2670998573303223, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8803461790084839, "num_tokens": 848326306.0, "step": 22339 }, { "epoch": 2.841877623711996, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.340089321136475, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8882246017456055, "num_tokens": 848360886.0, "step": 22340 }, { "epoch": 2.8420048339905866, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4894609451293945, "learning_rate": 1e-05, "loss": 0.4441, "mean_token_accuracy": 0.877474308013916, "num_tokens": 848399364.0, "step": 22341 }, { "epoch": 2.842132044269177, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.380773544311523, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8851702213287354, "num_tokens": 848436978.0, "step": 22342 }, { "epoch": 2.8422592545477676, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.7296292781829834, "learning_rate": 1e-05, "loss": 0.4874, "mean_token_accuracy": 0.8661352396011353, "num_tokens": 848474661.0, "step": 22343 }, { "epoch": 2.842386464826358, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.238348007202148, "learning_rate": 1e-05, "loss": 0.4152, "mean_token_accuracy": 0.8888586759567261, "num_tokens": 848514469.0, "step": 22344 }, { "epoch": 2.8425136751049482, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.8025381565093994, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.88292396068573, "num_tokens": 848548606.0, "step": 22345 }, { "epoch": 2.842640885383539, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.042797565460205, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8717389106750488, "num_tokens": 848585309.0, "step": 22346 }, { "epoch": 2.8427680956621293, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.749412775039673, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8800599575042725, "num_tokens": 848625062.0, "step": 22347 }, { "epoch": 2.8428953059407203, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8692615032196045, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8768210411071777, "num_tokens": 848661914.0, "step": 22348 }, { "epoch": 2.8430225162193103, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.6812355518341064, "learning_rate": 1e-05, "loss": 0.5327, "mean_token_accuracy": 0.8492670059204102, "num_tokens": 848704343.0, "step": 22349 }, { "epoch": 2.843149726497901, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.743762731552124, "learning_rate": 1e-05, "loss": 0.3968, "mean_token_accuracy": 0.893265426158905, "num_tokens": 848737415.0, "step": 22350 }, { "epoch": 2.8432769367764914, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6021621227264404, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8770142793655396, "num_tokens": 848770795.0, "step": 22351 }, { "epoch": 2.843404147055082, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6695611476898193, "learning_rate": 1e-05, "loss": 0.4064, "mean_token_accuracy": 0.8848183751106262, "num_tokens": 848815086.0, "step": 22352 }, { "epoch": 2.8435313573336725, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.515814781188965, "learning_rate": 1e-05, "loss": 0.376, "mean_token_accuracy": 0.8961728811264038, "num_tokens": 848853545.0, "step": 22353 }, { "epoch": 2.843658567612263, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6650214195251465, "learning_rate": 1e-05, "loss": 0.4125, "mean_token_accuracy": 0.8845297694206238, "num_tokens": 848886047.0, "step": 22354 }, { "epoch": 2.8437857778908535, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.438035488128662, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8826942443847656, "num_tokens": 848929147.0, "step": 22355 }, { "epoch": 2.843912988169444, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.7159969806671143, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.877086341381073, "num_tokens": 848962912.0, "step": 22356 }, { "epoch": 2.8440401984480346, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.360323429107666, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8781909942626953, "num_tokens": 848997954.0, "step": 22357 }, { "epoch": 2.844167408726625, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7995874881744385, "learning_rate": 1e-05, "loss": 0.4444, "mean_token_accuracy": 0.8752039670944214, "num_tokens": 849044418.0, "step": 22358 }, { "epoch": 2.8442946190052156, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2860071659088135, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8772112131118774, "num_tokens": 849085591.0, "step": 22359 }, { "epoch": 2.844421829283806, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9548330307006836, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8753985166549683, "num_tokens": 849128046.0, "step": 22360 }, { "epoch": 2.8445490395623967, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2812232971191406, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8725842237472534, "num_tokens": 849165231.0, "step": 22361 }, { "epoch": 2.844676249840987, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.131321430206299, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8803527355194092, "num_tokens": 849203414.0, "step": 22362 }, { "epoch": 2.8448034601195777, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3527486324310303, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8835513591766357, "num_tokens": 849242463.0, "step": 22363 }, { "epoch": 2.8449306703981683, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.263474464416504, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8827760219573975, "num_tokens": 849280098.0, "step": 22364 }, { "epoch": 2.845057880676759, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.5536258220672607, "learning_rate": 1e-05, "loss": 0.4199, "mean_token_accuracy": 0.8826513290405273, "num_tokens": 849317460.0, "step": 22365 }, { "epoch": 2.8451850909553493, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.208504676818848, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8831086754798889, "num_tokens": 849354852.0, "step": 22366 }, { "epoch": 2.84531230123394, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6938514709472656, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8783177137374878, "num_tokens": 849397132.0, "step": 22367 }, { "epoch": 2.84543951151253, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.032627582550049, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8813270330429077, "num_tokens": 849441751.0, "step": 22368 }, { "epoch": 2.845566721791121, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.7203478813171387, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8722478747367859, "num_tokens": 849480177.0, "step": 22369 }, { "epoch": 2.845693932069711, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8417975902557373, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8715741634368896, "num_tokens": 849520211.0, "step": 22370 }, { "epoch": 2.845821142348302, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6916632652282715, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.879096508026123, "num_tokens": 849551591.0, "step": 22371 }, { "epoch": 2.845948352626892, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.670650005340576, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8710103034973145, "num_tokens": 849592700.0, "step": 22372 }, { "epoch": 2.846075562905483, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.642740488052368, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.8692273497581482, "num_tokens": 849631220.0, "step": 22373 }, { "epoch": 2.846202773184073, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5615415573120117, "learning_rate": 1e-05, "loss": 0.4158, "mean_token_accuracy": 0.8848809003829956, "num_tokens": 849670349.0, "step": 22374 }, { "epoch": 2.8463299834626636, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.614365577697754, "learning_rate": 1e-05, "loss": 0.5158, "mean_token_accuracy": 0.8531407117843628, "num_tokens": 849711000.0, "step": 22375 }, { "epoch": 2.846457193741254, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4942920207977295, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8653854131698608, "num_tokens": 849751199.0, "step": 22376 }, { "epoch": 2.8465844040198447, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.592921733856201, "learning_rate": 1e-05, "loss": 0.3694, "mean_token_accuracy": 0.9015007019042969, "num_tokens": 849785416.0, "step": 22377 }, { "epoch": 2.8467116142984352, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.449615478515625, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8687770962715149, "num_tokens": 849818553.0, "step": 22378 }, { "epoch": 2.8468388245770258, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5879292488098145, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8848560452461243, "num_tokens": 849854994.0, "step": 22379 }, { "epoch": 2.8469660348556163, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4133710861206055, "learning_rate": 1e-05, "loss": 0.4736, "mean_token_accuracy": 0.8608377575874329, "num_tokens": 849896603.0, "step": 22380 }, { "epoch": 2.847093245134207, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6116487979888916, "learning_rate": 1e-05, "loss": 0.4719, "mean_token_accuracy": 0.8676114082336426, "num_tokens": 849935999.0, "step": 22381 }, { "epoch": 2.8472204554127973, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3545236587524414, "learning_rate": 1e-05, "loss": 0.4065, "mean_token_accuracy": 0.8839172124862671, "num_tokens": 849976858.0, "step": 22382 }, { "epoch": 2.847347665691388, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6682538986206055, "learning_rate": 1e-05, "loss": 0.399, "mean_token_accuracy": 0.8921098709106445, "num_tokens": 850012802.0, "step": 22383 }, { "epoch": 2.8474748759699784, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.310029983520508, "learning_rate": 1e-05, "loss": 0.4502, "mean_token_accuracy": 0.8721351623535156, "num_tokens": 850047330.0, "step": 22384 }, { "epoch": 2.847602086248569, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7741472721099854, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8729285001754761, "num_tokens": 850087543.0, "step": 22385 }, { "epoch": 2.8477292965271594, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.2214903831481934, "learning_rate": 1e-05, "loss": 0.4058, "mean_token_accuracy": 0.8834505677223206, "num_tokens": 850126864.0, "step": 22386 }, { "epoch": 2.84785650680575, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.961756706237793, "learning_rate": 1e-05, "loss": 0.4031, "mean_token_accuracy": 0.890258252620697, "num_tokens": 850162476.0, "step": 22387 }, { "epoch": 2.8479837170843405, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.18745756149292, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8764014840126038, "num_tokens": 850197606.0, "step": 22388 }, { "epoch": 2.848110927362931, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.234079837799072, "learning_rate": 1e-05, "loss": 0.4519, "mean_token_accuracy": 0.8781879544258118, "num_tokens": 850232223.0, "step": 22389 }, { "epoch": 2.8482381376415216, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.2889342308044434, "learning_rate": 1e-05, "loss": 0.4126, "mean_token_accuracy": 0.8857935667037964, "num_tokens": 850268936.0, "step": 22390 }, { "epoch": 2.848365347920112, "ewc_loss": 0.107421875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.47778844833374, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8725311160087585, "num_tokens": 850301652.0, "step": 22391 }, { "epoch": 2.8484925581987026, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.6175224781036377, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.881266713142395, "num_tokens": 850335033.0, "step": 22392 }, { "epoch": 2.8486197684772927, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 7.104873657226562e-05, "grad_norm": 4.45825719833374, "learning_rate": 1e-05, "loss": 0.4516, "mean_token_accuracy": 0.8777896165847778, "num_tokens": 850369677.0, "step": 22393 }, { "epoch": 2.8487469787558837, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.864198923110962, "learning_rate": 1e-05, "loss": 0.4944, "mean_token_accuracy": 0.8645229935646057, "num_tokens": 850411464.0, "step": 22394 }, { "epoch": 2.8488741890344738, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.247391223907471, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8805960416793823, "num_tokens": 850447578.0, "step": 22395 }, { "epoch": 2.8490013993130647, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.864379644393921, "learning_rate": 1e-05, "loss": 0.4237, "mean_token_accuracy": 0.8839446306228638, "num_tokens": 850489689.0, "step": 22396 }, { "epoch": 2.849128609591655, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.030704498291016, "learning_rate": 1e-05, "loss": 0.4227, "mean_token_accuracy": 0.8857585191726685, "num_tokens": 850528598.0, "step": 22397 }, { "epoch": 2.849255819870246, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.780055522918701, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8804693222045898, "num_tokens": 850566953.0, "step": 22398 }, { "epoch": 2.849383030148836, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8699183464050293, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8861932754516602, "num_tokens": 850604286.0, "step": 22399 }, { "epoch": 2.8495102404274264, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.681086540222168, "learning_rate": 1e-05, "loss": 0.461, "mean_token_accuracy": 0.8749794363975525, "num_tokens": 850640724.0, "step": 22400 }, { "epoch": 2.849637450706017, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.772808074951172, "learning_rate": 1e-05, "loss": 0.4457, "mean_token_accuracy": 0.8771599531173706, "num_tokens": 850682377.0, "step": 22401 }, { "epoch": 2.8497646609846075, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.568326234817505, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8770188093185425, "num_tokens": 850721637.0, "step": 22402 }, { "epoch": 2.849891871263198, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7519469261169434, "learning_rate": 1e-05, "loss": 0.4528, "mean_token_accuracy": 0.8724634647369385, "num_tokens": 850754884.0, "step": 22403 }, { "epoch": 2.8500190815417885, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.4707930088043213, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8738605976104736, "num_tokens": 850793706.0, "step": 22404 }, { "epoch": 2.850146291820379, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.789844274520874, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8676729202270508, "num_tokens": 850827708.0, "step": 22405 }, { "epoch": 2.8502735020989696, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3716554641723633, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8708237409591675, "num_tokens": 850870808.0, "step": 22406 }, { "epoch": 2.85040071237756, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8905820846557617, "learning_rate": 1e-05, "loss": 0.4308, "mean_token_accuracy": 0.8805372714996338, "num_tokens": 850906129.0, "step": 22407 }, { "epoch": 2.8505279226561506, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3135361671447754, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8801589012145996, "num_tokens": 850944021.0, "step": 22408 }, { "epoch": 2.850655132934741, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.065002918243408, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8839007019996643, "num_tokens": 850983040.0, "step": 22409 }, { "epoch": 2.8507823432133317, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3419220447540283, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8790466785430908, "num_tokens": 851017400.0, "step": 22410 }, { "epoch": 2.850909553491922, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.218780040740967, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8797022700309753, "num_tokens": 851055559.0, "step": 22411 }, { "epoch": 2.8510367637705127, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.4866974353790283, "learning_rate": 1e-05, "loss": 0.4084, "mean_token_accuracy": 0.8880285024642944, "num_tokens": 851088810.0, "step": 22412 }, { "epoch": 2.8511639740491033, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.254405498504639, "learning_rate": 1e-05, "loss": 0.4556, "mean_token_accuracy": 0.8779441118240356, "num_tokens": 851121531.0, "step": 22413 }, { "epoch": 2.851291184327694, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.673018217086792, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8781434297561646, "num_tokens": 851160601.0, "step": 22414 }, { "epoch": 2.8514183946062843, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.097815990447998, "learning_rate": 1e-05, "loss": 0.4723, "mean_token_accuracy": 0.8673672080039978, "num_tokens": 851195887.0, "step": 22415 }, { "epoch": 2.851545604884875, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003528594970703125, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.751208782196045, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8838917016983032, "num_tokens": 851231761.0, "step": 22416 }, { "epoch": 2.8516728151634654, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.8670661449432373, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8617846965789795, "num_tokens": 851270765.0, "step": 22417 }, { "epoch": 2.8518000254420555, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.7310333251953125, "learning_rate": 1e-05, "loss": 0.501, "mean_token_accuracy": 0.8643754720687866, "num_tokens": 851312200.0, "step": 22418 }, { "epoch": 2.8519272357206464, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.660616636276245, "learning_rate": 1e-05, "loss": 0.4526, "mean_token_accuracy": 0.8734612464904785, "num_tokens": 851350854.0, "step": 22419 }, { "epoch": 2.8520544459992365, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7026805877685547, "learning_rate": 1e-05, "loss": 0.4687, "mean_token_accuracy": 0.8657963871955872, "num_tokens": 851389242.0, "step": 22420 }, { "epoch": 2.8521816562778275, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5152108669281006, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8762240409851074, "num_tokens": 851425450.0, "step": 22421 }, { "epoch": 2.8523088665564176, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.695101022720337, "learning_rate": 1e-05, "loss": 0.4362, "mean_token_accuracy": 0.8770073652267456, "num_tokens": 851459470.0, "step": 22422 }, { "epoch": 2.852436076835008, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.419053077697754, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8688565492630005, "num_tokens": 851500849.0, "step": 22423 }, { "epoch": 2.8525632871135986, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.717224597930908, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8787609338760376, "num_tokens": 851540048.0, "step": 22424 }, { "epoch": 2.852690497392189, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3601975440979004, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8759885430335999, "num_tokens": 851573343.0, "step": 22425 }, { "epoch": 2.8528177076707797, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7807669639587402, "learning_rate": 1e-05, "loss": 0.4812, "mean_token_accuracy": 0.8616129159927368, "num_tokens": 851613658.0, "step": 22426 }, { "epoch": 2.85294491794937, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2906415462493896, "learning_rate": 1e-05, "loss": 0.3968, "mean_token_accuracy": 0.8885701298713684, "num_tokens": 851649961.0, "step": 22427 }, { "epoch": 2.8530721282279607, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.9089057445526123, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.8770818114280701, "num_tokens": 851685465.0, "step": 22428 }, { "epoch": 2.8531993385065513, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.252736806869507, "learning_rate": 1e-05, "loss": 0.3799, "mean_token_accuracy": 0.8937880992889404, "num_tokens": 851718434.0, "step": 22429 }, { "epoch": 2.853326548785142, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.091142654418945, "learning_rate": 1e-05, "loss": 0.4306, "mean_token_accuracy": 0.8850531578063965, "num_tokens": 851756207.0, "step": 22430 }, { "epoch": 2.8534537590637323, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.299002170562744, "learning_rate": 1e-05, "loss": 0.4563, "mean_token_accuracy": 0.8738363981246948, "num_tokens": 851795398.0, "step": 22431 }, { "epoch": 2.853580969342323, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.269446849822998, "learning_rate": 1e-05, "loss": 0.4853, "mean_token_accuracy": 0.8672590851783752, "num_tokens": 851832569.0, "step": 22432 }, { "epoch": 2.8537081796209134, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.4646849632263184, "learning_rate": 1e-05, "loss": 0.4349, "mean_token_accuracy": 0.8826366662979126, "num_tokens": 851870749.0, "step": 22433 }, { "epoch": 2.853835389899504, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.315662860870361, "learning_rate": 1e-05, "loss": 0.4432, "mean_token_accuracy": 0.882036030292511, "num_tokens": 851912591.0, "step": 22434 }, { "epoch": 2.8539626001780944, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6761672496795654, "learning_rate": 1e-05, "loss": 0.4518, "mean_token_accuracy": 0.8775181770324707, "num_tokens": 851948792.0, "step": 22435 }, { "epoch": 2.854089810456685, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.190764904022217, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8699343204498291, "num_tokens": 851988811.0, "step": 22436 }, { "epoch": 2.8542170207352755, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.755060911178589, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8841835856437683, "num_tokens": 852024781.0, "step": 22437 }, { "epoch": 2.854344231013866, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.003880023956299, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8836871385574341, "num_tokens": 852056676.0, "step": 22438 }, { "epoch": 2.8544714412924566, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.725510835647583, "learning_rate": 1e-05, "loss": 0.4814, "mean_token_accuracy": 0.8669931888580322, "num_tokens": 852095500.0, "step": 22439 }, { "epoch": 2.854598651571047, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.814107894897461, "learning_rate": 1e-05, "loss": 0.4097, "mean_token_accuracy": 0.8888288140296936, "num_tokens": 852133658.0, "step": 22440 }, { "epoch": 2.8547258618496376, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6672396659851074, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8619276285171509, "num_tokens": 852173963.0, "step": 22441 }, { "epoch": 2.854853072128228, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.703855037689209, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8766699433326721, "num_tokens": 852209957.0, "step": 22442 }, { "epoch": 2.8549802824068182, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.583326816558838, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.8790969848632812, "num_tokens": 852252281.0, "step": 22443 }, { "epoch": 2.855107492685409, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.645015239715576, "learning_rate": 1e-05, "loss": 0.3833, "mean_token_accuracy": 0.8950859904289246, "num_tokens": 852285287.0, "step": 22444 }, { "epoch": 2.8552347029639993, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5353801250457764, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8692137598991394, "num_tokens": 852326596.0, "step": 22445 }, { "epoch": 2.8553619132425903, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6224067211151123, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8846761584281921, "num_tokens": 852358062.0, "step": 22446 }, { "epoch": 2.8554891235211803, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4667129516601562, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8688560724258423, "num_tokens": 852395828.0, "step": 22447 }, { "epoch": 2.855616333799771, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.740643262863159, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8741645812988281, "num_tokens": 852439466.0, "step": 22448 }, { "epoch": 2.8557435440783614, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3560116291046143, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8802869319915771, "num_tokens": 852477992.0, "step": 22449 }, { "epoch": 2.855870754356952, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7852420806884766, "learning_rate": 1e-05, "loss": 0.4242, "mean_token_accuracy": 0.8848238587379456, "num_tokens": 852514814.0, "step": 22450 }, { "epoch": 2.8559979646355425, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.276308059692383, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8809465169906616, "num_tokens": 852553866.0, "step": 22451 }, { "epoch": 2.856125174914133, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.948472023010254, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8683119416236877, "num_tokens": 852594291.0, "step": 22452 }, { "epoch": 2.8562523851927235, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.2630727291107178, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8835779428482056, "num_tokens": 852630586.0, "step": 22453 }, { "epoch": 2.856379595471314, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.138947486877441, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.8745750188827515, "num_tokens": 852670013.0, "step": 22454 }, { "epoch": 2.8565068057499046, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3526909351348877, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8769546747207642, "num_tokens": 852704097.0, "step": 22455 }, { "epoch": 2.856634016028495, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.278650760650635, "learning_rate": 1e-05, "loss": 0.4633, "mean_token_accuracy": 0.8718786835670471, "num_tokens": 852739771.0, "step": 22456 }, { "epoch": 2.8567612263070856, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.552769899368286, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8681987524032593, "num_tokens": 852778604.0, "step": 22457 }, { "epoch": 2.856888436585676, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.235240459442139, "learning_rate": 1e-05, "loss": 0.5181, "mean_token_accuracy": 0.8571293354034424, "num_tokens": 852814966.0, "step": 22458 }, { "epoch": 2.8570156468642667, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7131078243255615, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8691006898880005, "num_tokens": 852849562.0, "step": 22459 }, { "epoch": 2.857142857142857, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.0497565269470215, "learning_rate": 1e-05, "loss": 0.4299, "mean_token_accuracy": 0.8827769756317139, "num_tokens": 852884183.0, "step": 22460 }, { "epoch": 2.8572700674214477, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7294564247131348, "learning_rate": 1e-05, "loss": 0.4893, "mean_token_accuracy": 0.8695802092552185, "num_tokens": 852925121.0, "step": 22461 }, { "epoch": 2.8573972777000383, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.8445582389831543, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.884067952632904, "num_tokens": 852965962.0, "step": 22462 }, { "epoch": 2.857524487978629, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.6895430088043213, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8758496046066284, "num_tokens": 853004886.0, "step": 22463 }, { "epoch": 2.8576516982572193, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.686713933944702, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8728751540184021, "num_tokens": 853038898.0, "step": 22464 }, { "epoch": 2.85777890853581, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6500120162963867, "learning_rate": 1e-05, "loss": 0.4864, "mean_token_accuracy": 0.8620510101318359, "num_tokens": 853079064.0, "step": 22465 }, { "epoch": 2.8579061188144, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.567298650741577, "learning_rate": 1e-05, "loss": 0.496, "mean_token_accuracy": 0.8597264885902405, "num_tokens": 853125704.0, "step": 22466 }, { "epoch": 2.858033329092991, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.614039659500122, "learning_rate": 1e-05, "loss": 0.4534, "mean_token_accuracy": 0.873796820640564, "num_tokens": 853168760.0, "step": 22467 }, { "epoch": 2.858160539371581, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.498772621154785, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8776894807815552, "num_tokens": 853207146.0, "step": 22468 }, { "epoch": 2.858287749650172, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5935590267181396, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.8840930461883545, "num_tokens": 853248764.0, "step": 22469 }, { "epoch": 2.858414959928762, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.450591564178467, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8837772011756897, "num_tokens": 853283325.0, "step": 22470 }, { "epoch": 2.858542170207353, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6140432357788086, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8772176504135132, "num_tokens": 853316579.0, "step": 22471 }, { "epoch": 2.858669380485943, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4034111499786377, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8813139200210571, "num_tokens": 853353109.0, "step": 22472 }, { "epoch": 2.8587965907645336, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.63004732131958, "learning_rate": 1e-05, "loss": 0.4383, "mean_token_accuracy": 0.8799328804016113, "num_tokens": 853394974.0, "step": 22473 }, { "epoch": 2.858923801043124, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3398361206054688, "learning_rate": 1e-05, "loss": 0.3832, "mean_token_accuracy": 0.89393550157547, "num_tokens": 853433819.0, "step": 22474 }, { "epoch": 2.8590510113217147, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7053306102752686, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8768401145935059, "num_tokens": 853471947.0, "step": 22475 }, { "epoch": 2.859178221600305, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.282193899154663, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8770337104797363, "num_tokens": 853505936.0, "step": 22476 }, { "epoch": 2.8593054318788957, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8271727561950684, "learning_rate": 1e-05, "loss": 0.3882, "mean_token_accuracy": 0.8972746133804321, "num_tokens": 853540957.0, "step": 22477 }, { "epoch": 2.8594326421574863, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2134640216827393, "learning_rate": 1e-05, "loss": 0.3915, "mean_token_accuracy": 0.891761302947998, "num_tokens": 853574117.0, "step": 22478 }, { "epoch": 2.859559852436077, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.025701999664307, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8717787265777588, "num_tokens": 853612424.0, "step": 22479 }, { "epoch": 2.8596870627146673, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.1949057579040527, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8753251433372498, "num_tokens": 853652659.0, "step": 22480 }, { "epoch": 2.859814272993258, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.29059362411499, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8844228982925415, "num_tokens": 853690763.0, "step": 22481 }, { "epoch": 2.8599414832718484, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.350446939468384, "learning_rate": 1e-05, "loss": 0.4007, "mean_token_accuracy": 0.8895518183708191, "num_tokens": 853727437.0, "step": 22482 }, { "epoch": 2.860068693550439, "ewc_loss": 0.107421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 7.200241088867188e-05, "grad_norm": 4.487801551818848, "learning_rate": 1e-05, "loss": 0.4879, "mean_token_accuracy": 0.8696364164352417, "num_tokens": 853769179.0, "step": 22483 }, { "epoch": 2.8601959038290294, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.6943392753601074, "learning_rate": 1e-05, "loss": 0.4042, "mean_token_accuracy": 0.891345202922821, "num_tokens": 853803374.0, "step": 22484 }, { "epoch": 2.86032311410762, "ewc_loss": 0.10498046875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.404631614685059, "learning_rate": 1e-05, "loss": 0.4735, "mean_token_accuracy": 0.8705847263336182, "num_tokens": 853834286.0, "step": 22485 }, { "epoch": 2.8604503243862105, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.8915560245513916, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8841719627380371, "num_tokens": 853872369.0, "step": 22486 }, { "epoch": 2.860577534664801, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.150272846221924, "learning_rate": 1e-05, "loss": 0.4891, "mean_token_accuracy": 0.8675881624221802, "num_tokens": 853916470.0, "step": 22487 }, { "epoch": 2.8607047449433916, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.8410677909851074, "learning_rate": 1e-05, "loss": 0.3905, "mean_token_accuracy": 0.8974393606185913, "num_tokens": 853956802.0, "step": 22488 }, { "epoch": 2.860831955221982, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.962531328201294, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8832592964172363, "num_tokens": 853990914.0, "step": 22489 }, { "epoch": 2.8609591655005726, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.739056348800659, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.883926510810852, "num_tokens": 854029707.0, "step": 22490 }, { "epoch": 2.8610863757791627, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8168158531188965, "learning_rate": 1e-05, "loss": 0.3762, "mean_token_accuracy": 0.8997963070869446, "num_tokens": 854065584.0, "step": 22491 }, { "epoch": 2.8612135860577537, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6274147033691406, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8771934509277344, "num_tokens": 854105937.0, "step": 22492 }, { "epoch": 2.8613407963363438, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.735477924346924, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8825438022613525, "num_tokens": 854146746.0, "step": 22493 }, { "epoch": 2.8614680066149347, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.547790765762329, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.8894264698028564, "num_tokens": 854179953.0, "step": 22494 }, { "epoch": 2.861595216893525, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7083427906036377, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8799656629562378, "num_tokens": 854220765.0, "step": 22495 }, { "epoch": 2.861722427172116, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4541690349578857, "learning_rate": 1e-05, "loss": 0.4268, "mean_token_accuracy": 0.8796034455299377, "num_tokens": 854258544.0, "step": 22496 }, { "epoch": 2.861849637450706, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7457337379455566, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.878117024898529, "num_tokens": 854297250.0, "step": 22497 }, { "epoch": 2.8619768477292964, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3595166206359863, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.876099169254303, "num_tokens": 854340283.0, "step": 22498 }, { "epoch": 2.862104058007887, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8387584686279297, "learning_rate": 1e-05, "loss": 0.3969, "mean_token_accuracy": 0.8930383920669556, "num_tokens": 854380145.0, "step": 22499 }, { "epoch": 2.8622312682864774, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.28737735748291, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8796936273574829, "num_tokens": 854422809.0, "step": 22500 }, { "epoch": 2.862358478565068, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 4.008401393890381, "learning_rate": 1e-05, "loss": 0.4209, "mean_token_accuracy": 0.885086715221405, "num_tokens": 854461926.0, "step": 22501 }, { "epoch": 2.8624856888436585, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2981371879577637, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8675381541252136, "num_tokens": 854502431.0, "step": 22502 }, { "epoch": 2.862612899122249, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.187958240509033, "learning_rate": 1e-05, "loss": 0.3961, "mean_token_accuracy": 0.8942316770553589, "num_tokens": 854537440.0, "step": 22503 }, { "epoch": 2.8627401094008396, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.419360637664795, "learning_rate": 1e-05, "loss": 0.4224, "mean_token_accuracy": 0.8854725360870361, "num_tokens": 854575121.0, "step": 22504 }, { "epoch": 2.86286731967943, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.280735015869141, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.8935226798057556, "num_tokens": 854603624.0, "step": 22505 }, { "epoch": 2.8629945299580206, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.6330573558807373, "learning_rate": 1e-05, "loss": 0.4829, "mean_token_accuracy": 0.861873984336853, "num_tokens": 854644616.0, "step": 22506 }, { "epoch": 2.863121740236611, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.153042793273926, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8766329288482666, "num_tokens": 854689772.0, "step": 22507 }, { "epoch": 2.8632489505152017, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.742039918899536, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8891019821166992, "num_tokens": 854725549.0, "step": 22508 }, { "epoch": 2.863376160793792, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9330496788024902, "learning_rate": 1e-05, "loss": 0.4271, "mean_token_accuracy": 0.8847647905349731, "num_tokens": 854768509.0, "step": 22509 }, { "epoch": 2.8635033710723827, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.750990629196167, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8739902973175049, "num_tokens": 854804822.0, "step": 22510 }, { "epoch": 2.8636305813509733, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7282445430755615, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8842779994010925, "num_tokens": 854837458.0, "step": 22511 }, { "epoch": 2.863757791629564, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7045252323150635, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8791902661323547, "num_tokens": 854877996.0, "step": 22512 }, { "epoch": 2.8638850019081543, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.614339828491211, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8712119460105896, "num_tokens": 854913078.0, "step": 22513 }, { "epoch": 2.864012212186745, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.655360221862793, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8855484127998352, "num_tokens": 854952300.0, "step": 22514 }, { "epoch": 2.8641394224653354, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4840826988220215, "learning_rate": 1e-05, "loss": 0.4062, "mean_token_accuracy": 0.8845850229263306, "num_tokens": 854992657.0, "step": 22515 }, { "epoch": 2.8642666327439255, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6575279235839844, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8804133534431458, "num_tokens": 855028147.0, "step": 22516 }, { "epoch": 2.8643938430225164, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4186348915100098, "learning_rate": 1e-05, "loss": 0.4414, "mean_token_accuracy": 0.8773453235626221, "num_tokens": 855069930.0, "step": 22517 }, { "epoch": 2.8645210533011065, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.691612482070923, "learning_rate": 1e-05, "loss": 0.4228, "mean_token_accuracy": 0.8823398351669312, "num_tokens": 855105224.0, "step": 22518 }, { "epoch": 2.8646482635796975, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.371528148651123, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8663475513458252, "num_tokens": 855139533.0, "step": 22519 }, { "epoch": 2.8647754738582876, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.742992877960205, "learning_rate": 1e-05, "loss": 0.4809, "mean_token_accuracy": 0.8672459125518799, "num_tokens": 855178380.0, "step": 22520 }, { "epoch": 2.864902684136878, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.302696943283081, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8765292763710022, "num_tokens": 855212700.0, "step": 22521 }, { "epoch": 2.8650298944154686, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8585429191589355, "learning_rate": 1e-05, "loss": 0.4161, "mean_token_accuracy": 0.88564133644104, "num_tokens": 855250340.0, "step": 22522 }, { "epoch": 2.865157104694059, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.246464729309082, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8802958726882935, "num_tokens": 855285341.0, "step": 22523 }, { "epoch": 2.8652843149726497, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0133748054504395, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8756096363067627, "num_tokens": 855331038.0, "step": 22524 }, { "epoch": 2.86541152525124, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2532730102539062, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8839045763015747, "num_tokens": 855363643.0, "step": 22525 }, { "epoch": 2.8655387355298307, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2240705490112305, "learning_rate": 1e-05, "loss": 0.4273, "mean_token_accuracy": 0.886585533618927, "num_tokens": 855397342.0, "step": 22526 }, { "epoch": 2.8656659458084213, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.3675169944763184, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8899350166320801, "num_tokens": 855432867.0, "step": 22527 }, { "epoch": 2.865793156087012, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.374414443969727, "learning_rate": 1e-05, "loss": 0.467, "mean_token_accuracy": 0.8739039897918701, "num_tokens": 855466091.0, "step": 22528 }, { "epoch": 2.8659203663656023, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6271321773529053, "learning_rate": 1e-05, "loss": 0.4792, "mean_token_accuracy": 0.8666529655456543, "num_tokens": 855501200.0, "step": 22529 }, { "epoch": 2.866047576644193, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.282167434692383, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8887776136398315, "num_tokens": 855543625.0, "step": 22530 }, { "epoch": 2.8661747869227834, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7629482746124268, "learning_rate": 1e-05, "loss": 0.4711, "mean_token_accuracy": 0.8725199103355408, "num_tokens": 855579653.0, "step": 22531 }, { "epoch": 2.866301997201374, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.116729259490967, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8795853853225708, "num_tokens": 855611081.0, "step": 22532 }, { "epoch": 2.8664292074799644, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7729218006134033, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8918125629425049, "num_tokens": 855642389.0, "step": 22533 }, { "epoch": 2.866556417758555, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.910269021987915, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8732972741127014, "num_tokens": 855686240.0, "step": 22534 }, { "epoch": 2.8666836280371455, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7016303539276123, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8736053705215454, "num_tokens": 855727979.0, "step": 22535 }, { "epoch": 2.866810838315736, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7630367279052734, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8678478002548218, "num_tokens": 855769119.0, "step": 22536 }, { "epoch": 2.8669380485943265, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6210744380950928, "learning_rate": 1e-05, "loss": 0.3896, "mean_token_accuracy": 0.890254020690918, "num_tokens": 855811472.0, "step": 22537 }, { "epoch": 2.867065258872917, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6899924278259277, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8805675506591797, "num_tokens": 855842739.0, "step": 22538 }, { "epoch": 2.8671924691515076, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.559471607208252, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8775200247764587, "num_tokens": 855874520.0, "step": 22539 }, { "epoch": 2.867319679430098, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.632876396179199, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8886575698852539, "num_tokens": 855915151.0, "step": 22540 }, { "epoch": 2.867446889708688, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.496060848236084, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8631266355514526, "num_tokens": 855948944.0, "step": 22541 }, { "epoch": 2.867574099987279, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6355745792388916, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8758159279823303, "num_tokens": 855981604.0, "step": 22542 }, { "epoch": 2.8677013102658693, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.419917106628418, "learning_rate": 1e-05, "loss": 0.3971, "mean_token_accuracy": 0.8894041180610657, "num_tokens": 856020407.0, "step": 22543 }, { "epoch": 2.8678285205444602, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.689443349838257, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8671150803565979, "num_tokens": 856056514.0, "step": 22544 }, { "epoch": 2.8679557308230503, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.348543882369995, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8772045969963074, "num_tokens": 856091385.0, "step": 22545 }, { "epoch": 2.868082941101641, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7747342586517334, "learning_rate": 1e-05, "loss": 0.3862, "mean_token_accuracy": 0.8962652683258057, "num_tokens": 856133077.0, "step": 22546 }, { "epoch": 2.8682101513802314, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.273402214050293, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.8840126991271973, "num_tokens": 856174682.0, "step": 22547 }, { "epoch": 2.868337361658822, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9433038234710693, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8625335693359375, "num_tokens": 856214486.0, "step": 22548 }, { "epoch": 2.8684645719374124, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.259246826171875, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8720921277999878, "num_tokens": 856246816.0, "step": 22549 }, { "epoch": 2.868591782216003, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.1424384117126465, "learning_rate": 1e-05, "loss": 0.4751, "mean_token_accuracy": 0.8722715377807617, "num_tokens": 856286232.0, "step": 22550 }, { "epoch": 2.8687189924945935, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.3196041584014893, "learning_rate": 1e-05, "loss": 0.4061, "mean_token_accuracy": 0.8856356739997864, "num_tokens": 856326451.0, "step": 22551 }, { "epoch": 2.868846202773184, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.3026604652404785, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8742387890815735, "num_tokens": 856367766.0, "step": 22552 }, { "epoch": 2.8689734130517746, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5613248348236084, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8743247985839844, "num_tokens": 856396085.0, "step": 22553 }, { "epoch": 2.869100623330365, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.250406265258789, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8923243284225464, "num_tokens": 856439835.0, "step": 22554 }, { "epoch": 2.8692278336089556, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.710143566131592, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8860193490982056, "num_tokens": 856470890.0, "step": 22555 }, { "epoch": 2.869355043887546, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.084230899810791, "learning_rate": 1e-05, "loss": 0.4816, "mean_token_accuracy": 0.8702062964439392, "num_tokens": 856510281.0, "step": 22556 }, { "epoch": 2.8694822541661367, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.7236876487731934, "learning_rate": 1e-05, "loss": 0.4784, "mean_token_accuracy": 0.8660101890563965, "num_tokens": 856558194.0, "step": 22557 }, { "epoch": 2.869609464444727, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.896224021911621, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8804243206977844, "num_tokens": 856594001.0, "step": 22558 }, { "epoch": 2.8697366747233177, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.691784381866455, "learning_rate": 1e-05, "loss": 0.4275, "mean_token_accuracy": 0.8806438446044922, "num_tokens": 856636424.0, "step": 22559 }, { "epoch": 2.8698638850019083, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7403459548950195, "learning_rate": 1e-05, "loss": 0.4634, "mean_token_accuracy": 0.869256854057312, "num_tokens": 856670561.0, "step": 22560 }, { "epoch": 2.869991095280499, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6479666233062744, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8699808716773987, "num_tokens": 856704852.0, "step": 22561 }, { "epoch": 2.8701183055590893, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6262426376342773, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8848122954368591, "num_tokens": 856743943.0, "step": 22562 }, { "epoch": 2.87024551583768, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5939183235168457, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8704594969749451, "num_tokens": 856782944.0, "step": 22563 }, { "epoch": 2.87037272611627, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5695626735687256, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8781386613845825, "num_tokens": 856821240.0, "step": 22564 }, { "epoch": 2.870499936394861, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.542893409729004, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8751348257064819, "num_tokens": 856860113.0, "step": 22565 }, { "epoch": 2.870627146673451, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5243842601776123, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.879250168800354, "num_tokens": 856902501.0, "step": 22566 }, { "epoch": 2.870754356952042, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.513615846633911, "learning_rate": 1e-05, "loss": 0.4085, "mean_token_accuracy": 0.8861191868782043, "num_tokens": 856940023.0, "step": 22567 }, { "epoch": 2.870881567230632, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5142505168914795, "learning_rate": 1e-05, "loss": 0.4985, "mean_token_accuracy": 0.8567445874214172, "num_tokens": 856980653.0, "step": 22568 }, { "epoch": 2.871008777509223, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.503917694091797, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.885585367679596, "num_tokens": 857008081.0, "step": 22569 }, { "epoch": 2.871135987787813, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.497687578201294, "learning_rate": 1e-05, "loss": 0.3908, "mean_token_accuracy": 0.8907980918884277, "num_tokens": 857040884.0, "step": 22570 }, { "epoch": 2.8712631980664036, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.473050355911255, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8748725056648254, "num_tokens": 857078519.0, "step": 22571 }, { "epoch": 2.871390408344994, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.490662097930908, "learning_rate": 1e-05, "loss": 0.434, "mean_token_accuracy": 0.8783711194992065, "num_tokens": 857119250.0, "step": 22572 }, { "epoch": 2.8715176186235847, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4564452171325684, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8808741569519043, "num_tokens": 857160003.0, "step": 22573 }, { "epoch": 2.871644828902175, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4894354343414307, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8803762197494507, "num_tokens": 857200918.0, "step": 22574 }, { "epoch": 2.8717720391807657, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4433135986328125, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8864729404449463, "num_tokens": 857241157.0, "step": 22575 }, { "epoch": 2.8718992494593563, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.505641222000122, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.879313588142395, "num_tokens": 857276237.0, "step": 22576 }, { "epoch": 2.872026459737947, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4218955039978027, "learning_rate": 1e-05, "loss": 0.3853, "mean_token_accuracy": 0.894416332244873, "num_tokens": 857311059.0, "step": 22577 }, { "epoch": 2.8721536700165373, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.548463821411133, "learning_rate": 1e-05, "loss": 0.4802, "mean_token_accuracy": 0.8653128147125244, "num_tokens": 857350326.0, "step": 22578 }, { "epoch": 2.872280880295128, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3857293128967285, "learning_rate": 1e-05, "loss": 0.4527, "mean_token_accuracy": 0.8692465424537659, "num_tokens": 857387708.0, "step": 22579 }, { "epoch": 2.8724080905737184, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5959653854370117, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8831934332847595, "num_tokens": 857427658.0, "step": 22580 }, { "epoch": 2.872535300852309, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.3196494579315186, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8754967451095581, "num_tokens": 857468529.0, "step": 22581 }, { "epoch": 2.8726625111308994, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.682363748550415, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.881585419178009, "num_tokens": 857509406.0, "step": 22582 }, { "epoch": 2.87278972140949, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2411389350891113, "learning_rate": 1e-05, "loss": 0.4318, "mean_token_accuracy": 0.8816583156585693, "num_tokens": 857548302.0, "step": 22583 }, { "epoch": 2.8729169316880805, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.852360486984253, "learning_rate": 1e-05, "loss": 0.3856, "mean_token_accuracy": 0.8955395221710205, "num_tokens": 857584022.0, "step": 22584 }, { "epoch": 2.873044141966671, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.1496517658233643, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8717727661132812, "num_tokens": 857627334.0, "step": 22585 }, { "epoch": 2.8731713522452615, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1649980545043945, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8788132667541504, "num_tokens": 857663317.0, "step": 22586 }, { "epoch": 2.873298562523852, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1681764125823975, "learning_rate": 1e-05, "loss": 0.4693, "mean_token_accuracy": 0.8644111156463623, "num_tokens": 857697358.0, "step": 22587 }, { "epoch": 2.8734257728024426, "ewc_loss": 0.109375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 7.390975952148438e-05, "grad_norm": 4.518864631652832, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8745631575584412, "num_tokens": 857737686.0, "step": 22588 }, { "epoch": 2.8735529830810327, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.4805057048797607, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8697022199630737, "num_tokens": 857774730.0, "step": 22589 }, { "epoch": 2.8736801933596237, "ewc_loss": 0.11181640625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 7.62939453125e-05, "grad_norm": 4.691420078277588, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.874405026435852, "num_tokens": 857811503.0, "step": 22590 }, { "epoch": 2.8738074036382137, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 3.9587342739105225, "learning_rate": 1e-05, "loss": 0.4727, "mean_token_accuracy": 0.8710986375808716, "num_tokens": 857850696.0, "step": 22591 }, { "epoch": 2.8739346139168047, "ewc_loss": 0.10595703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.438527584075928, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8843957781791687, "num_tokens": 857894849.0, "step": 22592 }, { "epoch": 2.874061824195395, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.029872417449951, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8839190006256104, "num_tokens": 857938876.0, "step": 22593 }, { "epoch": 2.8741890344739858, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.197126388549805, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8838781118392944, "num_tokens": 857973424.0, "step": 22594 }, { "epoch": 2.874316244752576, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.921097755432129, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8712781667709351, "num_tokens": 858008105.0, "step": 22595 }, { "epoch": 2.8744434550311664, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.013012409210205, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8823652267456055, "num_tokens": 858043898.0, "step": 22596 }, { "epoch": 2.874570665309757, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7636845111846924, "learning_rate": 1e-05, "loss": 0.4875, "mean_token_accuracy": 0.8650844097137451, "num_tokens": 858080671.0, "step": 22597 }, { "epoch": 2.8746978755883474, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8960118293762207, "learning_rate": 1e-05, "loss": 0.3938, "mean_token_accuracy": 0.8927683234214783, "num_tokens": 858118619.0, "step": 22598 }, { "epoch": 2.874825085866938, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6144907474517822, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8769786357879639, "num_tokens": 858160207.0, "step": 22599 }, { "epoch": 2.8749522961455285, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8801066875457764, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8724309206008911, "num_tokens": 858201634.0, "step": 22600 }, { "epoch": 2.875079506424119, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.4998648166656494, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8702594041824341, "num_tokens": 858241618.0, "step": 22601 }, { "epoch": 2.8752067167027096, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.9264657497406006, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.876836895942688, "num_tokens": 858275393.0, "step": 22602 }, { "epoch": 2.8753339269813, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.408266544342041, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.874385416507721, "num_tokens": 858316470.0, "step": 22603 }, { "epoch": 2.8754611372598906, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.059117794036865, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8850154876708984, "num_tokens": 858350567.0, "step": 22604 }, { "epoch": 2.875588347538481, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3783841133117676, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8778982162475586, "num_tokens": 858393998.0, "step": 22605 }, { "epoch": 2.8757155578170717, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.226787567138672, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8910640478134155, "num_tokens": 858430987.0, "step": 22606 }, { "epoch": 2.875842768095662, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.5231738090515137, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8735324144363403, "num_tokens": 858464230.0, "step": 22607 }, { "epoch": 2.8759699783742527, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.259377479553223, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8732563257217407, "num_tokens": 858505965.0, "step": 22608 }, { "epoch": 2.8760971886528433, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.6909170150756836, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8817386627197266, "num_tokens": 858542629.0, "step": 22609 }, { "epoch": 2.876224398931434, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.10085391998291, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8895784020423889, "num_tokens": 858583377.0, "step": 22610 }, { "epoch": 2.8763516092100243, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.766054391860962, "learning_rate": 1e-05, "loss": 0.4036, "mean_token_accuracy": 0.8925752639770508, "num_tokens": 858620546.0, "step": 22611 }, { "epoch": 2.876478819488615, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.886575698852539, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8718217611312866, "num_tokens": 858659469.0, "step": 22612 }, { "epoch": 2.8766060297672054, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7536211013793945, "learning_rate": 1e-05, "loss": 0.4355, "mean_token_accuracy": 0.8786709308624268, "num_tokens": 858694519.0, "step": 22613 }, { "epoch": 2.8767332400457954, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6657216548919678, "learning_rate": 1e-05, "loss": 0.4092, "mean_token_accuracy": 0.8874379396438599, "num_tokens": 858728371.0, "step": 22614 }, { "epoch": 2.8768604503243864, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7226760387420654, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8851920366287231, "num_tokens": 858765195.0, "step": 22615 }, { "epoch": 2.8769876606029765, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.5047433376312256, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8883402347564697, "num_tokens": 858804129.0, "step": 22616 }, { "epoch": 2.8771148708815675, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.711869716644287, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8867855668067932, "num_tokens": 858841267.0, "step": 22617 }, { "epoch": 2.8772420811601576, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.4034106731414795, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8774899840354919, "num_tokens": 858882405.0, "step": 22618 }, { "epoch": 2.877369291438748, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.766522169113159, "learning_rate": 1e-05, "loss": 0.4235, "mean_token_accuracy": 0.8819013833999634, "num_tokens": 858920176.0, "step": 22619 }, { "epoch": 2.8774965017173386, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.320671558380127, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8736405372619629, "num_tokens": 858959065.0, "step": 22620 }, { "epoch": 2.877623711995929, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8685412406921387, "learning_rate": 1e-05, "loss": 0.4103, "mean_token_accuracy": 0.8883940577507019, "num_tokens": 858997497.0, "step": 22621 }, { "epoch": 2.8777509222745197, "ewc_loss": 0.08203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.2493340969085693, "learning_rate": 1e-05, "loss": 0.4051, "mean_token_accuracy": 0.8849064111709595, "num_tokens": 859036425.0, "step": 22622 }, { "epoch": 2.87787813255311, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.057413578033447, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8777143359184265, "num_tokens": 859073602.0, "step": 22623 }, { "epoch": 2.8780053428317007, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.290198802947998, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8758313655853271, "num_tokens": 859107576.0, "step": 22624 }, { "epoch": 2.8781325531102913, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.243764400482178, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.870909571647644, "num_tokens": 859141866.0, "step": 22625 }, { "epoch": 2.878259763388882, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.420416831970215, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8879945278167725, "num_tokens": 859180592.0, "step": 22626 }, { "epoch": 2.8783869736674723, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.338947772979736, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.8740498423576355, "num_tokens": 859216365.0, "step": 22627 }, { "epoch": 2.878514183946063, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.6574926376342773, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.8759982585906982, "num_tokens": 859254167.0, "step": 22628 }, { "epoch": 2.8786413942246534, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.226166248321533, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8861566185951233, "num_tokens": 859288980.0, "step": 22629 }, { "epoch": 2.878768604503244, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7596545219421387, "learning_rate": 1e-05, "loss": 0.3987, "mean_token_accuracy": 0.8949301242828369, "num_tokens": 859326741.0, "step": 22630 }, { "epoch": 2.8788958147818344, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.024375915527344, "learning_rate": 1e-05, "loss": 0.4524, "mean_token_accuracy": 0.8736838102340698, "num_tokens": 859369530.0, "step": 22631 }, { "epoch": 2.879023025060425, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.741387367248535, "learning_rate": 1e-05, "loss": 0.4895, "mean_token_accuracy": 0.8624331951141357, "num_tokens": 859409870.0, "step": 22632 }, { "epoch": 2.8791502353390155, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8751392364501953, "learning_rate": 1e-05, "loss": 0.4806, "mean_token_accuracy": 0.8635509014129639, "num_tokens": 859443419.0, "step": 22633 }, { "epoch": 2.879277445617606, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.687767744064331, "learning_rate": 1e-05, "loss": 0.4175, "mean_token_accuracy": 0.885058581829071, "num_tokens": 859474832.0, "step": 22634 }, { "epoch": 2.8794046558961965, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7201244831085205, "learning_rate": 1e-05, "loss": 0.4553, "mean_token_accuracy": 0.8761967420578003, "num_tokens": 859517668.0, "step": 22635 }, { "epoch": 2.879531866174787, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6050193309783936, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8826392889022827, "num_tokens": 859555954.0, "step": 22636 }, { "epoch": 2.8796590764533776, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.655949592590332, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8729740381240845, "num_tokens": 859591221.0, "step": 22637 }, { "epoch": 2.879786286731968, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5309290885925293, "learning_rate": 1e-05, "loss": 0.3812, "mean_token_accuracy": 0.8943195343017578, "num_tokens": 859631879.0, "step": 22638 }, { "epoch": 2.879913497010558, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.639484167098999, "learning_rate": 1e-05, "loss": 0.4706, "mean_token_accuracy": 0.8724923729896545, "num_tokens": 859669259.0, "step": 22639 }, { "epoch": 2.880040707289149, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.45147705078125, "learning_rate": 1e-05, "loss": 0.4041, "mean_token_accuracy": 0.885527491569519, "num_tokens": 859706168.0, "step": 22640 }, { "epoch": 2.8801679175677393, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6633968353271484, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8793036341667175, "num_tokens": 859743182.0, "step": 22641 }, { "epoch": 2.8802951278463302, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3815643787384033, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8826694488525391, "num_tokens": 859786120.0, "step": 22642 }, { "epoch": 2.8804223381249203, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7386834621429443, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.885377049446106, "num_tokens": 859824097.0, "step": 22643 }, { "epoch": 2.880549548403511, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3279781341552734, "learning_rate": 1e-05, "loss": 0.4589, "mean_token_accuracy": 0.8680070638656616, "num_tokens": 859860081.0, "step": 22644 }, { "epoch": 2.8806767586821014, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8547308444976807, "learning_rate": 1e-05, "loss": 0.456, "mean_token_accuracy": 0.8730775117874146, "num_tokens": 859898731.0, "step": 22645 }, { "epoch": 2.880803968960692, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.266089916229248, "learning_rate": 1e-05, "loss": 0.4267, "mean_token_accuracy": 0.87906813621521, "num_tokens": 859938406.0, "step": 22646 }, { "epoch": 2.8809311792392824, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.032978534698486, "learning_rate": 1e-05, "loss": 0.4011, "mean_token_accuracy": 0.8922906517982483, "num_tokens": 859974270.0, "step": 22647 }, { "epoch": 2.881058389517873, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.2908377647399902, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.883608341217041, "num_tokens": 860007261.0, "step": 22648 }, { "epoch": 2.8811855997964635, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.208367347717285, "learning_rate": 1e-05, "loss": 0.4458, "mean_token_accuracy": 0.8788785338401794, "num_tokens": 860051170.0, "step": 22649 }, { "epoch": 2.881312810075054, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.430788278579712, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8840264678001404, "num_tokens": 860082771.0, "step": 22650 }, { "epoch": 2.8814400203536445, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.274565696716309, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8666850328445435, "num_tokens": 860124184.0, "step": 22651 }, { "epoch": 2.881567230632235, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.6395246982574463, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.866332471370697, "num_tokens": 860158453.0, "step": 22652 }, { "epoch": 2.8816944409108256, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.151477336883545, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8869712948799133, "num_tokens": 860195403.0, "step": 22653 }, { "epoch": 2.881821651189416, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.711432933807373, "learning_rate": 1e-05, "loss": 0.4608, "mean_token_accuracy": 0.8695504665374756, "num_tokens": 860232813.0, "step": 22654 }, { "epoch": 2.8819488614680067, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.9699244499206543, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8755972385406494, "num_tokens": 860269542.0, "step": 22655 }, { "epoch": 2.882076071746597, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.719728469848633, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8729021549224854, "num_tokens": 860305386.0, "step": 22656 }, { "epoch": 2.8822032820251877, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7781362533569336, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.879142701625824, "num_tokens": 860343137.0, "step": 22657 }, { "epoch": 2.8823304923037782, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.670696496963501, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8726997375488281, "num_tokens": 860382968.0, "step": 22658 }, { "epoch": 2.8824577025823688, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6300127506256104, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8735494613647461, "num_tokens": 860426985.0, "step": 22659 }, { "epoch": 2.8825849128609593, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6380269527435303, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8694038391113281, "num_tokens": 860460950.0, "step": 22660 }, { "epoch": 2.88271212313955, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5359952449798584, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8852009773254395, "num_tokens": 860499328.0, "step": 22661 }, { "epoch": 2.88283933341814, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6112256050109863, "learning_rate": 1e-05, "loss": 0.4411, "mean_token_accuracy": 0.8792108297348022, "num_tokens": 860532411.0, "step": 22662 }, { "epoch": 2.882966543696731, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4739511013031006, "learning_rate": 1e-05, "loss": 0.4099, "mean_token_accuracy": 0.8892530202865601, "num_tokens": 860567565.0, "step": 22663 }, { "epoch": 2.883093753975321, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.588390588760376, "learning_rate": 1e-05, "loss": 0.4566, "mean_token_accuracy": 0.871421217918396, "num_tokens": 860611687.0, "step": 22664 }, { "epoch": 2.883220964253912, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.435387372970581, "learning_rate": 1e-05, "loss": 0.4839, "mean_token_accuracy": 0.8637340068817139, "num_tokens": 860652220.0, "step": 22665 }, { "epoch": 2.883348174532502, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5902068614959717, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8858356475830078, "num_tokens": 860689641.0, "step": 22666 }, { "epoch": 2.883475384811093, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.395958185195923, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8833245038986206, "num_tokens": 860727490.0, "step": 22667 }, { "epoch": 2.883602595089683, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.631342887878418, "learning_rate": 1e-05, "loss": 0.4365, "mean_token_accuracy": 0.8744446635246277, "num_tokens": 860768625.0, "step": 22668 }, { "epoch": 2.8837298053682736, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3432228565216064, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8852659463882446, "num_tokens": 860806294.0, "step": 22669 }, { "epoch": 2.883857015646864, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.71710205078125, "learning_rate": 1e-05, "loss": 0.4831, "mean_token_accuracy": 0.8627908229827881, "num_tokens": 860842672.0, "step": 22670 }, { "epoch": 2.8839842259254547, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2794535160064697, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.8702821135520935, "num_tokens": 860879882.0, "step": 22671 }, { "epoch": 2.884111436204045, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.844088315963745, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8793982267379761, "num_tokens": 860916099.0, "step": 22672 }, { "epoch": 2.8842386464826357, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.193661689758301, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8790118098258972, "num_tokens": 860957982.0, "step": 22673 }, { "epoch": 2.8843658567612263, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.07423210144043, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8870738744735718, "num_tokens": 860995528.0, "step": 22674 }, { "epoch": 2.884493067039817, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.1990230083465576, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.876238226890564, "num_tokens": 861034676.0, "step": 22675 }, { "epoch": 2.8846202773184073, "ewc_loss": 0.10498046875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.380457878112793, "learning_rate": 1e-05, "loss": 0.4777, "mean_token_accuracy": 0.8707880973815918, "num_tokens": 861069767.0, "step": 22676 }, { "epoch": 2.884747487596998, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4187629222869873, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8806126713752747, "num_tokens": 861109787.0, "step": 22677 }, { "epoch": 2.8848746978755884, "ewc_loss": 0.1083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.532398223876953, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8824611306190491, "num_tokens": 861144603.0, "step": 22678 }, { "epoch": 2.885001908154179, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.7793939113616943, "learning_rate": 1e-05, "loss": 0.4264, "mean_token_accuracy": 0.8859703540802002, "num_tokens": 861184750.0, "step": 22679 }, { "epoch": 2.8851291184327694, "ewc_loss": 0.10498046875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.374353408813477, "learning_rate": 1e-05, "loss": 0.4554, "mean_token_accuracy": 0.876535177230835, "num_tokens": 861223591.0, "step": 22680 }, { "epoch": 2.88525632871136, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.9364960193634033, "learning_rate": 1e-05, "loss": 0.453, "mean_token_accuracy": 0.8753489255905151, "num_tokens": 861260456.0, "step": 22681 }, { "epoch": 2.8853835389899505, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.127100467681885, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8768698573112488, "num_tokens": 861298266.0, "step": 22682 }, { "epoch": 2.885510749268541, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.8467533588409424, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.882992148399353, "num_tokens": 861334646.0, "step": 22683 }, { "epoch": 2.8856379595471315, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.942899465560913, "learning_rate": 1e-05, "loss": 0.4636, "mean_token_accuracy": 0.872219443321228, "num_tokens": 861370734.0, "step": 22684 }, { "epoch": 2.885765169825722, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.727480173110962, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8949825167655945, "num_tokens": 861403111.0, "step": 22685 }, { "epoch": 2.8858923801043126, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8200666904449463, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8815493583679199, "num_tokens": 861440668.0, "step": 22686 }, { "epoch": 2.8860195903829027, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6174163818359375, "learning_rate": 1e-05, "loss": 0.4714, "mean_token_accuracy": 0.8703183531761169, "num_tokens": 861478586.0, "step": 22687 }, { "epoch": 2.8861468006614936, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.763371229171753, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8785393238067627, "num_tokens": 861511637.0, "step": 22688 }, { "epoch": 2.8862740109400837, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5189061164855957, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8778648376464844, "num_tokens": 861545421.0, "step": 22689 }, { "epoch": 2.8864012212186747, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7707951068878174, "learning_rate": 1e-05, "loss": 0.4684, "mean_token_accuracy": 0.8732650279998779, "num_tokens": 861586230.0, "step": 22690 }, { "epoch": 2.886528431497265, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4138669967651367, "learning_rate": 1e-05, "loss": 0.4098, "mean_token_accuracy": 0.8861193656921387, "num_tokens": 861620556.0, "step": 22691 }, { "epoch": 2.8866556417758558, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8429455757141113, "learning_rate": 1e-05, "loss": 0.4396, "mean_token_accuracy": 0.8759444355964661, "num_tokens": 861655826.0, "step": 22692 }, { "epoch": 2.886782852054446, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.350372552871704, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8820899724960327, "num_tokens": 861699160.0, "step": 22693 }, { "epoch": 2.8869100623330364, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.9666507244110107, "learning_rate": 1e-05, "loss": 0.4833, "mean_token_accuracy": 0.8709602355957031, "num_tokens": 861736784.0, "step": 22694 }, { "epoch": 2.887037272611627, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.3030643463134766, "learning_rate": 1e-05, "loss": 0.3986, "mean_token_accuracy": 0.889447808265686, "num_tokens": 861775359.0, "step": 22695 }, { "epoch": 2.8871644828902174, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.14861536026001, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8673827648162842, "num_tokens": 861813852.0, "step": 22696 }, { "epoch": 2.887291693168808, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.4021592140197754, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8781518936157227, "num_tokens": 861853730.0, "step": 22697 }, { "epoch": 2.8874189034473985, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.260270595550537, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8758800625801086, "num_tokens": 861891405.0, "step": 22698 }, { "epoch": 2.887546113725989, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.5967905521392822, "learning_rate": 1e-05, "loss": 0.4069, "mean_token_accuracy": 0.8886591196060181, "num_tokens": 861929595.0, "step": 22699 }, { "epoch": 2.8876733240045795, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.1937103271484375, "learning_rate": 1e-05, "loss": 0.4775, "mean_token_accuracy": 0.8670445680618286, "num_tokens": 861963263.0, "step": 22700 }, { "epoch": 2.88780053428317, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.733173131942749, "learning_rate": 1e-05, "loss": 0.4769, "mean_token_accuracy": 0.8705253601074219, "num_tokens": 862002259.0, "step": 22701 }, { "epoch": 2.8879277445617606, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.988560438156128, "learning_rate": 1e-05, "loss": 0.5028, "mean_token_accuracy": 0.8623040318489075, "num_tokens": 862039869.0, "step": 22702 }, { "epoch": 2.888054954840351, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.763617992401123, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8793869018554688, "num_tokens": 862068979.0, "step": 22703 }, { "epoch": 2.8881821651189417, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.743673801422119, "learning_rate": 1e-05, "loss": 0.4798, "mean_token_accuracy": 0.8651275634765625, "num_tokens": 862117384.0, "step": 22704 }, { "epoch": 2.888309375397532, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7292120456695557, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8685204982757568, "num_tokens": 862157961.0, "step": 22705 }, { "epoch": 2.8884365856761227, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5587692260742188, "learning_rate": 1e-05, "loss": 0.4353, "mean_token_accuracy": 0.8802609443664551, "num_tokens": 862199023.0, "step": 22706 }, { "epoch": 2.8885637959547132, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7245614528656006, "learning_rate": 1e-05, "loss": 0.4346, "mean_token_accuracy": 0.8792573809623718, "num_tokens": 862237839.0, "step": 22707 }, { "epoch": 2.8886910062333038, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.436215877532959, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8748177289962769, "num_tokens": 862273060.0, "step": 22708 }, { "epoch": 2.8888182165118943, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7456462383270264, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.8684784770011902, "num_tokens": 862316469.0, "step": 22709 }, { "epoch": 2.888945426790485, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3515164852142334, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8711721897125244, "num_tokens": 862359367.0, "step": 22710 }, { "epoch": 2.8890726370690754, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8138458728790283, "learning_rate": 1e-05, "loss": 0.418, "mean_token_accuracy": 0.8856254816055298, "num_tokens": 862397113.0, "step": 22711 }, { "epoch": 2.8891998473476654, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.294961929321289, "learning_rate": 1e-05, "loss": 0.4517, "mean_token_accuracy": 0.8727755546569824, "num_tokens": 862435627.0, "step": 22712 }, { "epoch": 2.8893270576262564, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.939541816711426, "learning_rate": 1e-05, "loss": 0.4371, "mean_token_accuracy": 0.8800491094589233, "num_tokens": 862476948.0, "step": 22713 }, { "epoch": 2.8894542679048465, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2701802253723145, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8769096732139587, "num_tokens": 862513429.0, "step": 22714 }, { "epoch": 2.8895814781834375, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.12029504776001, "learning_rate": 1e-05, "loss": 0.4728, "mean_token_accuracy": 0.8671160936355591, "num_tokens": 862548570.0, "step": 22715 }, { "epoch": 2.8897086884620276, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.400935411453247, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8769051432609558, "num_tokens": 862586265.0, "step": 22716 }, { "epoch": 2.889835898740618, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.196383953094482, "learning_rate": 1e-05, "loss": 0.4899, "mean_token_accuracy": 0.8644099235534668, "num_tokens": 862620552.0, "step": 22717 }, { "epoch": 2.8899631090192086, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.467470407485962, "learning_rate": 1e-05, "loss": 0.4013, "mean_token_accuracy": 0.8901417255401611, "num_tokens": 862656449.0, "step": 22718 }, { "epoch": 2.890090319297799, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.208944797515869, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8708301186561584, "num_tokens": 862697612.0, "step": 22719 }, { "epoch": 2.8902175295763897, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.606335401535034, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.8670074939727783, "num_tokens": 862742304.0, "step": 22720 }, { "epoch": 2.89034473985498, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.117232322692871, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.8804294466972351, "num_tokens": 862781738.0, "step": 22721 }, { "epoch": 2.8904719501335707, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.682666778564453, "learning_rate": 1e-05, "loss": 0.5147, "mean_token_accuracy": 0.8508557081222534, "num_tokens": 862823124.0, "step": 22722 }, { "epoch": 2.8905991604121613, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9682204723358154, "learning_rate": 1e-05, "loss": 0.4835, "mean_token_accuracy": 0.8667654991149902, "num_tokens": 862857706.0, "step": 22723 }, { "epoch": 2.890726370690752, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.665057897567749, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8597261309623718, "num_tokens": 862899966.0, "step": 22724 }, { "epoch": 2.8908535809693423, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8293533325195312, "learning_rate": 1e-05, "loss": 0.4135, "mean_token_accuracy": 0.8788290023803711, "num_tokens": 862931090.0, "step": 22725 }, { "epoch": 2.890980791247933, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6183836460113525, "learning_rate": 1e-05, "loss": 0.4005, "mean_token_accuracy": 0.8845617771148682, "num_tokens": 862962833.0, "step": 22726 }, { "epoch": 2.8911080015265234, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.711325168609619, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8734225034713745, "num_tokens": 863003769.0, "step": 22727 }, { "epoch": 2.891235211805114, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.562176465988159, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8783761262893677, "num_tokens": 863038130.0, "step": 22728 }, { "epoch": 2.8913624220837044, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.645374298095703, "learning_rate": 1e-05, "loss": 0.396, "mean_token_accuracy": 0.891990065574646, "num_tokens": 863077943.0, "step": 22729 }, { "epoch": 2.891489632362295, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5111539363861084, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8753457069396973, "num_tokens": 863113853.0, "step": 22730 }, { "epoch": 2.8916168426408855, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6259093284606934, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8842511177062988, "num_tokens": 863149407.0, "step": 22731 }, { "epoch": 2.891744052919476, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.459481716156006, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8726347088813782, "num_tokens": 863183324.0, "step": 22732 }, { "epoch": 2.8918712631980665, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6362287998199463, "learning_rate": 1e-05, "loss": 0.3904, "mean_token_accuracy": 0.892667829990387, "num_tokens": 863220400.0, "step": 22733 }, { "epoch": 2.891998473476657, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3896284103393555, "learning_rate": 1e-05, "loss": 0.41, "mean_token_accuracy": 0.8861566781997681, "num_tokens": 863259565.0, "step": 22734 }, { "epoch": 2.8921256837552476, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.703056573867798, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8713291883468628, "num_tokens": 863293480.0, "step": 22735 }, { "epoch": 2.892252894033838, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3346917629241943, "learning_rate": 1e-05, "loss": 0.4449, "mean_token_accuracy": 0.8774213790893555, "num_tokens": 863331702.0, "step": 22736 }, { "epoch": 2.892380104312428, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7947933673858643, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.873167872428894, "num_tokens": 863366195.0, "step": 22737 }, { "epoch": 2.892507314591019, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.274685859680176, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.8723247647285461, "num_tokens": 863406674.0, "step": 22738 }, { "epoch": 2.8926345248696093, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.9489853382110596, "learning_rate": 1e-05, "loss": 0.4525, "mean_token_accuracy": 0.8771740198135376, "num_tokens": 863441837.0, "step": 22739 }, { "epoch": 2.8927617351482002, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.258669853210449, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8851435780525208, "num_tokens": 863476461.0, "step": 22740 }, { "epoch": 2.8928889454267903, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.127987384796143, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8697333335876465, "num_tokens": 863513715.0, "step": 22741 }, { "epoch": 2.893016155705381, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3285255432128906, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8825913667678833, "num_tokens": 863553172.0, "step": 22742 }, { "epoch": 2.8931433659839714, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.300748348236084, "learning_rate": 1e-05, "loss": 0.489, "mean_token_accuracy": 0.864503026008606, "num_tokens": 863585353.0, "step": 22743 }, { "epoch": 2.893270576262562, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.535893678665161, "learning_rate": 1e-05, "loss": 0.4717, "mean_token_accuracy": 0.8714859485626221, "num_tokens": 863627510.0, "step": 22744 }, { "epoch": 2.8933977865411524, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.260318756103516, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8702198266983032, "num_tokens": 863670683.0, "step": 22745 }, { "epoch": 2.893524996819743, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.6978211402893066, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8726801872253418, "num_tokens": 863710793.0, "step": 22746 }, { "epoch": 2.8936522070983335, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.095513820648193, "learning_rate": 1e-05, "loss": 0.4251, "mean_token_accuracy": 0.8842318654060364, "num_tokens": 863745963.0, "step": 22747 }, { "epoch": 2.893779417376924, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.7352542877197266, "learning_rate": 1e-05, "loss": 0.4075, "mean_token_accuracy": 0.8904116153717041, "num_tokens": 863783729.0, "step": 22748 }, { "epoch": 2.8939066276555145, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.8912689685821533, "learning_rate": 1e-05, "loss": 0.4332, "mean_token_accuracy": 0.8796985745429993, "num_tokens": 863820070.0, "step": 22749 }, { "epoch": 2.894033837934105, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7042036056518555, "learning_rate": 1e-05, "loss": 0.4185, "mean_token_accuracy": 0.882347583770752, "num_tokens": 863853438.0, "step": 22750 }, { "epoch": 2.8941610482126956, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7177274227142334, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.880378782749176, "num_tokens": 863894151.0, "step": 22751 }, { "epoch": 2.894288258491286, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.657926082611084, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8845146894454956, "num_tokens": 863929712.0, "step": 22752 }, { "epoch": 2.8944154687698767, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5970144271850586, "learning_rate": 1e-05, "loss": 0.491, "mean_token_accuracy": 0.8600287437438965, "num_tokens": 863970236.0, "step": 22753 }, { "epoch": 2.894542679048467, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6123032569885254, "learning_rate": 1e-05, "loss": 0.4648, "mean_token_accuracy": 0.8683121800422668, "num_tokens": 864009164.0, "step": 22754 }, { "epoch": 2.8946698893270577, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5180654525756836, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8740275502204895, "num_tokens": 864048941.0, "step": 22755 }, { "epoch": 2.8947970996056482, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5988688468933105, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8737839460372925, "num_tokens": 864084041.0, "step": 22756 }, { "epoch": 2.8949243098842388, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4698984622955322, "learning_rate": 1e-05, "loss": 0.4118, "mean_token_accuracy": 0.8815914392471313, "num_tokens": 864118370.0, "step": 22757 }, { "epoch": 2.8950515201628293, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5991687774658203, "learning_rate": 1e-05, "loss": 0.394, "mean_token_accuracy": 0.8907419443130493, "num_tokens": 864148888.0, "step": 22758 }, { "epoch": 2.89517873044142, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4220974445343018, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8828849196434021, "num_tokens": 864185850.0, "step": 22759 }, { "epoch": 2.89530594072001, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.631871223449707, "learning_rate": 1e-05, "loss": 0.484, "mean_token_accuracy": 0.8611833453178406, "num_tokens": 864224914.0, "step": 22760 }, { "epoch": 2.895433150998601, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.357635498046875, "learning_rate": 1e-05, "loss": 0.4369, "mean_token_accuracy": 0.8741887807846069, "num_tokens": 864266430.0, "step": 22761 }, { "epoch": 2.895560361277191, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6943604946136475, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8747930526733398, "num_tokens": 864299769.0, "step": 22762 }, { "epoch": 2.895687571555782, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.291834831237793, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8717555999755859, "num_tokens": 864340713.0, "step": 22763 }, { "epoch": 2.895814781834372, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.8178958892822266, "learning_rate": 1e-05, "loss": 0.4733, "mean_token_accuracy": 0.8686199188232422, "num_tokens": 864382930.0, "step": 22764 }, { "epoch": 2.895941992112963, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2079415321350098, "learning_rate": 1e-05, "loss": 0.4026, "mean_token_accuracy": 0.8871666789054871, "num_tokens": 864424632.0, "step": 22765 }, { "epoch": 2.896069202391553, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.031270503997803, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8895448446273804, "num_tokens": 864463444.0, "step": 22766 }, { "epoch": 2.8961964126701436, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.1997461318969727, "learning_rate": 1e-05, "loss": 0.4932, "mean_token_accuracy": 0.8588879108428955, "num_tokens": 864506651.0, "step": 22767 }, { "epoch": 2.896323622948734, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.305253505706787, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8828830122947693, "num_tokens": 864544075.0, "step": 22768 }, { "epoch": 2.8964508332273247, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.353839635848999, "learning_rate": 1e-05, "loss": 0.3903, "mean_token_accuracy": 0.8951246738433838, "num_tokens": 864579814.0, "step": 22769 }, { "epoch": 2.896578043505915, "ewc_loss": 0.10791015625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 7.2479248046875e-05, "grad_norm": 4.509615898132324, "learning_rate": 1e-05, "loss": 0.5148, "mean_token_accuracy": 0.8598827719688416, "num_tokens": 864616690.0, "step": 22770 }, { "epoch": 2.8967052537845057, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.6906940937042236, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8827031850814819, "num_tokens": 864662035.0, "step": 22771 }, { "epoch": 2.8968324640630962, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.4326558113098145, "learning_rate": 1e-05, "loss": 0.4892, "mean_token_accuracy": 0.8669600486755371, "num_tokens": 864695219.0, "step": 22772 }, { "epoch": 2.8969596743416868, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.907850980758667, "learning_rate": 1e-05, "loss": 0.4486, "mean_token_accuracy": 0.8800257444381714, "num_tokens": 864731990.0, "step": 22773 }, { "epoch": 2.8970868846202773, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.178158283233643, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8804827928543091, "num_tokens": 864769642.0, "step": 22774 }, { "epoch": 2.897214094898868, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.873353958129883, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8725588917732239, "num_tokens": 864805329.0, "step": 22775 }, { "epoch": 2.8973413051774584, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.9600636959075928, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8877461552619934, "num_tokens": 864847748.0, "step": 22776 }, { "epoch": 2.897468515456049, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.758167028427124, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.865720272064209, "num_tokens": 864884621.0, "step": 22777 }, { "epoch": 2.8975957257346394, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8050856590270996, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.886001467704773, "num_tokens": 864927398.0, "step": 22778 }, { "epoch": 2.89772293601323, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6642305850982666, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8889055252075195, "num_tokens": 864961038.0, "step": 22779 }, { "epoch": 2.8978501462918205, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.721151351928711, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.8734490275382996, "num_tokens": 865000307.0, "step": 22780 }, { "epoch": 2.897977356570411, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5573208332061768, "learning_rate": 1e-05, "loss": 0.3991, "mean_token_accuracy": 0.8883508443832397, "num_tokens": 865035898.0, "step": 22781 }, { "epoch": 2.8981045668490015, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6879425048828125, "learning_rate": 1e-05, "loss": 0.3798, "mean_token_accuracy": 0.8979580402374268, "num_tokens": 865076325.0, "step": 22782 }, { "epoch": 2.898231777127592, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4665181636810303, "learning_rate": 1e-05, "loss": 0.4497, "mean_token_accuracy": 0.8729219436645508, "num_tokens": 865117290.0, "step": 22783 }, { "epoch": 2.8983589874061826, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7185568809509277, "learning_rate": 1e-05, "loss": 0.4212, "mean_token_accuracy": 0.8837615847587585, "num_tokens": 865153003.0, "step": 22784 }, { "epoch": 2.8984861976847727, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.368711233139038, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.8729127049446106, "num_tokens": 865195113.0, "step": 22785 }, { "epoch": 2.8986134079633636, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.8081212043762207, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8849908113479614, "num_tokens": 865231319.0, "step": 22786 }, { "epoch": 2.8987406182419537, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.3028340339660645, "learning_rate": 1e-05, "loss": 0.4189, "mean_token_accuracy": 0.8793675899505615, "num_tokens": 865265862.0, "step": 22787 }, { "epoch": 2.8988678285205447, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.99440598487854, "learning_rate": 1e-05, "loss": 0.4571, "mean_token_accuracy": 0.8759362697601318, "num_tokens": 865303873.0, "step": 22788 }, { "epoch": 2.898995038799135, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2819719314575195, "learning_rate": 1e-05, "loss": 0.4047, "mean_token_accuracy": 0.8867161273956299, "num_tokens": 865335163.0, "step": 22789 }, { "epoch": 2.8991222490777258, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.176340103149414, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8772534132003784, "num_tokens": 865373498.0, "step": 22790 }, { "epoch": 2.899249459356316, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.37306809425354, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8785660266876221, "num_tokens": 865411042.0, "step": 22791 }, { "epoch": 2.8993766696349064, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.312139511108398, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.879488468170166, "num_tokens": 865447676.0, "step": 22792 }, { "epoch": 2.899503879913497, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.5958733558654785, "learning_rate": 1e-05, "loss": 0.4545, "mean_token_accuracy": 0.8737797737121582, "num_tokens": 865490905.0, "step": 22793 }, { "epoch": 2.8996310901920874, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.255251884460449, "learning_rate": 1e-05, "loss": 0.4582, "mean_token_accuracy": 0.8813818097114563, "num_tokens": 865530572.0, "step": 22794 }, { "epoch": 2.899758300470678, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7616889476776123, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8822080492973328, "num_tokens": 865565999.0, "step": 22795 }, { "epoch": 2.8998855107492685, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.055091381072998, "learning_rate": 1e-05, "loss": 0.4583, "mean_token_accuracy": 0.871845006942749, "num_tokens": 865602455.0, "step": 22796 }, { "epoch": 2.900012721027859, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7779276371002197, "learning_rate": 1e-05, "loss": 0.3831, "mean_token_accuracy": 0.8993188142776489, "num_tokens": 865637723.0, "step": 22797 }, { "epoch": 2.9001399313064495, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.843824863433838, "learning_rate": 1e-05, "loss": 0.4591, "mean_token_accuracy": 0.8704887628555298, "num_tokens": 865669668.0, "step": 22798 }, { "epoch": 2.90026714158504, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7325689792633057, "learning_rate": 1e-05, "loss": 0.4386, "mean_token_accuracy": 0.8790254592895508, "num_tokens": 865708255.0, "step": 22799 }, { "epoch": 2.9003943518636306, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.657658815383911, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8684283494949341, "num_tokens": 865749607.0, "step": 22800 }, { "epoch": 2.900521562142221, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.726457357406616, "learning_rate": 1e-05, "loss": 0.47, "mean_token_accuracy": 0.8706228733062744, "num_tokens": 865786372.0, "step": 22801 }, { "epoch": 2.9006487724208116, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5194363594055176, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8823691606521606, "num_tokens": 865822939.0, "step": 22802 }, { "epoch": 2.900775982699402, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.681224822998047, "learning_rate": 1e-05, "loss": 0.3988, "mean_token_accuracy": 0.8888617157936096, "num_tokens": 865859012.0, "step": 22803 }, { "epoch": 2.9009031929779927, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.433361768722534, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8850045204162598, "num_tokens": 865897641.0, "step": 22804 }, { "epoch": 2.9010304032565832, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6820685863494873, "learning_rate": 1e-05, "loss": 0.4445, "mean_token_accuracy": 0.8764573335647583, "num_tokens": 865938270.0, "step": 22805 }, { "epoch": 2.9011576135351738, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3756589889526367, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8821040987968445, "num_tokens": 865971066.0, "step": 22806 }, { "epoch": 2.9012848238137643, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.728950023651123, "learning_rate": 1e-05, "loss": 0.4107, "mean_token_accuracy": 0.8875246047973633, "num_tokens": 866008595.0, "step": 22807 }, { "epoch": 2.901412034092355, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3251869678497314, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8811072707176208, "num_tokens": 866046650.0, "step": 22808 }, { "epoch": 2.9015392443709453, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8293046951293945, "learning_rate": 1e-05, "loss": 0.4652, "mean_token_accuracy": 0.8696976900100708, "num_tokens": 866086305.0, "step": 22809 }, { "epoch": 2.9016664546495354, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2543416023254395, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8767486810684204, "num_tokens": 866125770.0, "step": 22810 }, { "epoch": 2.9017936649281264, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.999727725982666, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8876266479492188, "num_tokens": 866160777.0, "step": 22811 }, { "epoch": 2.9019208752067165, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003566741943359375, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.2528624534606934, "learning_rate": 1e-05, "loss": 0.4044, "mean_token_accuracy": 0.8864042162895203, "num_tokens": 866195212.0, "step": 22812 }, { "epoch": 2.9020480854853075, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.191686630249023, "learning_rate": 1e-05, "loss": 0.4278, "mean_token_accuracy": 0.8852946162223816, "num_tokens": 866236252.0, "step": 22813 }, { "epoch": 2.9021752957638975, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.353074312210083, "learning_rate": 1e-05, "loss": 0.4032, "mean_token_accuracy": 0.8899574279785156, "num_tokens": 866275705.0, "step": 22814 }, { "epoch": 2.902302506042488, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.326875686645508, "learning_rate": 1e-05, "loss": 0.4631, "mean_token_accuracy": 0.8772827386856079, "num_tokens": 866316458.0, "step": 22815 }, { "epoch": 2.9024297163210786, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.589153528213501, "learning_rate": 1e-05, "loss": 0.4577, "mean_token_accuracy": 0.8721212148666382, "num_tokens": 866351107.0, "step": 22816 }, { "epoch": 2.902556926599669, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.273254871368408, "learning_rate": 1e-05, "loss": 0.4034, "mean_token_accuracy": 0.8955416679382324, "num_tokens": 866385381.0, "step": 22817 }, { "epoch": 2.9026841368782597, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7286810874938965, "learning_rate": 1e-05, "loss": 0.4226, "mean_token_accuracy": 0.8860867619514465, "num_tokens": 866427018.0, "step": 22818 }, { "epoch": 2.90281134715685, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.090622425079346, "learning_rate": 1e-05, "loss": 0.4004, "mean_token_accuracy": 0.8938812613487244, "num_tokens": 866463973.0, "step": 22819 }, { "epoch": 2.9029385574354407, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7552483081817627, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8755874037742615, "num_tokens": 866500900.0, "step": 22820 }, { "epoch": 2.9030657677140312, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.887040138244629, "learning_rate": 1e-05, "loss": 0.3966, "mean_token_accuracy": 0.8943794369697571, "num_tokens": 866539757.0, "step": 22821 }, { "epoch": 2.9031929779926218, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6979291439056396, "learning_rate": 1e-05, "loss": 0.4399, "mean_token_accuracy": 0.8783717155456543, "num_tokens": 866577454.0, "step": 22822 }, { "epoch": 2.9033201882712123, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.738818407058716, "learning_rate": 1e-05, "loss": 0.4012, "mean_token_accuracy": 0.8920832872390747, "num_tokens": 866615032.0, "step": 22823 }, { "epoch": 2.903447398549803, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6442363262176514, "learning_rate": 1e-05, "loss": 0.4476, "mean_token_accuracy": 0.8739684820175171, "num_tokens": 866649700.0, "step": 22824 }, { "epoch": 2.9035746088283934, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6375486850738525, "learning_rate": 1e-05, "loss": 0.4462, "mean_token_accuracy": 0.8759952187538147, "num_tokens": 866684720.0, "step": 22825 }, { "epoch": 2.903701819106984, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.57875919342041, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8778671622276306, "num_tokens": 866722127.0, "step": 22826 }, { "epoch": 2.9038290293855744, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.588510036468506, "learning_rate": 1e-05, "loss": 0.4272, "mean_token_accuracy": 0.882485032081604, "num_tokens": 866757093.0, "step": 22827 }, { "epoch": 2.903956239664165, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5207810401916504, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8724143505096436, "num_tokens": 866796229.0, "step": 22828 }, { "epoch": 2.9040834499427555, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.567032814025879, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8805853128433228, "num_tokens": 866831867.0, "step": 22829 }, { "epoch": 2.904210660221346, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.464940071105957, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8704564571380615, "num_tokens": 866876246.0, "step": 22830 }, { "epoch": 2.9043378704999365, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.591160774230957, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8783579468727112, "num_tokens": 866909555.0, "step": 22831 }, { "epoch": 2.904465080778527, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4137003421783447, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8763991594314575, "num_tokens": 866945208.0, "step": 22832 }, { "epoch": 2.9045922910571176, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.6414642333984375, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8835197687149048, "num_tokens": 866976832.0, "step": 22833 }, { "epoch": 2.904719501335708, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3428025245666504, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8747369050979614, "num_tokens": 867013964.0, "step": 22834 }, { "epoch": 2.904846711614298, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.7259907722473145, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8758062124252319, "num_tokens": 867054361.0, "step": 22835 }, { "epoch": 2.904973921892889, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.259237051010132, "learning_rate": 1e-05, "loss": 0.3757, "mean_token_accuracy": 0.8954474925994873, "num_tokens": 867098732.0, "step": 22836 }, { "epoch": 2.9051011321714793, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.914841890335083, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8638224601745605, "num_tokens": 867133839.0, "step": 22837 }, { "epoch": 2.9052283424500702, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2031161785125732, "learning_rate": 1e-05, "loss": 0.4928, "mean_token_accuracy": 0.8573460578918457, "num_tokens": 867175419.0, "step": 22838 }, { "epoch": 2.9053555527286603, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.149206638336182, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8725712299346924, "num_tokens": 867216585.0, "step": 22839 }, { "epoch": 2.905482763007251, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.253138542175293, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8841679096221924, "num_tokens": 867257228.0, "step": 22840 }, { "epoch": 2.9056099732858414, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.400997161865234, "learning_rate": 1e-05, "loss": 0.4629, "mean_token_accuracy": 0.8773789405822754, "num_tokens": 867297504.0, "step": 22841 }, { "epoch": 2.905737183564432, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.521043539047241, "learning_rate": 1e-05, "loss": 0.4421, "mean_token_accuracy": 0.8802536725997925, "num_tokens": 867336799.0, "step": 22842 }, { "epoch": 2.9058643938430224, "ewc_loss": 0.107421875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 4.467381954193115, "learning_rate": 1e-05, "loss": 0.4544, "mean_token_accuracy": 0.878178060054779, "num_tokens": 867371337.0, "step": 22843 }, { "epoch": 2.905991604121613, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.809420585632324, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8822705745697021, "num_tokens": 867411870.0, "step": 22844 }, { "epoch": 2.9061188144002035, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.275134086608887, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8675621151924133, "num_tokens": 867450278.0, "step": 22845 }, { "epoch": 2.906246024678794, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 3.8699700832366943, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8728686571121216, "num_tokens": 867490316.0, "step": 22846 }, { "epoch": 2.9063732349573845, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.058078765869141, "learning_rate": 1e-05, "loss": 0.4205, "mean_token_accuracy": 0.8865671753883362, "num_tokens": 867526243.0, "step": 22847 }, { "epoch": 2.906500445235975, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.797619104385376, "learning_rate": 1e-05, "loss": 0.4575, "mean_token_accuracy": 0.8767258524894714, "num_tokens": 867568757.0, "step": 22848 }, { "epoch": 2.9066276555145656, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8889732360839844, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8848667740821838, "num_tokens": 867602164.0, "step": 22849 }, { "epoch": 2.906754865793156, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6937077045440674, "learning_rate": 1e-05, "loss": 0.4547, "mean_token_accuracy": 0.8724578619003296, "num_tokens": 867640722.0, "step": 22850 }, { "epoch": 2.9068820760717466, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.773876428604126, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.881252646446228, "num_tokens": 867679568.0, "step": 22851 }, { "epoch": 2.907009286350337, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5927398204803467, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8716033697128296, "num_tokens": 867714991.0, "step": 22852 }, { "epoch": 2.9071364966289277, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7329044342041016, "learning_rate": 1e-05, "loss": 0.4521, "mean_token_accuracy": 0.8731539249420166, "num_tokens": 867753180.0, "step": 22853 }, { "epoch": 2.9072637069075182, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4873385429382324, "learning_rate": 1e-05, "loss": 0.424, "mean_token_accuracy": 0.8803947567939758, "num_tokens": 867789535.0, "step": 22854 }, { "epoch": 2.9073909171861088, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.762430429458618, "learning_rate": 1e-05, "loss": 0.4179, "mean_token_accuracy": 0.88131183385849, "num_tokens": 867821830.0, "step": 22855 }, { "epoch": 2.9075181274646993, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.393430709838867, "learning_rate": 1e-05, "loss": 0.4192, "mean_token_accuracy": 0.8834902048110962, "num_tokens": 867859878.0, "step": 22856 }, { "epoch": 2.90764533774329, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.846468687057495, "learning_rate": 1e-05, "loss": 0.4494, "mean_token_accuracy": 0.8739930391311646, "num_tokens": 867897212.0, "step": 22857 }, { "epoch": 2.90777254802188, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.336750030517578, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8659130334854126, "num_tokens": 867929899.0, "step": 22858 }, { "epoch": 2.907899758300471, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.976837635040283, "learning_rate": 1e-05, "loss": 0.4143, "mean_token_accuracy": 0.8850665092468262, "num_tokens": 867962313.0, "step": 22859 }, { "epoch": 2.908026968579061, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.3277106285095215, "learning_rate": 1e-05, "loss": 0.5077, "mean_token_accuracy": 0.8565784692764282, "num_tokens": 867998886.0, "step": 22860 }, { "epoch": 2.908154178857652, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.130686283111572, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8781673908233643, "num_tokens": 868039151.0, "step": 22861 }, { "epoch": 2.908281389136242, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.4086833000183105, "learning_rate": 1e-05, "loss": 0.4236, "mean_token_accuracy": 0.8821418285369873, "num_tokens": 868071021.0, "step": 22862 }, { "epoch": 2.908408599414833, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.21084451675415, "learning_rate": 1e-05, "loss": 0.4413, "mean_token_accuracy": 0.8818567991256714, "num_tokens": 868109872.0, "step": 22863 }, { "epoch": 2.908535809693423, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.5676257610321045, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8792481422424316, "num_tokens": 868150258.0, "step": 22864 }, { "epoch": 2.9086630199720136, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.135040283203125, "learning_rate": 1e-05, "loss": 0.4745, "mean_token_accuracy": 0.86761075258255, "num_tokens": 868195202.0, "step": 22865 }, { "epoch": 2.908790230250604, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6905319690704346, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.888410210609436, "num_tokens": 868226458.0, "step": 22866 }, { "epoch": 2.9089174405291947, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.952045440673828, "learning_rate": 1e-05, "loss": 0.443, "mean_token_accuracy": 0.8764213919639587, "num_tokens": 868262908.0, "step": 22867 }, { "epoch": 2.909044650807785, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7057597637176514, "learning_rate": 1e-05, "loss": 0.4675, "mean_token_accuracy": 0.8697580099105835, "num_tokens": 868300013.0, "step": 22868 }, { "epoch": 2.9091718610863757, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.7377266883850098, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8799781799316406, "num_tokens": 868342794.0, "step": 22869 }, { "epoch": 2.9092990713649662, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.696729898452759, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8748314380645752, "num_tokens": 868382009.0, "step": 22870 }, { "epoch": 2.9094262816435568, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5769386291503906, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8661963939666748, "num_tokens": 868417770.0, "step": 22871 }, { "epoch": 2.9095534919221473, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.67995548248291, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.8843361735343933, "num_tokens": 868455447.0, "step": 22872 }, { "epoch": 2.909680702200738, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.45809268951416, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8741201162338257, "num_tokens": 868500388.0, "step": 22873 }, { "epoch": 2.9098079124793284, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.700460910797119, "learning_rate": 1e-05, "loss": 0.441, "mean_token_accuracy": 0.8794890642166138, "num_tokens": 868535707.0, "step": 22874 }, { "epoch": 2.909935122757919, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3820128440856934, "learning_rate": 1e-05, "loss": 0.4927, "mean_token_accuracy": 0.8618517518043518, "num_tokens": 868576634.0, "step": 22875 }, { "epoch": 2.9100623330365094, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.735031843185425, "learning_rate": 1e-05, "loss": 0.4155, "mean_token_accuracy": 0.8855996131896973, "num_tokens": 868619914.0, "step": 22876 }, { "epoch": 2.9101895433151, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.328455686569214, "learning_rate": 1e-05, "loss": 0.4207, "mean_token_accuracy": 0.8820818066596985, "num_tokens": 868652279.0, "step": 22877 }, { "epoch": 2.9103167535936905, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.826092481613159, "learning_rate": 1e-05, "loss": 0.4585, "mean_token_accuracy": 0.8716758489608765, "num_tokens": 868689418.0, "step": 22878 }, { "epoch": 2.910443963872281, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.259899377822876, "learning_rate": 1e-05, "loss": 0.3811, "mean_token_accuracy": 0.8932924270629883, "num_tokens": 868726695.0, "step": 22879 }, { "epoch": 2.9105711741508715, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.9829766750335693, "learning_rate": 1e-05, "loss": 0.4506, "mean_token_accuracy": 0.8725099563598633, "num_tokens": 868766120.0, "step": 22880 }, { "epoch": 2.910698384429462, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2465920448303223, "learning_rate": 1e-05, "loss": 0.3955, "mean_token_accuracy": 0.8901411294937134, "num_tokens": 868806041.0, "step": 22881 }, { "epoch": 2.9108255947080526, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.173141956329346, "learning_rate": 1e-05, "loss": 0.4501, "mean_token_accuracy": 0.8774436712265015, "num_tokens": 868852980.0, "step": 22882 }, { "epoch": 2.9109528049866427, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.3615360260009766, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8681367635726929, "num_tokens": 868890951.0, "step": 22883 }, { "epoch": 2.9110800152652336, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.320934295654297, "learning_rate": 1e-05, "loss": 0.4836, "mean_token_accuracy": 0.87043696641922, "num_tokens": 868926412.0, "step": 22884 }, { "epoch": 2.9112072255438237, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.5639116764068604, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.8713994026184082, "num_tokens": 868965181.0, "step": 22885 }, { "epoch": 2.9113344358224147, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.278576374053955, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.881562352180481, "num_tokens": 869004635.0, "step": 22886 }, { "epoch": 2.9114616461010048, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.713141918182373, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8798598051071167, "num_tokens": 869043833.0, "step": 22887 }, { "epoch": 2.9115888563795957, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.139164447784424, "learning_rate": 1e-05, "loss": 0.4776, "mean_token_accuracy": 0.8688015937805176, "num_tokens": 869078868.0, "step": 22888 }, { "epoch": 2.911716066658186, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.756112813949585, "learning_rate": 1e-05, "loss": 0.4378, "mean_token_accuracy": 0.8820496797561646, "num_tokens": 869113444.0, "step": 22889 }, { "epoch": 2.9118432769367764, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.940030574798584, "learning_rate": 1e-05, "loss": 0.4203, "mean_token_accuracy": 0.8835016489028931, "num_tokens": 869147741.0, "step": 22890 }, { "epoch": 2.911970487215367, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7151060104370117, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8754051327705383, "num_tokens": 869180247.0, "step": 22891 }, { "epoch": 2.9120976974939574, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.773432970046997, "learning_rate": 1e-05, "loss": 0.3929, "mean_token_accuracy": 0.8942264914512634, "num_tokens": 869221602.0, "step": 22892 }, { "epoch": 2.912224907772548, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6430156230926514, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.879358172416687, "num_tokens": 869259601.0, "step": 22893 }, { "epoch": 2.9123521180511385, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6637253761291504, "learning_rate": 1e-05, "loss": 0.3951, "mean_token_accuracy": 0.8912429809570312, "num_tokens": 869303366.0, "step": 22894 }, { "epoch": 2.912479328329729, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5882840156555176, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8630921840667725, "num_tokens": 869338579.0, "step": 22895 }, { "epoch": 2.9126065386083195, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6072864532470703, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8654753565788269, "num_tokens": 869384129.0, "step": 22896 }, { "epoch": 2.91273374888691, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.533959150314331, "learning_rate": 1e-05, "loss": 0.4338, "mean_token_accuracy": 0.8777938485145569, "num_tokens": 869421752.0, "step": 22897 }, { "epoch": 2.9128609591655006, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.584165573120117, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8737823963165283, "num_tokens": 869459309.0, "step": 22898 }, { "epoch": 2.912988169444091, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4761204719543457, "learning_rate": 1e-05, "loss": 0.4037, "mean_token_accuracy": 0.8883394002914429, "num_tokens": 869497718.0, "step": 22899 }, { "epoch": 2.9131153797226816, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5838708877563477, "learning_rate": 1e-05, "loss": 0.4638, "mean_token_accuracy": 0.871023952960968, "num_tokens": 869540522.0, "step": 22900 }, { "epoch": 2.913242590001272, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.432222604751587, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8702868223190308, "num_tokens": 869579610.0, "step": 22901 }, { "epoch": 2.9133698002798627, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6205406188964844, "learning_rate": 1e-05, "loss": 0.4109, "mean_token_accuracy": 0.8874168395996094, "num_tokens": 869619186.0, "step": 22902 }, { "epoch": 2.9134970105584532, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3675119876861572, "learning_rate": 1e-05, "loss": 0.446, "mean_token_accuracy": 0.871582567691803, "num_tokens": 869657415.0, "step": 22903 }, { "epoch": 2.9136242208370438, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6962382793426514, "learning_rate": 1e-05, "loss": 0.4508, "mean_token_accuracy": 0.8744871020317078, "num_tokens": 869698319.0, "step": 22904 }, { "epoch": 2.9137514311156343, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2884199619293213, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8878257274627686, "num_tokens": 869738630.0, "step": 22905 }, { "epoch": 2.913878641394225, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.8317062854766846, "learning_rate": 1e-05, "loss": 0.4716, "mean_token_accuracy": 0.8661529421806335, "num_tokens": 869780713.0, "step": 22906 }, { "epoch": 2.9140058516728153, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.7206878662109375e-05, "grad_norm": 3.2172935009002686, "learning_rate": 1e-05, "loss": 0.3914, "mean_token_accuracy": 0.8914022445678711, "num_tokens": 869817046.0, "step": 22907 }, { "epoch": 2.9141330619514054, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.0559611320495605, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8868637084960938, "num_tokens": 869855099.0, "step": 22908 }, { "epoch": 2.9142602722299964, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.2346744537353516, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8766897320747375, "num_tokens": 869893207.0, "step": 22909 }, { "epoch": 2.9143874825085865, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.3094682693481445, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.879539966583252, "num_tokens": 869929925.0, "step": 22910 }, { "epoch": 2.9145146927871775, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.3890204429626465, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8826854228973389, "num_tokens": 869974607.0, "step": 22911 }, { "epoch": 2.9146419030657675, "ewc_loss": 0.107421875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 7.152557373046875e-05, "grad_norm": 4.466081142425537, "learning_rate": 1e-05, "loss": 0.4471, "mean_token_accuracy": 0.8772745728492737, "num_tokens": 870013056.0, "step": 22912 }, { "epoch": 2.914769113344358, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.7225277423858643, "learning_rate": 1e-05, "loss": 0.473, "mean_token_accuracy": 0.8689934015274048, "num_tokens": 870053735.0, "step": 22913 }, { "epoch": 2.9148963236229486, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.33767557144165, "learning_rate": 1e-05, "loss": 0.4336, "mean_token_accuracy": 0.8862239122390747, "num_tokens": 870093607.0, "step": 22914 }, { "epoch": 2.915023533901539, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 3.8552744388580322, "learning_rate": 1e-05, "loss": 0.4417, "mean_token_accuracy": 0.8790678977966309, "num_tokens": 870133118.0, "step": 22915 }, { "epoch": 2.9151507441801296, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.125669002532959, "learning_rate": 1e-05, "loss": 0.4398, "mean_token_accuracy": 0.8803553581237793, "num_tokens": 870169782.0, "step": 22916 }, { "epoch": 2.91527795445872, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.8151886463165283, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8783752918243408, "num_tokens": 870206891.0, "step": 22917 }, { "epoch": 2.9154051647373107, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9330766201019287, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8703614473342896, "num_tokens": 870253035.0, "step": 22918 }, { "epoch": 2.9155323750159012, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7256548404693604, "learning_rate": 1e-05, "loss": 0.4473, "mean_token_accuracy": 0.8783367872238159, "num_tokens": 870292286.0, "step": 22919 }, { "epoch": 2.9156595852944918, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.798335552215576, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8688690662384033, "num_tokens": 870335594.0, "step": 22920 }, { "epoch": 2.9157867955730823, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.615257501602173, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8797527551651001, "num_tokens": 870379465.0, "step": 22921 }, { "epoch": 2.915914005851673, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.734236240386963, "learning_rate": 1e-05, "loss": 0.4072, "mean_token_accuracy": 0.8891578912734985, "num_tokens": 870418060.0, "step": 22922 }, { "epoch": 2.9160412161302633, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5220723152160645, "learning_rate": 1e-05, "loss": 0.4018, "mean_token_accuracy": 0.8847893476486206, "num_tokens": 870455565.0, "step": 22923 }, { "epoch": 2.916168426408854, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7278809547424316, "learning_rate": 1e-05, "loss": 0.4438, "mean_token_accuracy": 0.8762050271034241, "num_tokens": 870492609.0, "step": 22924 }, { "epoch": 2.9162956366874444, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.43892765045166, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8771299719810486, "num_tokens": 870528150.0, "step": 22925 }, { "epoch": 2.916422846966035, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.762814998626709, "learning_rate": 1e-05, "loss": 0.42, "mean_token_accuracy": 0.8828126788139343, "num_tokens": 870568188.0, "step": 22926 }, { "epoch": 2.9165500572446255, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.369415283203125, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.88162761926651, "num_tokens": 870601459.0, "step": 22927 }, { "epoch": 2.916677267523216, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8726024627685547, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.880936861038208, "num_tokens": 870638976.0, "step": 22928 }, { "epoch": 2.9168044778018065, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3153207302093506, "learning_rate": 1e-05, "loss": 0.4587, "mean_token_accuracy": 0.8651829957962036, "num_tokens": 870671163.0, "step": 22929 }, { "epoch": 2.916931688080397, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.017121315002441, "learning_rate": 1e-05, "loss": 0.4561, "mean_token_accuracy": 0.8788821697235107, "num_tokens": 870705806.0, "step": 22930 }, { "epoch": 2.917058898358987, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.317265510559082, "learning_rate": 1e-05, "loss": 0.4437, "mean_token_accuracy": 0.8732045888900757, "num_tokens": 870743045.0, "step": 22931 }, { "epoch": 2.917186108637578, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.17348051071167, "learning_rate": 1e-05, "loss": 0.4174, "mean_token_accuracy": 0.887099027633667, "num_tokens": 870780965.0, "step": 22932 }, { "epoch": 2.917313318916168, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.444303274154663, "learning_rate": 1e-05, "loss": 0.4009, "mean_token_accuracy": 0.8910435438156128, "num_tokens": 870814694.0, "step": 22933 }, { "epoch": 2.917440529194759, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.24482536315918, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8734251856803894, "num_tokens": 870845654.0, "step": 22934 }, { "epoch": 2.9175677394733492, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.6161630153656006, "learning_rate": 1e-05, "loss": 0.4366, "mean_token_accuracy": 0.8794165253639221, "num_tokens": 870881720.0, "step": 22935 }, { "epoch": 2.91769494975194, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.112793922424316, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8830900192260742, "num_tokens": 870927480.0, "step": 22936 }, { "epoch": 2.9178221600305303, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7077126502990723, "learning_rate": 1e-05, "loss": 0.4442, "mean_token_accuracy": 0.8835135698318481, "num_tokens": 870962407.0, "step": 22937 }, { "epoch": 2.917949370309121, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9287631511688232, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8887786865234375, "num_tokens": 870994638.0, "step": 22938 }, { "epoch": 2.9180765805877114, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7003653049468994, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.8836997747421265, "num_tokens": 871032638.0, "step": 22939 }, { "epoch": 2.918203790866302, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7379233837127686, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8775882124900818, "num_tokens": 871070955.0, "step": 22940 }, { "epoch": 2.9183310011448924, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6748549938201904, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8772559762001038, "num_tokens": 871109282.0, "step": 22941 }, { "epoch": 2.918458211423483, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.583300828933716, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8693710565567017, "num_tokens": 871151214.0, "step": 22942 }, { "epoch": 2.9185854217020735, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.659761428833008, "learning_rate": 1e-05, "loss": 0.4709, "mean_token_accuracy": 0.8698675632476807, "num_tokens": 871192909.0, "step": 22943 }, { "epoch": 2.918712631980664, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.480659008026123, "learning_rate": 1e-05, "loss": 0.4498, "mean_token_accuracy": 0.8755666017532349, "num_tokens": 871234952.0, "step": 22944 }, { "epoch": 2.9188398422592545, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.665070056915283, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8787195086479187, "num_tokens": 871270789.0, "step": 22945 }, { "epoch": 2.918967052537845, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.407799005508423, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8715576529502869, "num_tokens": 871312425.0, "step": 22946 }, { "epoch": 2.9190942628164356, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6922221183776855, "learning_rate": 1e-05, "loss": 0.4019, "mean_token_accuracy": 0.8917333483695984, "num_tokens": 871347685.0, "step": 22947 }, { "epoch": 2.919221473095026, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.343921184539795, "learning_rate": 1e-05, "loss": 0.4263, "mean_token_accuracy": 0.8786532282829285, "num_tokens": 871381789.0, "step": 22948 }, { "epoch": 2.9193486833736166, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.762908697128296, "learning_rate": 1e-05, "loss": 0.4002, "mean_token_accuracy": 0.8900914192199707, "num_tokens": 871417578.0, "step": 22949 }, { "epoch": 2.919475893652207, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2785470485687256, "learning_rate": 1e-05, "loss": 0.4296, "mean_token_accuracy": 0.877956211566925, "num_tokens": 871454816.0, "step": 22950 }, { "epoch": 2.9196031039307977, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.904341220855713, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8795434236526489, "num_tokens": 871488898.0, "step": 22951 }, { "epoch": 2.9197303142093882, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.222754955291748, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8847296237945557, "num_tokens": 871531895.0, "step": 22952 }, { "epoch": 2.9198575244879788, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.124330520629883, "learning_rate": 1e-05, "loss": 0.4262, "mean_token_accuracy": 0.8867223262786865, "num_tokens": 871566151.0, "step": 22953 }, { "epoch": 2.9199847347665693, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.289395332336426, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8754076361656189, "num_tokens": 871603777.0, "step": 22954 }, { "epoch": 2.92011194504516, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.323770046234131, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8844234347343445, "num_tokens": 871636118.0, "step": 22955 }, { "epoch": 2.92023915532375, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.489044427871704, "learning_rate": 1e-05, "loss": 0.438, "mean_token_accuracy": 0.8781308531761169, "num_tokens": 871670948.0, "step": 22956 }, { "epoch": 2.920366365602341, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.369468688964844, "learning_rate": 1e-05, "loss": 0.4788, "mean_token_accuracy": 0.8723534345626831, "num_tokens": 871707987.0, "step": 22957 }, { "epoch": 2.920493575880931, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7054426670074463, "learning_rate": 1e-05, "loss": 0.4295, "mean_token_accuracy": 0.8860875368118286, "num_tokens": 871753861.0, "step": 22958 }, { "epoch": 2.920620786159522, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.244639873504639, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8746628761291504, "num_tokens": 871792667.0, "step": 22959 }, { "epoch": 2.920747996438112, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.8058362007141113, "learning_rate": 1e-05, "loss": 0.5051, "mean_token_accuracy": 0.8606012463569641, "num_tokens": 871830994.0, "step": 22960 }, { "epoch": 2.920875206716703, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.037454128265381, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8661808967590332, "num_tokens": 871869984.0, "step": 22961 }, { "epoch": 2.921002416995293, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.773063898086548, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.866777777671814, "num_tokens": 871906757.0, "step": 22962 }, { "epoch": 2.9211296272738836, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.843581438064575, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.8848826885223389, "num_tokens": 871945387.0, "step": 22963 }, { "epoch": 2.921256837552474, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.688436269760132, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8795545101165771, "num_tokens": 871981456.0, "step": 22964 }, { "epoch": 2.9213840478310646, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.7043633460998535, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8849338889122009, "num_tokens": 872022953.0, "step": 22965 }, { "epoch": 2.921511258109655, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6125094890594482, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.879561185836792, "num_tokens": 872059103.0, "step": 22966 }, { "epoch": 2.9216384683882457, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6469151973724365, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8791472911834717, "num_tokens": 872091910.0, "step": 22967 }, { "epoch": 2.9217656786668362, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5439140796661377, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8784105777740479, "num_tokens": 872130863.0, "step": 22968 }, { "epoch": 2.9218928889454268, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.623898506164551, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8689166307449341, "num_tokens": 872168104.0, "step": 22969 }, { "epoch": 2.9220200992240173, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.5047693252563477, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8741837739944458, "num_tokens": 872197554.0, "step": 22970 }, { "epoch": 2.922147309502608, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6053571701049805, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8789564967155457, "num_tokens": 872232283.0, "step": 22971 }, { "epoch": 2.9222745197811983, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4439682960510254, "learning_rate": 1e-05, "loss": 0.4173, "mean_token_accuracy": 0.8802029490470886, "num_tokens": 872263459.0, "step": 22972 }, { "epoch": 2.922401730059789, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6158151626586914, "learning_rate": 1e-05, "loss": 0.4004, "mean_token_accuracy": 0.8906466960906982, "num_tokens": 872303056.0, "step": 22973 }, { "epoch": 2.9225289403383794, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.386256694793701, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.875623345375061, "num_tokens": 872341030.0, "step": 22974 }, { "epoch": 2.92265615061697, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.675475835800171, "learning_rate": 1e-05, "loss": 0.4619, "mean_token_accuracy": 0.8731268644332886, "num_tokens": 872384006.0, "step": 22975 }, { "epoch": 2.9227833608955605, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.321160316467285, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8688066005706787, "num_tokens": 872425575.0, "step": 22976 }, { "epoch": 2.922910571174151, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.7828285694122314, "learning_rate": 1e-05, "loss": 0.4258, "mean_token_accuracy": 0.8824366331100464, "num_tokens": 872463483.0, "step": 22977 }, { "epoch": 2.9230377814527415, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2709500789642334, "learning_rate": 1e-05, "loss": 0.4419, "mean_token_accuracy": 0.8765096664428711, "num_tokens": 872496066.0, "step": 22978 }, { "epoch": 2.923164991731332, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9450063705444336, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8820698857307434, "num_tokens": 872529330.0, "step": 22979 }, { "epoch": 2.9232922020099226, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.2347161769866943, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8827178478240967, "num_tokens": 872563989.0, "step": 22980 }, { "epoch": 2.9234194122885127, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.1422200202941895, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8852145671844482, "num_tokens": 872601506.0, "step": 22981 }, { "epoch": 2.9235466225671036, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.3085429668426514, "learning_rate": 1e-05, "loss": 0.4231, "mean_token_accuracy": 0.8831389546394348, "num_tokens": 872633258.0, "step": 22982 }, { "epoch": 2.9236738328456937, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.345199108123779, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.866294801235199, "num_tokens": 872665803.0, "step": 22983 }, { "epoch": 2.9238010431242847, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.9604644775390625e-05, "grad_norm": 3.5232441425323486, "learning_rate": 1e-05, "loss": 0.444, "mean_token_accuracy": 0.8775990009307861, "num_tokens": 872703607.0, "step": 22984 }, { "epoch": 2.9239282534028748, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.361218452453613, "learning_rate": 1e-05, "loss": 0.4622, "mean_token_accuracy": 0.8763502836227417, "num_tokens": 872741966.0, "step": 22985 }, { "epoch": 2.9240554636814657, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.761273145675659, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.8812718391418457, "num_tokens": 872773052.0, "step": 22986 }, { "epoch": 2.924182673960056, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.181415557861328, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8686582446098328, "num_tokens": 872813667.0, "step": 22987 }, { "epoch": 2.9243098842386464, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.782099723815918, "learning_rate": 1e-05, "loss": 0.4908, "mean_token_accuracy": 0.8673627376556396, "num_tokens": 872854320.0, "step": 22988 }, { "epoch": 2.924437094517237, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.9872946739196777, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8786463737487793, "num_tokens": 872895190.0, "step": 22989 }, { "epoch": 2.9245643047958274, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7327401638031006, "learning_rate": 1e-05, "loss": 0.4394, "mean_token_accuracy": 0.8811594843864441, "num_tokens": 872932288.0, "step": 22990 }, { "epoch": 2.924691515074418, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8356218338012695, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8834753036499023, "num_tokens": 872963617.0, "step": 22991 }, { "epoch": 2.9248187253530085, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.6651198863983154, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8716226816177368, "num_tokens": 872997635.0, "step": 22992 }, { "epoch": 2.924945935631599, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7167985439300537, "learning_rate": 1e-05, "loss": 0.4881, "mean_token_accuracy": 0.8622478246688843, "num_tokens": 873037275.0, "step": 22993 }, { "epoch": 2.9250731459101895, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.57780122756958, "learning_rate": 1e-05, "loss": 0.4592, "mean_token_accuracy": 0.8706986904144287, "num_tokens": 873074791.0, "step": 22994 }, { "epoch": 2.92520035618878, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6692373752593994, "learning_rate": 1e-05, "loss": 0.4159, "mean_token_accuracy": 0.8850792646408081, "num_tokens": 873111521.0, "step": 22995 }, { "epoch": 2.9253275664673706, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5025479793548584, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8589414358139038, "num_tokens": 873150007.0, "step": 22996 }, { "epoch": 2.925454776745961, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.660161256790161, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8790773153305054, "num_tokens": 873188360.0, "step": 22997 }, { "epoch": 2.9255819870245516, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4289374351501465, "learning_rate": 1e-05, "loss": 0.3962, "mean_token_accuracy": 0.8864521384239197, "num_tokens": 873222962.0, "step": 22998 }, { "epoch": 2.925709197303142, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.694333076477051, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8862518072128296, "num_tokens": 873260579.0, "step": 22999 }, { "epoch": 2.9258364075817327, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3725998401641846, "learning_rate": 1e-05, "loss": 0.4017, "mean_token_accuracy": 0.8878788948059082, "num_tokens": 873292862.0, "step": 23000 }, { "epoch": 2.925963617860323, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7871387004852295, "learning_rate": 1e-05, "loss": 0.5113, "mean_token_accuracy": 0.8542929887771606, "num_tokens": 873331824.0, "step": 23001 }, { "epoch": 2.9260908281389137, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3054983615875244, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8780013918876648, "num_tokens": 873364902.0, "step": 23002 }, { "epoch": 2.9262180384175043, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.9132587909698486, "learning_rate": 1e-05, "loss": 0.3964, "mean_token_accuracy": 0.8936837911605835, "num_tokens": 873398523.0, "step": 23003 }, { "epoch": 2.926345248696095, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2640762329101562, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8808920979499817, "num_tokens": 873435779.0, "step": 23004 }, { "epoch": 2.9264724589746853, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.091209411621094, "learning_rate": 1e-05, "loss": 0.421, "mean_token_accuracy": 0.8895964622497559, "num_tokens": 873476706.0, "step": 23005 }, { "epoch": 2.9265996692532754, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3287787437438965, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8800659775733948, "num_tokens": 873507762.0, "step": 23006 }, { "epoch": 2.9267268795318664, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.241515159606934, "learning_rate": 1e-05, "loss": 0.4163, "mean_token_accuracy": 0.8865993022918701, "num_tokens": 873545692.0, "step": 23007 }, { "epoch": 2.9268540898104565, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.478135585784912, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8816968202590942, "num_tokens": 873589131.0, "step": 23008 }, { "epoch": 2.9269813000890474, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.274352550506592, "learning_rate": 1e-05, "loss": 0.5009, "mean_token_accuracy": 0.8629869818687439, "num_tokens": 873628593.0, "step": 23009 }, { "epoch": 2.9271085103676375, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.667520523071289, "learning_rate": 1e-05, "loss": 0.4637, "mean_token_accuracy": 0.8720718026161194, "num_tokens": 873671525.0, "step": 23010 }, { "epoch": 2.927235720646228, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.128822326660156, "learning_rate": 1e-05, "loss": 0.4569, "mean_token_accuracy": 0.8793630599975586, "num_tokens": 873710618.0, "step": 23011 }, { "epoch": 2.9273629309248186, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.745774507522583, "learning_rate": 1e-05, "loss": 0.5002, "mean_token_accuracy": 0.862135648727417, "num_tokens": 873751731.0, "step": 23012 }, { "epoch": 2.927490141203409, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9223744869232178, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.883776068687439, "num_tokens": 873791948.0, "step": 23013 }, { "epoch": 2.9276173514819996, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.718096971511841, "learning_rate": 1e-05, "loss": 0.4616, "mean_token_accuracy": 0.8707091808319092, "num_tokens": 873834193.0, "step": 23014 }, { "epoch": 2.92774456176059, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7543435096740723, "learning_rate": 1e-05, "loss": 0.4614, "mean_token_accuracy": 0.8715960383415222, "num_tokens": 873870035.0, "step": 23015 }, { "epoch": 2.9278717720391807, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.6801645755767822, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8841579556465149, "num_tokens": 873902007.0, "step": 23016 }, { "epoch": 2.9279989823177712, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.602015733718872, "learning_rate": 1e-05, "loss": 0.4938, "mean_token_accuracy": 0.859968364238739, "num_tokens": 873947063.0, "step": 23017 }, { "epoch": 2.9281261925963618, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6360199451446533, "learning_rate": 1e-05, "loss": 0.4713, "mean_token_accuracy": 0.8686925768852234, "num_tokens": 873987542.0, "step": 23018 }, { "epoch": 2.9282534028749523, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5547077655792236, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8790940046310425, "num_tokens": 874020492.0, "step": 23019 }, { "epoch": 2.928380613153543, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5765535831451416, "learning_rate": 1e-05, "loss": 0.3995, "mean_token_accuracy": 0.8881056308746338, "num_tokens": 874061902.0, "step": 23020 }, { "epoch": 2.9285078234321333, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.505392551422119, "learning_rate": 1e-05, "loss": 0.3883, "mean_token_accuracy": 0.893215000629425, "num_tokens": 874093277.0, "step": 23021 }, { "epoch": 2.928635033710724, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5367794036865234, "learning_rate": 1e-05, "loss": 0.4043, "mean_token_accuracy": 0.8868510723114014, "num_tokens": 874136543.0, "step": 23022 }, { "epoch": 2.9287622439893144, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.4752750396728516, "learning_rate": 1e-05, "loss": 0.4003, "mean_token_accuracy": 0.8885959982872009, "num_tokens": 874180933.0, "step": 23023 }, { "epoch": 2.928889454267905, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5299017429351807, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.883115291595459, "num_tokens": 874221980.0, "step": 23024 }, { "epoch": 2.9290166645464955, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4769482612609863, "learning_rate": 1e-05, "loss": 0.4283, "mean_token_accuracy": 0.8804528713226318, "num_tokens": 874255814.0, "step": 23025 }, { "epoch": 2.929143874825086, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5270700454711914, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8714419007301331, "num_tokens": 874296230.0, "step": 23026 }, { "epoch": 2.9292710851036765, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.464120388031006, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8668051958084106, "num_tokens": 874331854.0, "step": 23027 }, { "epoch": 2.929398295382267, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.512502908706665, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8834760785102844, "num_tokens": 874373348.0, "step": 23028 }, { "epoch": 2.929525505660857, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.439455032348633, "learning_rate": 1e-05, "loss": 0.4129, "mean_token_accuracy": 0.8817446827888489, "num_tokens": 874415226.0, "step": 23029 }, { "epoch": 2.929652715939448, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5369820594787598, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8608245849609375, "num_tokens": 874451894.0, "step": 23030 }, { "epoch": 2.929779926218038, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.409621477127075, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8840703964233398, "num_tokens": 874490248.0, "step": 23031 }, { "epoch": 2.929907136496629, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5647616386413574, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8816475868225098, "num_tokens": 874527421.0, "step": 23032 }, { "epoch": 2.9300343467752192, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.356632709503174, "learning_rate": 1e-05, "loss": 0.4532, "mean_token_accuracy": 0.8717942237854004, "num_tokens": 874570834.0, "step": 23033 }, { "epoch": 2.93016155705381, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6527302265167236, "learning_rate": 1e-05, "loss": 0.466, "mean_token_accuracy": 0.8693569898605347, "num_tokens": 874607956.0, "step": 23034 }, { "epoch": 2.9302887673324003, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.287428855895996, "learning_rate": 1e-05, "loss": 0.4459, "mean_token_accuracy": 0.8735437393188477, "num_tokens": 874644599.0, "step": 23035 }, { "epoch": 2.930415977610991, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7837438583374023, "learning_rate": 1e-05, "loss": 0.4649, "mean_token_accuracy": 0.8720166683197021, "num_tokens": 874682214.0, "step": 23036 }, { "epoch": 2.9305431878895813, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.1857967376708984, "learning_rate": 1e-05, "loss": 0.4625, "mean_token_accuracy": 0.8654218912124634, "num_tokens": 874719298.0, "step": 23037 }, { "epoch": 2.930670398168172, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.0205183029174805, "learning_rate": 1e-05, "loss": 0.4196, "mean_token_accuracy": 0.8880540728569031, "num_tokens": 874757391.0, "step": 23038 }, { "epoch": 2.9307976084467624, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1447067260742188, "learning_rate": 1e-05, "loss": 0.4147, "mean_token_accuracy": 0.8823965787887573, "num_tokens": 874788307.0, "step": 23039 }, { "epoch": 2.930924818725353, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.360591411590576, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8760838508605957, "num_tokens": 874825648.0, "step": 23040 }, { "epoch": 2.9310520290039435, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.306882381439209, "learning_rate": 1e-05, "loss": 0.45, "mean_token_accuracy": 0.8740787506103516, "num_tokens": 874862275.0, "step": 23041 }, { "epoch": 2.931179239282534, "ewc_loss": 0.11328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 7.724761962890625e-05, "grad_norm": 4.673043251037598, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.873033881187439, "num_tokens": 874898411.0, "step": 23042 }, { "epoch": 2.9313064495611245, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.7794227600097656, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8793735504150391, "num_tokens": 874935599.0, "step": 23043 }, { "epoch": 2.931433659839715, "ewc_loss": 0.109375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.5714545249938965, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8758577108383179, "num_tokens": 874966302.0, "step": 23044 }, { "epoch": 2.9315608701183056, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.044719219207764, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8922570943832397, "num_tokens": 875003158.0, "step": 23045 }, { "epoch": 2.931688080396896, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2751784324646, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8795838356018066, "num_tokens": 875042163.0, "step": 23046 }, { "epoch": 2.9318152906754866, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.969805955886841, "learning_rate": 1e-05, "loss": 0.4611, "mean_token_accuracy": 0.8753822445869446, "num_tokens": 875081957.0, "step": 23047 }, { "epoch": 2.931942500954077, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.069930553436279, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.8799234628677368, "num_tokens": 875122286.0, "step": 23048 }, { "epoch": 2.9320697112326677, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.8173577785491943, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8634147047996521, "num_tokens": 875163320.0, "step": 23049 }, { "epoch": 2.932196921511258, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9293501377105713, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8859193325042725, "num_tokens": 875203819.0, "step": 23050 }, { "epoch": 2.9323241317898487, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.683000087738037, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8804395794868469, "num_tokens": 875239028.0, "step": 23051 }, { "epoch": 2.9324513420684393, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8725509643554688, "learning_rate": 1e-05, "loss": 0.4771, "mean_token_accuracy": 0.8684922456741333, "num_tokens": 875276324.0, "step": 23052 }, { "epoch": 2.93257855234703, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5428411960601807, "learning_rate": 1e-05, "loss": 0.4138, "mean_token_accuracy": 0.8843012452125549, "num_tokens": 875316918.0, "step": 23053 }, { "epoch": 2.93270576262562, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.9049243927001953, "learning_rate": 1e-05, "loss": 0.4597, "mean_token_accuracy": 0.8712402582168579, "num_tokens": 875351009.0, "step": 23054 }, { "epoch": 2.932832972904211, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4424445629119873, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.888083815574646, "num_tokens": 875386988.0, "step": 23055 }, { "epoch": 2.932960183182801, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.97758412361145, "learning_rate": 1e-05, "loss": 0.4307, "mean_token_accuracy": 0.8813555240631104, "num_tokens": 875429607.0, "step": 23056 }, { "epoch": 2.933087393461392, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.3962435722351074, "learning_rate": 1e-05, "loss": 0.422, "mean_token_accuracy": 0.8835906386375427, "num_tokens": 875461071.0, "step": 23057 }, { "epoch": 2.933214603739982, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.126897811889648, "learning_rate": 1e-05, "loss": 0.4321, "mean_token_accuracy": 0.8839337229728699, "num_tokens": 875498146.0, "step": 23058 }, { "epoch": 2.933341814018573, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.4236295223236084, "learning_rate": 1e-05, "loss": 0.4377, "mean_token_accuracy": 0.8787966966629028, "num_tokens": 875536155.0, "step": 23059 }, { "epoch": 2.933469024297163, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.271782398223877, "learning_rate": 1e-05, "loss": 0.459, "mean_token_accuracy": 0.8742367029190063, "num_tokens": 875569868.0, "step": 23060 }, { "epoch": 2.9335962345757536, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.595899820327759, "learning_rate": 1e-05, "loss": 0.4422, "mean_token_accuracy": 0.8766621351242065, "num_tokens": 875609298.0, "step": 23061 }, { "epoch": 2.933723444854344, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.217804908752441, "learning_rate": 1e-05, "loss": 0.4027, "mean_token_accuracy": 0.8905559182167053, "num_tokens": 875642837.0, "step": 23062 }, { "epoch": 2.9338506551329346, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.746021032333374, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8738605976104736, "num_tokens": 875683273.0, "step": 23063 }, { "epoch": 2.933977865411525, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.032160758972168, "learning_rate": 1e-05, "loss": 0.4789, "mean_token_accuracy": 0.8640585541725159, "num_tokens": 875717461.0, "step": 23064 }, { "epoch": 2.9341050756901157, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.7852814197540283, "learning_rate": 1e-05, "loss": 0.4699, "mean_token_accuracy": 0.8675323724746704, "num_tokens": 875756975.0, "step": 23065 }, { "epoch": 2.934232285968706, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.77616286277771, "learning_rate": 1e-05, "loss": 0.4176, "mean_token_accuracy": 0.8829623460769653, "num_tokens": 875793414.0, "step": 23066 }, { "epoch": 2.9343594962472968, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7839550971984863, "learning_rate": 1e-05, "loss": 0.4918, "mean_token_accuracy": 0.8588234186172485, "num_tokens": 875822477.0, "step": 23067 }, { "epoch": 2.9344867065258873, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5734148025512695, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.8796212077140808, "num_tokens": 875856220.0, "step": 23068 }, { "epoch": 2.934613916804478, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.753000497817993, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.879640519618988, "num_tokens": 875890572.0, "step": 23069 }, { "epoch": 2.9347411270830683, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4351561069488525, "learning_rate": 1e-05, "loss": 0.4187, "mean_token_accuracy": 0.8868811130523682, "num_tokens": 875928231.0, "step": 23070 }, { "epoch": 2.934868337361659, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.785623073577881, "learning_rate": 1e-05, "loss": 0.4659, "mean_token_accuracy": 0.8710446357727051, "num_tokens": 875966211.0, "step": 23071 }, { "epoch": 2.9349955476402494, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.326587200164795, "learning_rate": 1e-05, "loss": 0.4233, "mean_token_accuracy": 0.879048764705658, "num_tokens": 876004857.0, "step": 23072 }, { "epoch": 2.93512275791884, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8873980045318604, "learning_rate": 1e-05, "loss": 0.4842, "mean_token_accuracy": 0.863430380821228, "num_tokens": 876041966.0, "step": 23073 }, { "epoch": 2.9352499681974304, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2846779823303223, "learning_rate": 1e-05, "loss": 0.4115, "mean_token_accuracy": 0.8834561109542847, "num_tokens": 876075732.0, "step": 23074 }, { "epoch": 2.935377178476021, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.036877155303955, "learning_rate": 1e-05, "loss": 0.4639, "mean_token_accuracy": 0.8736919164657593, "num_tokens": 876116582.0, "step": 23075 }, { "epoch": 2.9355043887546115, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.286526679992676, "learning_rate": 1e-05, "loss": 0.3841, "mean_token_accuracy": 0.8934888243675232, "num_tokens": 876150122.0, "step": 23076 }, { "epoch": 2.935631599033202, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.206752777099609, "learning_rate": 1e-05, "loss": 0.4222, "mean_token_accuracy": 0.8882477283477783, "num_tokens": 876187902.0, "step": 23077 }, { "epoch": 2.9357588093117926, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.420062780380249, "learning_rate": 1e-05, "loss": 0.4496, "mean_token_accuracy": 0.8745710253715515, "num_tokens": 876224304.0, "step": 23078 }, { "epoch": 2.9358860195903826, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.3018951416015625, "learning_rate": 1e-05, "loss": 0.5012, "mean_token_accuracy": 0.8622303009033203, "num_tokens": 876261436.0, "step": 23079 }, { "epoch": 2.9360132298689736, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.612736225128174, "learning_rate": 1e-05, "loss": 0.4613, "mean_token_accuracy": 0.8705720901489258, "num_tokens": 876299996.0, "step": 23080 }, { "epoch": 2.9361404401475637, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2226152420043945, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8686432838439941, "num_tokens": 876337793.0, "step": 23081 }, { "epoch": 2.9362676504261547, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7307958602905273, "learning_rate": 1e-05, "loss": 0.4387, "mean_token_accuracy": 0.8818762302398682, "num_tokens": 876371837.0, "step": 23082 }, { "epoch": 2.9363948607047448, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.033756256103516, "learning_rate": 1e-05, "loss": 0.426, "mean_token_accuracy": 0.8859813213348389, "num_tokens": 876414402.0, "step": 23083 }, { "epoch": 2.9365220709833357, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7268495559692383, "learning_rate": 1e-05, "loss": 0.3931, "mean_token_accuracy": 0.8942409753799438, "num_tokens": 876451790.0, "step": 23084 }, { "epoch": 2.936649281261926, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8830151557922363, "learning_rate": 1e-05, "loss": 0.4764, "mean_token_accuracy": 0.8637298345565796, "num_tokens": 876486891.0, "step": 23085 }, { "epoch": 2.9367764915405163, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.6734910011291504, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8871133327484131, "num_tokens": 876525560.0, "step": 23086 }, { "epoch": 2.936903701819107, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7365872859954834, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8711657524108887, "num_tokens": 876569037.0, "step": 23087 }, { "epoch": 2.9370309120976974, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6151199340820312, "learning_rate": 1e-05, "loss": 0.4466, "mean_token_accuracy": 0.8737002611160278, "num_tokens": 876605354.0, "step": 23088 }, { "epoch": 2.937158122376288, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6543660163879395, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.8859700560569763, "num_tokens": 876644251.0, "step": 23089 }, { "epoch": 2.9372853326548785, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5444962978363037, "learning_rate": 1e-05, "loss": 0.4117, "mean_token_accuracy": 0.8882796764373779, "num_tokens": 876686558.0, "step": 23090 }, { "epoch": 2.937412542933469, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6425089836120605, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8785799741744995, "num_tokens": 876718629.0, "step": 23091 }, { "epoch": 2.9375397532120595, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.4833500385284424, "learning_rate": 1e-05, "loss": 0.4646, "mean_token_accuracy": 0.8655929565429688, "num_tokens": 876758037.0, "step": 23092 }, { "epoch": 2.93766696349065, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.651867628097534, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8796067833900452, "num_tokens": 876795201.0, "step": 23093 }, { "epoch": 2.9377941737692406, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.4083447456359863, "learning_rate": 1e-05, "loss": 0.3941, "mean_token_accuracy": 0.8902854919433594, "num_tokens": 876833465.0, "step": 23094 }, { "epoch": 2.937921384047831, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.705929756164551, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8778873682022095, "num_tokens": 876869915.0, "step": 23095 }, { "epoch": 2.9380485943264216, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3328001499176025, "learning_rate": 1e-05, "loss": 0.4104, "mean_token_accuracy": 0.8849552869796753, "num_tokens": 876906811.0, "step": 23096 }, { "epoch": 2.938175804605012, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.8195717334747314, "learning_rate": 1e-05, "loss": 0.4623, "mean_token_accuracy": 0.8701964616775513, "num_tokens": 876944247.0, "step": 23097 }, { "epoch": 2.9383030148836027, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2776219844818115, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.881540060043335, "num_tokens": 876978557.0, "step": 23098 }, { "epoch": 2.938430225162193, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.992347240447998, "learning_rate": 1e-05, "loss": 0.4914, "mean_token_accuracy": 0.8615947961807251, "num_tokens": 877021241.0, "step": 23099 }, { "epoch": 2.9385574354407837, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2729666233062744, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8722643256187439, "num_tokens": 877060777.0, "step": 23100 }, { "epoch": 2.9386846457193743, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.190842628479004, "learning_rate": 1e-05, "loss": 0.4858, "mean_token_accuracy": 0.8688427209854126, "num_tokens": 877096955.0, "step": 23101 }, { "epoch": 2.938811855997965, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.39294695854187, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8757472038269043, "num_tokens": 877130202.0, "step": 23102 }, { "epoch": 2.9389390662765553, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.292809963226318, "learning_rate": 1e-05, "loss": 0.4663, "mean_token_accuracy": 0.8711766600608826, "num_tokens": 877168777.0, "step": 23103 }, { "epoch": 2.9390662765551454, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.5801265239715576, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.8751773238182068, "num_tokens": 877209737.0, "step": 23104 }, { "epoch": 2.9391934868337364, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.229593276977539, "learning_rate": 1e-05, "loss": 0.4452, "mean_token_accuracy": 0.8795332312583923, "num_tokens": 877245826.0, "step": 23105 }, { "epoch": 2.9393206971123265, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.724076747894287, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8836732506752014, "num_tokens": 877285249.0, "step": 23106 }, { "epoch": 2.9394479073909174, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.039379119873047, "learning_rate": 1e-05, "loss": 0.3958, "mean_token_accuracy": 0.8929517269134521, "num_tokens": 877321971.0, "step": 23107 }, { "epoch": 2.9395751176695075, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7390928268432617, "learning_rate": 1e-05, "loss": 0.4832, "mean_token_accuracy": 0.8743849992752075, "num_tokens": 877361479.0, "step": 23108 }, { "epoch": 2.939702327948098, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.8461387157440186, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8775790929794312, "num_tokens": 877400166.0, "step": 23109 }, { "epoch": 2.9398295382266886, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6939988136291504, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8800866603851318, "num_tokens": 877437755.0, "step": 23110 }, { "epoch": 2.939956748505279, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7003238201141357, "learning_rate": 1e-05, "loss": 0.4846, "mean_token_accuracy": 0.8644631505012512, "num_tokens": 877471496.0, "step": 23111 }, { "epoch": 2.9400839587838696, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.653261423110962, "learning_rate": 1e-05, "loss": 0.3985, "mean_token_accuracy": 0.89194256067276, "num_tokens": 877504576.0, "step": 23112 }, { "epoch": 2.94021116906246, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5929887294769287, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8735777735710144, "num_tokens": 877542795.0, "step": 23113 }, { "epoch": 2.9403383793410507, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6002771854400635, "learning_rate": 1e-05, "loss": 0.4504, "mean_token_accuracy": 0.8731014132499695, "num_tokens": 877581731.0, "step": 23114 }, { "epoch": 2.940465589619641, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5346550941467285, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8721229434013367, "num_tokens": 877618405.0, "step": 23115 }, { "epoch": 2.9405927998982317, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.564518928527832, "learning_rate": 1e-05, "loss": 0.4475, "mean_token_accuracy": 0.879888653755188, "num_tokens": 877657792.0, "step": 23116 }, { "epoch": 2.9407200101768223, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5087270736694336, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8884830474853516, "num_tokens": 877688965.0, "step": 23117 }, { "epoch": 2.940847220455413, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.525132894515991, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.8882031440734863, "num_tokens": 877726595.0, "step": 23118 }, { "epoch": 2.9409744307340033, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.482625961303711, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8794883489608765, "num_tokens": 877771713.0, "step": 23119 }, { "epoch": 2.941101641012594, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.528320789337158, "learning_rate": 1e-05, "loss": 0.4194, "mean_token_accuracy": 0.883444607257843, "num_tokens": 877805048.0, "step": 23120 }, { "epoch": 2.9412288512911844, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4641363620758057, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8840281963348389, "num_tokens": 877840469.0, "step": 23121 }, { "epoch": 2.941356061569775, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.5302443504333496, "learning_rate": 1e-05, "loss": 0.4485, "mean_token_accuracy": 0.870681881904602, "num_tokens": 877876747.0, "step": 23122 }, { "epoch": 2.9414832718483654, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.443800449371338, "learning_rate": 1e-05, "loss": 0.4202, "mean_token_accuracy": 0.8815810680389404, "num_tokens": 877913871.0, "step": 23123 }, { "epoch": 2.941610482126956, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1021575927734375e-05, "grad_norm": 3.5464015007019043, "learning_rate": 1e-05, "loss": 0.5185, "mean_token_accuracy": 0.8649595379829407, "num_tokens": 877954195.0, "step": 23124 }, { "epoch": 2.9417376924055465, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.41667103767395, "learning_rate": 1e-05, "loss": 0.4658, "mean_token_accuracy": 0.8646673560142517, "num_tokens": 877989124.0, "step": 23125 }, { "epoch": 2.941864902684137, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.558664321899414, "learning_rate": 1e-05, "loss": 0.442, "mean_token_accuracy": 0.8772639036178589, "num_tokens": 878030188.0, "step": 23126 }, { "epoch": 2.941992112962727, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3562324047088623, "learning_rate": 1e-05, "loss": 0.3729, "mean_token_accuracy": 0.897564172744751, "num_tokens": 878073839.0, "step": 23127 }, { "epoch": 2.942119323241318, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6392340660095215, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8812028169631958, "num_tokens": 878111446.0, "step": 23128 }, { "epoch": 2.942246533519908, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.290583372116089, "learning_rate": 1e-05, "loss": 0.4865, "mean_token_accuracy": 0.858802318572998, "num_tokens": 878154543.0, "step": 23129 }, { "epoch": 2.942373743798499, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7641539573669434, "learning_rate": 1e-05, "loss": 0.4286, "mean_token_accuracy": 0.8855339884757996, "num_tokens": 878194265.0, "step": 23130 }, { "epoch": 2.9425009540770892, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.216285228729248, "learning_rate": 1e-05, "loss": 0.4827, "mean_token_accuracy": 0.8638114333152771, "num_tokens": 878231475.0, "step": 23131 }, { "epoch": 2.94262816435568, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9683494567871094, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8804993033409119, "num_tokens": 878269492.0, "step": 23132 }, { "epoch": 2.9427553746342703, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.1407389640808105, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8655193448066711, "num_tokens": 878307047.0, "step": 23133 }, { "epoch": 2.942882584912861, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.292645454406738, "learning_rate": 1e-05, "loss": 0.4642, "mean_token_accuracy": 0.8747284412384033, "num_tokens": 878343971.0, "step": 23134 }, { "epoch": 2.9430097951914513, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.2575957775115967, "learning_rate": 1e-05, "loss": 0.4217, "mean_token_accuracy": 0.8841026425361633, "num_tokens": 878380858.0, "step": 23135 }, { "epoch": 2.943137005470042, "ewc_loss": 0.1103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 4.6104512214660645, "learning_rate": 1e-05, "loss": 0.4487, "mean_token_accuracy": 0.8849303722381592, "num_tokens": 878420854.0, "step": 23136 }, { "epoch": 2.9432642157486324, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 3.6889209747314453, "learning_rate": 1e-05, "loss": 0.482, "mean_token_accuracy": 0.8675220012664795, "num_tokens": 878455465.0, "step": 23137 }, { "epoch": 2.943391426027223, "ewc_loss": 0.109375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.594200611114502, "learning_rate": 1e-05, "loss": 0.4284, "mean_token_accuracy": 0.8893743753433228, "num_tokens": 878487122.0, "step": 23138 }, { "epoch": 2.9435186363058135, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.021411895751953, "learning_rate": 1e-05, "loss": 0.4993, "mean_token_accuracy": 0.863987922668457, "num_tokens": 878523881.0, "step": 23139 }, { "epoch": 2.943645846584404, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.2996931076049805, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8682156801223755, "num_tokens": 878564351.0, "step": 23140 }, { "epoch": 2.9437730568629945, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 3.9648969173431396, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.8890318870544434, "num_tokens": 878605759.0, "step": 23141 }, { "epoch": 2.943900267141585, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.1075358390808105, "learning_rate": 1e-05, "loss": 0.4223, "mean_token_accuracy": 0.8869376182556152, "num_tokens": 878644975.0, "step": 23142 }, { "epoch": 2.9440274774201756, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.8334193229675293, "learning_rate": 1e-05, "loss": 0.4294, "mean_token_accuracy": 0.8851087689399719, "num_tokens": 878679882.0, "step": 23143 }, { "epoch": 2.944154687698766, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.940547466278076, "learning_rate": 1e-05, "loss": 0.3976, "mean_token_accuracy": 0.8925023078918457, "num_tokens": 878723539.0, "step": 23144 }, { "epoch": 2.9442818979773566, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.684347152709961, "learning_rate": 1e-05, "loss": 0.4114, "mean_token_accuracy": 0.8861113786697388, "num_tokens": 878761993.0, "step": 23145 }, { "epoch": 2.944409108255947, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.8867995738983154, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8782550096511841, "num_tokens": 878794355.0, "step": 23146 }, { "epoch": 2.9445363185345377, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5739219188690186, "learning_rate": 1e-05, "loss": 0.4953, "mean_token_accuracy": 0.8621850609779358, "num_tokens": 878828477.0, "step": 23147 }, { "epoch": 2.944663528813128, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.886497735977173, "learning_rate": 1e-05, "loss": 0.4692, "mean_token_accuracy": 0.8673298358917236, "num_tokens": 878868134.0, "step": 23148 }, { "epoch": 2.9447907390917187, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.4497077465057373, "learning_rate": 1e-05, "loss": 0.4266, "mean_token_accuracy": 0.8821792602539062, "num_tokens": 878908912.0, "step": 23149 }, { "epoch": 2.9449179493703093, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.971902847290039, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8789281845092773, "num_tokens": 878947354.0, "step": 23150 }, { "epoch": 2.9450451596489, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.3804380893707275, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8835462331771851, "num_tokens": 878985538.0, "step": 23151 }, { "epoch": 2.94517236992749, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.119792461395264, "learning_rate": 1e-05, "loss": 0.4722, "mean_token_accuracy": 0.8697687983512878, "num_tokens": 879028884.0, "step": 23152 }, { "epoch": 2.945299580206081, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.416797637939453, "learning_rate": 1e-05, "loss": 0.416, "mean_token_accuracy": 0.8815748691558838, "num_tokens": 879064388.0, "step": 23153 }, { "epoch": 2.945426790484671, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2555742263793945, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8843114376068115, "num_tokens": 879100353.0, "step": 23154 }, { "epoch": 2.945554000763262, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.573194742202759, "learning_rate": 1e-05, "loss": 0.4204, "mean_token_accuracy": 0.8849775195121765, "num_tokens": 879138447.0, "step": 23155 }, { "epoch": 2.945681211041852, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.242099285125732, "learning_rate": 1e-05, "loss": 0.4947, "mean_token_accuracy": 0.8650038242340088, "num_tokens": 879172825.0, "step": 23156 }, { "epoch": 2.945808421320443, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7520737648010254, "learning_rate": 1e-05, "loss": 0.4543, "mean_token_accuracy": 0.8727656602859497, "num_tokens": 879212086.0, "step": 23157 }, { "epoch": 2.945935631599033, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.030071258544922, "learning_rate": 1e-05, "loss": 0.4375, "mean_token_accuracy": 0.8807436227798462, "num_tokens": 879247689.0, "step": 23158 }, { "epoch": 2.9460628418776236, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.785146713256836, "learning_rate": 1e-05, "loss": 0.4188, "mean_token_accuracy": 0.8879384994506836, "num_tokens": 879283510.0, "step": 23159 }, { "epoch": 2.946190052156214, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7870397567749023, "learning_rate": 1e-05, "loss": 0.4433, "mean_token_accuracy": 0.8771712183952332, "num_tokens": 879322037.0, "step": 23160 }, { "epoch": 2.9463172624348046, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.751577854156494, "learning_rate": 1e-05, "loss": 0.4645, "mean_token_accuracy": 0.8682225346565247, "num_tokens": 879366400.0, "step": 23161 }, { "epoch": 2.946444472713395, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.620739698410034, "learning_rate": 1e-05, "loss": 0.5057, "mean_token_accuracy": 0.8579474687576294, "num_tokens": 879397768.0, "step": 23162 }, { "epoch": 2.9465716829919857, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.7247490882873535, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8779304027557373, "num_tokens": 879436751.0, "step": 23163 }, { "epoch": 2.946698893270576, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.466978073120117, "learning_rate": 1e-05, "loss": 0.4859, "mean_token_accuracy": 0.8629298210144043, "num_tokens": 879477750.0, "step": 23164 }, { "epoch": 2.9468261035491667, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.73309063911438, "learning_rate": 1e-05, "loss": 0.4415, "mean_token_accuracy": 0.8758286833763123, "num_tokens": 879516049.0, "step": 23165 }, { "epoch": 2.9469533138277573, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3826498985290527, "learning_rate": 1e-05, "loss": 0.474, "mean_token_accuracy": 0.8641234040260315, "num_tokens": 879553176.0, "step": 23166 }, { "epoch": 2.947080524106348, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7920010089874268, "learning_rate": 1e-05, "loss": 0.3928, "mean_token_accuracy": 0.8914831876754761, "num_tokens": 879588435.0, "step": 23167 }, { "epoch": 2.9472077343849383, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3058650493621826, "learning_rate": 1e-05, "loss": 0.394, "mean_token_accuracy": 0.8916559219360352, "num_tokens": 879623232.0, "step": 23168 }, { "epoch": 2.947334944663529, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.903756856918335, "learning_rate": 1e-05, "loss": 0.4763, "mean_token_accuracy": 0.8684551119804382, "num_tokens": 879669470.0, "step": 23169 }, { "epoch": 2.9474621549421194, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.258848190307617, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8890878558158875, "num_tokens": 879708173.0, "step": 23170 }, { "epoch": 2.94758936522071, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.086897850036621, "learning_rate": 1e-05, "loss": 0.4785, "mean_token_accuracy": 0.8717594146728516, "num_tokens": 879749086.0, "step": 23171 }, { "epoch": 2.9477165754993004, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.3021633625030518, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.879438579082489, "num_tokens": 879786521.0, "step": 23172 }, { "epoch": 2.947843785777891, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.276519298553467, "learning_rate": 1e-05, "loss": 0.4315, "mean_token_accuracy": 0.8866232633590698, "num_tokens": 879820332.0, "step": 23173 }, { "epoch": 2.9479709960564815, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.471266746520996, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8761056661605835, "num_tokens": 879859224.0, "step": 23174 }, { "epoch": 2.948098206335072, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.333951473236084, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8830921649932861, "num_tokens": 879892177.0, "step": 23175 }, { "epoch": 2.9482254166136626, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.69356369972229, "learning_rate": 1e-05, "loss": 0.3942, "mean_token_accuracy": 0.8973467946052551, "num_tokens": 879924195.0, "step": 23176 }, { "epoch": 2.9483526268922526, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.196433067321777, "learning_rate": 1e-05, "loss": 0.4581, "mean_token_accuracy": 0.8759140968322754, "num_tokens": 879965679.0, "step": 23177 }, { "epoch": 2.9484798371708436, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.76983642578125, "learning_rate": 1e-05, "loss": 0.4225, "mean_token_accuracy": 0.8831372261047363, "num_tokens": 880006441.0, "step": 23178 }, { "epoch": 2.9486070474494337, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.010136127471924, "learning_rate": 1e-05, "loss": 0.4826, "mean_token_accuracy": 0.8670421838760376, "num_tokens": 880042823.0, "step": 23179 }, { "epoch": 2.9487342577280247, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7590456008911133, "learning_rate": 1e-05, "loss": 0.4384, "mean_token_accuracy": 0.8776907920837402, "num_tokens": 880073949.0, "step": 23180 }, { "epoch": 2.9488614680066147, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8090929985046387, "learning_rate": 1e-05, "loss": 0.3876, "mean_token_accuracy": 0.8970202207565308, "num_tokens": 880112921.0, "step": 23181 }, { "epoch": 2.9489886782852053, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.683741569519043, "learning_rate": 1e-05, "loss": 0.4686, "mean_token_accuracy": 0.8700045347213745, "num_tokens": 880152384.0, "step": 23182 }, { "epoch": 2.949115888563796, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.704686164855957, "learning_rate": 1e-05, "loss": 0.4514, "mean_token_accuracy": 0.878318190574646, "num_tokens": 880187107.0, "step": 23183 }, { "epoch": 2.9492430988423863, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.5972816944122314, "learning_rate": 1e-05, "loss": 0.4276, "mean_token_accuracy": 0.8807407021522522, "num_tokens": 880226072.0, "step": 23184 }, { "epoch": 2.949370309120977, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6431162357330322, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8806248903274536, "num_tokens": 880259607.0, "step": 23185 }, { "epoch": 2.9494975193995674, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.221366882324219e-05, "grad_norm": 3.5249109268188477, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8707402944564819, "num_tokens": 880301238.0, "step": 23186 }, { "epoch": 2.949624729678158, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.640528678894043, "learning_rate": 1e-05, "loss": 0.451, "mean_token_accuracy": 0.8711583614349365, "num_tokens": 880336449.0, "step": 23187 }, { "epoch": 2.9497519399567484, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4627599716186523, "learning_rate": 1e-05, "loss": 0.4484, "mean_token_accuracy": 0.8726962804794312, "num_tokens": 880373527.0, "step": 23188 }, { "epoch": 2.949879150235339, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6610965728759766, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8789262771606445, "num_tokens": 880412676.0, "step": 23189 }, { "epoch": 2.9500063605139295, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3855996131896973, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8814505338668823, "num_tokens": 880448171.0, "step": 23190 }, { "epoch": 2.95013357079252, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.743082046508789, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8754457235336304, "num_tokens": 880487056.0, "step": 23191 }, { "epoch": 2.9502607810711106, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3170394897460938, "learning_rate": 1e-05, "loss": 0.376, "mean_token_accuracy": 0.8972076177597046, "num_tokens": 880520990.0, "step": 23192 }, { "epoch": 2.950387991349701, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8734939098358154, "learning_rate": 1e-05, "loss": 0.475, "mean_token_accuracy": 0.8662710189819336, "num_tokens": 880557919.0, "step": 23193 }, { "epoch": 2.9505152016282916, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.268233060836792, "learning_rate": 1e-05, "loss": 0.3996, "mean_token_accuracy": 0.8883557915687561, "num_tokens": 880590578.0, "step": 23194 }, { "epoch": 2.950642411906882, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.054189205169678, "learning_rate": 1e-05, "loss": 0.4343, "mean_token_accuracy": 0.8805767297744751, "num_tokens": 880631607.0, "step": 23195 }, { "epoch": 2.9507696221854727, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.275176763534546, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8821890354156494, "num_tokens": 880674301.0, "step": 23196 }, { "epoch": 2.950896832464063, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.258090972900391, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8759647011756897, "num_tokens": 880714254.0, "step": 23197 }, { "epoch": 2.9510240427426537, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.4425904750823975, "learning_rate": 1e-05, "loss": 0.3789, "mean_token_accuracy": 0.8980880379676819, "num_tokens": 880751562.0, "step": 23198 }, { "epoch": 2.9511512530212443, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.352694511413574, "learning_rate": 1e-05, "loss": 0.4847, "mean_token_accuracy": 0.8666386008262634, "num_tokens": 880785062.0, "step": 23199 }, { "epoch": 2.951278463299835, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.692387104034424, "learning_rate": 1e-05, "loss": 0.4564, "mean_token_accuracy": 0.8744823932647705, "num_tokens": 880820673.0, "step": 23200 }, { "epoch": 2.9514056735784253, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.2086944580078125, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8738753795623779, "num_tokens": 880860707.0, "step": 23201 }, { "epoch": 2.9515328838570154, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.772836208343506, "learning_rate": 1e-05, "loss": 0.4312, "mean_token_accuracy": 0.8818987607955933, "num_tokens": 880903504.0, "step": 23202 }, { "epoch": 2.9516600941356064, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.997138261795044, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8908397555351257, "num_tokens": 880943414.0, "step": 23203 }, { "epoch": 2.9517873044141965, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.7631444931030273, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8857001066207886, "num_tokens": 880977736.0, "step": 23204 }, { "epoch": 2.9519145146927874, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.783463954925537, "learning_rate": 1e-05, "loss": 0.3928, "mean_token_accuracy": 0.8935838937759399, "num_tokens": 881019793.0, "step": 23205 }, { "epoch": 2.9520417249713775, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7041969299316406, "learning_rate": 1e-05, "loss": 0.4057, "mean_token_accuracy": 0.8903650045394897, "num_tokens": 881057703.0, "step": 23206 }, { "epoch": 2.952168935249968, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6413722038269043, "learning_rate": 1e-05, "loss": 0.4169, "mean_token_accuracy": 0.8847611546516418, "num_tokens": 881093649.0, "step": 23207 }, { "epoch": 2.9522961455285586, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6725351810455322, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.8605117201805115, "num_tokens": 881125620.0, "step": 23208 }, { "epoch": 2.952423355807149, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.536017656326294, "learning_rate": 1e-05, "loss": 0.3959, "mean_token_accuracy": 0.8914117813110352, "num_tokens": 881165313.0, "step": 23209 }, { "epoch": 2.9525505660857396, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6327567100524902, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8816406726837158, "num_tokens": 881195390.0, "step": 23210 }, { "epoch": 2.95267777636433, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.4721107482910156, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8714281916618347, "num_tokens": 881242005.0, "step": 23211 }, { "epoch": 2.9528049866429207, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.6117193698883057, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8824593424797058, "num_tokens": 881284587.0, "step": 23212 }, { "epoch": 2.952932196921511, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.420015335083008, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8835219740867615, "num_tokens": 881324174.0, "step": 23213 }, { "epoch": 2.9530594072001017, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6528494358062744, "learning_rate": 1e-05, "loss": 0.4884, "mean_token_accuracy": 0.8607566356658936, "num_tokens": 881360727.0, "step": 23214 }, { "epoch": 2.9531866174786923, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.356417417526245, "learning_rate": 1e-05, "loss": 0.4314, "mean_token_accuracy": 0.8791447281837463, "num_tokens": 881400482.0, "step": 23215 }, { "epoch": 2.953313827757283, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.7131214141845703, "learning_rate": 1e-05, "loss": 0.4368, "mean_token_accuracy": 0.8786687850952148, "num_tokens": 881440157.0, "step": 23216 }, { "epoch": 2.9534410380358733, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2693755626678467, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8863624334335327, "num_tokens": 881486814.0, "step": 23217 }, { "epoch": 2.953568248314464, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8532042503356934, "learning_rate": 1e-05, "loss": 0.458, "mean_token_accuracy": 0.8743757605552673, "num_tokens": 881526997.0, "step": 23218 }, { "epoch": 2.9536954585930544, "ewc_loss": 0.08251953125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.673004150390625e-05, "grad_norm": 3.203364849090576, "learning_rate": 1e-05, "loss": 0.4229, "mean_token_accuracy": 0.8795739412307739, "num_tokens": 881569103.0, "step": 23219 }, { "epoch": 2.953822668871645, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.089686393737793, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8834497332572937, "num_tokens": 881604734.0, "step": 23220 }, { "epoch": 2.9539498791502354, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2166712284088135, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8812234997749329, "num_tokens": 881645788.0, "step": 23221 }, { "epoch": 2.954077089428826, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.3780107498168945, "learning_rate": 1e-05, "loss": 0.409, "mean_token_accuracy": 0.8898152112960815, "num_tokens": 881674738.0, "step": 23222 }, { "epoch": 2.9542042997074165, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.4579710960388184, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8786523938179016, "num_tokens": 881711685.0, "step": 23223 }, { "epoch": 2.954331509986007, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 7.05718994140625e-05, "grad_norm": 4.474823951721191, "learning_rate": 1e-05, "loss": 0.4073, "mean_token_accuracy": 0.8932642936706543, "num_tokens": 881748165.0, "step": 23224 }, { "epoch": 2.954458720264597, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.764540672302246, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.884843647480011, "num_tokens": 881781332.0, "step": 23225 }, { "epoch": 2.954585930543188, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.3348493576049805, "learning_rate": 1e-05, "loss": 0.4454, "mean_token_accuracy": 0.8809750080108643, "num_tokens": 881814526.0, "step": 23226 }, { "epoch": 2.954713140821778, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.876239776611328, "learning_rate": 1e-05, "loss": 0.487, "mean_token_accuracy": 0.8672550320625305, "num_tokens": 881856625.0, "step": 23227 }, { "epoch": 2.954840351100369, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.104933738708496, "learning_rate": 1e-05, "loss": 0.4333, "mean_token_accuracy": 0.8852903246879578, "num_tokens": 881895092.0, "step": 23228 }, { "epoch": 2.954967561378959, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.809363603591919, "learning_rate": 1e-05, "loss": 0.4405, "mean_token_accuracy": 0.8805639743804932, "num_tokens": 881936789.0, "step": 23229 }, { "epoch": 2.95509477165755, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9366772174835205, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8877930641174316, "num_tokens": 881973471.0, "step": 23230 }, { "epoch": 2.9552219819361403, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.718823194503784, "learning_rate": 1e-05, "loss": 0.4531, "mean_token_accuracy": 0.8734015226364136, "num_tokens": 882011039.0, "step": 23231 }, { "epoch": 2.955349192214731, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8073341846466064, "learning_rate": 1e-05, "loss": 0.432, "mean_token_accuracy": 0.8794086575508118, "num_tokens": 882049221.0, "step": 23232 }, { "epoch": 2.9554764024933213, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6118342876434326, "learning_rate": 1e-05, "loss": 0.417, "mean_token_accuracy": 0.8840653896331787, "num_tokens": 882086527.0, "step": 23233 }, { "epoch": 2.955603612771912, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7529184818267822, "learning_rate": 1e-05, "loss": 0.371, "mean_token_accuracy": 0.8994925022125244, "num_tokens": 882118375.0, "step": 23234 }, { "epoch": 2.9557308230505024, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.539114236831665, "learning_rate": 1e-05, "loss": 0.4395, "mean_token_accuracy": 0.8770511150360107, "num_tokens": 882150604.0, "step": 23235 }, { "epoch": 2.955858033329093, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.734697103500366, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8787182569503784, "num_tokens": 882185822.0, "step": 23236 }, { "epoch": 2.9559852436076834, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4346675872802734, "learning_rate": 1e-05, "loss": 0.4279, "mean_token_accuracy": 0.8817380666732788, "num_tokens": 882224829.0, "step": 23237 }, { "epoch": 2.956112453886274, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.789963483810425, "learning_rate": 1e-05, "loss": 0.4807, "mean_token_accuracy": 0.8615537285804749, "num_tokens": 882263107.0, "step": 23238 }, { "epoch": 2.9562396641648645, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3743796348571777, "learning_rate": 1e-05, "loss": 0.4292, "mean_token_accuracy": 0.8787714242935181, "num_tokens": 882294302.0, "step": 23239 }, { "epoch": 2.956366874443455, "ewc_loss": 0.09375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.8858346939086914, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8830455541610718, "num_tokens": 882327081.0, "step": 23240 }, { "epoch": 2.9564940847220456, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.316758155822754, "learning_rate": 1e-05, "loss": 0.4164, "mean_token_accuracy": 0.8834617733955383, "num_tokens": 882361793.0, "step": 23241 }, { "epoch": 2.956621295000636, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.019155025482178, "learning_rate": 1e-05, "loss": 0.3876, "mean_token_accuracy": 0.895514190196991, "num_tokens": 882395255.0, "step": 23242 }, { "epoch": 2.9567485052792266, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3333263397216797, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8802784085273743, "num_tokens": 882431915.0, "step": 23243 }, { "epoch": 2.956875715557817, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.177276611328125, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8760398626327515, "num_tokens": 882472688.0, "step": 23244 }, { "epoch": 2.9570029258364077, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.4554052352905273, "learning_rate": 1e-05, "loss": 0.427, "mean_token_accuracy": 0.8806207180023193, "num_tokens": 882506851.0, "step": 23245 }, { "epoch": 2.957130136114998, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.242587566375732, "learning_rate": 1e-05, "loss": 0.4866, "mean_token_accuracy": 0.8717648983001709, "num_tokens": 882541935.0, "step": 23246 }, { "epoch": 2.9572573463935887, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.625516176223755, "learning_rate": 1e-05, "loss": 0.4172, "mean_token_accuracy": 0.8844594955444336, "num_tokens": 882577860.0, "step": 23247 }, { "epoch": 2.9573845566721793, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.122325897216797, "learning_rate": 1e-05, "loss": 0.4955, "mean_token_accuracy": 0.8642435669898987, "num_tokens": 882616182.0, "step": 23248 }, { "epoch": 2.95751176695077, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.726743459701538, "learning_rate": 1e-05, "loss": 0.5008, "mean_token_accuracy": 0.8598709106445312, "num_tokens": 882658847.0, "step": 23249 }, { "epoch": 2.95763897722936, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.89963960647583, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.8759253025054932, "num_tokens": 882699842.0, "step": 23250 }, { "epoch": 2.957766187507951, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7288219928741455, "learning_rate": 1e-05, "loss": 0.4122, "mean_token_accuracy": 0.885987401008606, "num_tokens": 882733143.0, "step": 23251 }, { "epoch": 2.957893397786541, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7096433639526367, "learning_rate": 1e-05, "loss": 0.4436, "mean_token_accuracy": 0.8739162683486938, "num_tokens": 882763233.0, "step": 23252 }, { "epoch": 2.958020608065132, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.689112424850464, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8801498413085938, "num_tokens": 882807612.0, "step": 23253 }, { "epoch": 2.958147818343722, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.5441765785217285, "learning_rate": 1e-05, "loss": 0.3686, "mean_token_accuracy": 0.9002059698104858, "num_tokens": 882842135.0, "step": 23254 }, { "epoch": 2.958275028622313, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6848270893096924, "learning_rate": 1e-05, "loss": 0.4198, "mean_token_accuracy": 0.8836712837219238, "num_tokens": 882876687.0, "step": 23255 }, { "epoch": 2.958402238900903, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.437331199645996, "learning_rate": 1e-05, "loss": 0.4423, "mean_token_accuracy": 0.8746457099914551, "num_tokens": 882919142.0, "step": 23256 }, { "epoch": 2.9585294491794936, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.715604782104492, "learning_rate": 1e-05, "loss": 0.4168, "mean_token_accuracy": 0.8873059749603271, "num_tokens": 882955301.0, "step": 23257 }, { "epoch": 2.958656659458084, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3665428161621094, "learning_rate": 1e-05, "loss": 0.3882, "mean_token_accuracy": 0.8941939473152161, "num_tokens": 882985436.0, "step": 23258 }, { "epoch": 2.9587838697366746, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.773729085922241, "learning_rate": 1e-05, "loss": 0.4595, "mean_token_accuracy": 0.8747228384017944, "num_tokens": 883028082.0, "step": 23259 }, { "epoch": 2.958911080015265, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3196496963500977, "learning_rate": 1e-05, "loss": 0.436, "mean_token_accuracy": 0.8775640726089478, "num_tokens": 883063771.0, "step": 23260 }, { "epoch": 2.9590382902938557, "ewc_loss": 0.09375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8764843940734863, "learning_rate": 1e-05, "loss": 0.4191, "mean_token_accuracy": 0.8821693658828735, "num_tokens": 883096742.0, "step": 23261 }, { "epoch": 2.959165500572446, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2618017196655273, "learning_rate": 1e-05, "loss": 0.4447, "mean_token_accuracy": 0.8749419450759888, "num_tokens": 883137916.0, "step": 23262 }, { "epoch": 2.9592927108510367, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.039809703826904, "learning_rate": 1e-05, "loss": 0.4586, "mean_token_accuracy": 0.8739302158355713, "num_tokens": 883172730.0, "step": 23263 }, { "epoch": 2.9594199211296273, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.2766807079315186, "learning_rate": 1e-05, "loss": 0.4392, "mean_token_accuracy": 0.8774418830871582, "num_tokens": 883212139.0, "step": 23264 }, { "epoch": 2.959547131408218, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.229608058929443, "learning_rate": 1e-05, "loss": 0.4238, "mean_token_accuracy": 0.8877766132354736, "num_tokens": 883244866.0, "step": 23265 }, { "epoch": 2.9596743416868083, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.4164443016052246, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8833174705505371, "num_tokens": 883277595.0, "step": 23266 }, { "epoch": 2.959801551965399, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.331310272216797, "learning_rate": 1e-05, "loss": 0.4552, "mean_token_accuracy": 0.877504825592041, "num_tokens": 883311246.0, "step": 23267 }, { "epoch": 2.9599287622439894, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6120874881744385, "learning_rate": 1e-05, "loss": 0.449, "mean_token_accuracy": 0.877889096736908, "num_tokens": 883353351.0, "step": 23268 }, { "epoch": 2.96005597252258, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.247817039489746, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8837031126022339, "num_tokens": 883392245.0, "step": 23269 }, { "epoch": 2.9601831828011704, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7403318881988525, "learning_rate": 1e-05, "loss": 0.4214, "mean_token_accuracy": 0.8842394351959229, "num_tokens": 883430404.0, "step": 23270 }, { "epoch": 2.960310393079761, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.00035858154296875, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.071204662322998, "learning_rate": 1e-05, "loss": 0.5145, "mean_token_accuracy": 0.8591135144233704, "num_tokens": 883477894.0, "step": 23271 }, { "epoch": 2.9604376033583515, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.742171049118042, "learning_rate": 1e-05, "loss": 0.4885, "mean_token_accuracy": 0.8626890778541565, "num_tokens": 883520396.0, "step": 23272 }, { "epoch": 2.960564813636942, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.897923707962036, "learning_rate": 1e-05, "loss": 0.4549, "mean_token_accuracy": 0.8775092959403992, "num_tokens": 883559120.0, "step": 23273 }, { "epoch": 2.9606920239155325, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.6808130741119385, "learning_rate": 1e-05, "loss": 0.4539, "mean_token_accuracy": 0.8730810284614563, "num_tokens": 883603061.0, "step": 23274 }, { "epoch": 2.9608192341941226, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7663135528564453, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8761435747146606, "num_tokens": 883637674.0, "step": 23275 }, { "epoch": 2.9609464444727136, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.615750789642334, "learning_rate": 1e-05, "loss": 0.4193, "mean_token_accuracy": 0.8828305006027222, "num_tokens": 883674631.0, "step": 23276 }, { "epoch": 2.9610736547513037, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.67107892036438, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8760948181152344, "num_tokens": 883716823.0, "step": 23277 }, { "epoch": 2.9612008650298947, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.000362396240234375, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5677947998046875, "learning_rate": 1e-05, "loss": 0.4536, "mean_token_accuracy": 0.8765595555305481, "num_tokens": 883751849.0, "step": 23278 }, { "epoch": 2.9613280753084847, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6186673641204834, "learning_rate": 1e-05, "loss": 0.4053, "mean_token_accuracy": 0.888417661190033, "num_tokens": 883790698.0, "step": 23279 }, { "epoch": 2.9614552855870753, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5029642581939697, "learning_rate": 1e-05, "loss": 0.3909, "mean_token_accuracy": 0.8928940296173096, "num_tokens": 883823238.0, "step": 23280 }, { "epoch": 2.961582495865666, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.614305019378662, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8808869123458862, "num_tokens": 883863036.0, "step": 23281 }, { "epoch": 2.9617097061442563, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.4477715492248535, "learning_rate": 1e-05, "loss": 0.4513, "mean_token_accuracy": 0.8683755397796631, "num_tokens": 883902556.0, "step": 23282 }, { "epoch": 2.961836916422847, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6710920333862305, "learning_rate": 1e-05, "loss": 0.4725, "mean_token_accuracy": 0.8652831315994263, "num_tokens": 883937039.0, "step": 23283 }, { "epoch": 2.9619641267014374, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.3876748085021973, "learning_rate": 1e-05, "loss": 0.4309, "mean_token_accuracy": 0.877757728099823, "num_tokens": 883969047.0, "step": 23284 }, { "epoch": 2.962091336980028, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.7180871963500977, "learning_rate": 1e-05, "loss": 0.4559, "mean_token_accuracy": 0.8742598295211792, "num_tokens": 884013431.0, "step": 23285 }, { "epoch": 2.9622185472586184, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.3179125785827637, "learning_rate": 1e-05, "loss": 0.4354, "mean_token_accuracy": 0.8761252760887146, "num_tokens": 884053714.0, "step": 23286 }, { "epoch": 2.962345757537209, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.832923650741577, "learning_rate": 1e-05, "loss": 0.4599, "mean_token_accuracy": 0.8730036616325378, "num_tokens": 884094624.0, "step": 23287 }, { "epoch": 2.9624729678157995, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2704226970672607, "learning_rate": 1e-05, "loss": 0.4367, "mean_token_accuracy": 0.876266598701477, "num_tokens": 884131759.0, "step": 23288 }, { "epoch": 2.96260017809439, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9995763301849365, "learning_rate": 1e-05, "loss": 0.4291, "mean_token_accuracy": 0.882934033870697, "num_tokens": 884175333.0, "step": 23289 }, { "epoch": 2.9627273883729806, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2621383666992188, "learning_rate": 1e-05, "loss": 0.3886, "mean_token_accuracy": 0.8925654292106628, "num_tokens": 884209778.0, "step": 23290 }, { "epoch": 2.962854598651571, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.204416751861572, "learning_rate": 1e-05, "loss": 0.4434, "mean_token_accuracy": 0.8792039752006531, "num_tokens": 884245581.0, "step": 23291 }, { "epoch": 2.9629818089301616, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.372602939605713, "learning_rate": 1e-05, "loss": 0.4081, "mean_token_accuracy": 0.8874099254608154, "num_tokens": 884284925.0, "step": 23292 }, { "epoch": 2.963109019208752, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.33048677444458, "learning_rate": 1e-05, "loss": 0.4316, "mean_token_accuracy": 0.8849895000457764, "num_tokens": 884326056.0, "step": 23293 }, { "epoch": 2.9632362294873427, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.602198362350464, "learning_rate": 1e-05, "loss": 0.4358, "mean_token_accuracy": 0.8795635104179382, "num_tokens": 884368169.0, "step": 23294 }, { "epoch": 2.963363439765933, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.274347305297852, "learning_rate": 1e-05, "loss": 0.4605, "mean_token_accuracy": 0.8790263533592224, "num_tokens": 884402785.0, "step": 23295 }, { "epoch": 2.9634906500445237, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.7487246990203857, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8815816640853882, "num_tokens": 884444755.0, "step": 23296 }, { "epoch": 2.9636178603231143, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 4.083303928375244, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8834502696990967, "num_tokens": 884486299.0, "step": 23297 }, { "epoch": 2.963745070601705, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.759326219558716, "learning_rate": 1e-05, "loss": 0.5068, "mean_token_accuracy": 0.8581284880638123, "num_tokens": 884532076.0, "step": 23298 }, { "epoch": 2.9638722808802953, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.8942058086395264, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8804411888122559, "num_tokens": 884569518.0, "step": 23299 }, { "epoch": 2.9639994911588854, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.7141225337982178, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8820675611495972, "num_tokens": 884601173.0, "step": 23300 }, { "epoch": 2.9641267014374764, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7347445487976074, "learning_rate": 1e-05, "loss": 0.429, "mean_token_accuracy": 0.8833156228065491, "num_tokens": 884637558.0, "step": 23301 }, { "epoch": 2.9642539117160664, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.643319845199585, "learning_rate": 1e-05, "loss": 0.4319, "mean_token_accuracy": 0.8805884122848511, "num_tokens": 884673377.0, "step": 23302 }, { "epoch": 2.9643811219946574, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6298635005950928, "learning_rate": 1e-05, "loss": 0.3952, "mean_token_accuracy": 0.8940656781196594, "num_tokens": 884710334.0, "step": 23303 }, { "epoch": 2.9645083322732475, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5809009075164795, "learning_rate": 1e-05, "loss": 0.4006, "mean_token_accuracy": 0.8912071585655212, "num_tokens": 884741515.0, "step": 23304 }, { "epoch": 2.964635542551838, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5971508026123047, "learning_rate": 1e-05, "loss": 0.455, "mean_token_accuracy": 0.8718879818916321, "num_tokens": 884782204.0, "step": 23305 }, { "epoch": 2.9647627528304286, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.508178234100342, "learning_rate": 1e-05, "loss": 0.4049, "mean_token_accuracy": 0.8877599239349365, "num_tokens": 884819613.0, "step": 23306 }, { "epoch": 2.964889963109019, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5911800861358643, "learning_rate": 1e-05, "loss": 0.4425, "mean_token_accuracy": 0.8780292868614197, "num_tokens": 884857381.0, "step": 23307 }, { "epoch": 2.9650171733876096, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.462066650390625, "learning_rate": 1e-05, "loss": 0.4968, "mean_token_accuracy": 0.8561298847198486, "num_tokens": 884893562.0, "step": 23308 }, { "epoch": 2.9651443836662, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.60705304145813, "learning_rate": 1e-05, "loss": 0.4407, "mean_token_accuracy": 0.8772890567779541, "num_tokens": 884934359.0, "step": 23309 }, { "epoch": 2.9652715939447907, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.394179344177246, "learning_rate": 1e-05, "loss": 0.3974, "mean_token_accuracy": 0.8913041949272156, "num_tokens": 884974446.0, "step": 23310 }, { "epoch": 2.965398804223381, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.656958818435669, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8725612163543701, "num_tokens": 885016447.0, "step": 23311 }, { "epoch": 2.9655260145019717, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.3433587551116943, "learning_rate": 1e-05, "loss": 0.447, "mean_token_accuracy": 0.8729557394981384, "num_tokens": 885055064.0, "step": 23312 }, { "epoch": 2.9656532247805623, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.771095037460327, "learning_rate": 1e-05, "loss": 0.4324, "mean_token_accuracy": 0.8830541372299194, "num_tokens": 885092380.0, "step": 23313 }, { "epoch": 2.965780435059153, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.2483267784118652, "learning_rate": 1e-05, "loss": 0.4025, "mean_token_accuracy": 0.8866187930107117, "num_tokens": 885130717.0, "step": 23314 }, { "epoch": 2.9659076453377433, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.939338207244873, "learning_rate": 1e-05, "loss": 0.4385, "mean_token_accuracy": 0.8814588785171509, "num_tokens": 885168291.0, "step": 23315 }, { "epoch": 2.966034855616334, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2225863933563232, "learning_rate": 1e-05, "loss": 0.4872, "mean_token_accuracy": 0.8612954616546631, "num_tokens": 885203839.0, "step": 23316 }, { "epoch": 2.9661620658949244, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.184157371520996, "learning_rate": 1e-05, "loss": 0.4671, "mean_token_accuracy": 0.8689698576927185, "num_tokens": 885239336.0, "step": 23317 }, { "epoch": 2.966289276173515, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.2955069541931152, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8605122566223145, "num_tokens": 885275964.0, "step": 23318 }, { "epoch": 2.9664164864521054, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.396512508392334, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8779492378234863, "num_tokens": 885307444.0, "step": 23319 }, { "epoch": 2.966543696730696, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.5482850074768066, "learning_rate": 1e-05, "loss": 0.4116, "mean_token_accuracy": 0.8899673819541931, "num_tokens": 885345379.0, "step": 23320 }, { "epoch": 2.9666709070092865, "ewc_loss": 0.10595703125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.406829357147217, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8849873542785645, "num_tokens": 885384105.0, "step": 23321 }, { "epoch": 2.966798117287877, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.797736644744873, "learning_rate": 1e-05, "loss": 0.4477, "mean_token_accuracy": 0.878145694732666, "num_tokens": 885420374.0, "step": 23322 }, { "epoch": 2.966925327566467, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.229343414306641, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8826196193695068, "num_tokens": 885453450.0, "step": 23323 }, { "epoch": 2.967052537845058, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.8480184078216553, "learning_rate": 1e-05, "loss": 0.4463, "mean_token_accuracy": 0.8835529088973999, "num_tokens": 885486331.0, "step": 23324 }, { "epoch": 2.967179748123648, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.01125955581665, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.891461968421936, "num_tokens": 885520474.0, "step": 23325 }, { "epoch": 2.967306958402239, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7588119506835938, "learning_rate": 1e-05, "loss": 0.4606, "mean_token_accuracy": 0.8753275275230408, "num_tokens": 885563128.0, "step": 23326 }, { "epoch": 2.967434168680829, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.843203067779541, "learning_rate": 1e-05, "loss": 0.3846, "mean_token_accuracy": 0.898537814617157, "num_tokens": 885604219.0, "step": 23327 }, { "epoch": 2.96756137895942, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.6678848266601562, "learning_rate": 1e-05, "loss": 0.428, "mean_token_accuracy": 0.8850956559181213, "num_tokens": 885639932.0, "step": 23328 }, { "epoch": 2.9676885892380103, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.759777545928955, "learning_rate": 1e-05, "loss": 0.4724, "mean_token_accuracy": 0.865769624710083, "num_tokens": 885675914.0, "step": 23329 }, { "epoch": 2.967815799516601, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.5697250366210938, "learning_rate": 1e-05, "loss": 0.4162, "mean_token_accuracy": 0.8861907720565796, "num_tokens": 885711910.0, "step": 23330 }, { "epoch": 2.9679430097951913, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.717427968978882, "learning_rate": 1e-05, "loss": 0.4052, "mean_token_accuracy": 0.8878554105758667, "num_tokens": 885751615.0, "step": 23331 }, { "epoch": 2.968070220073782, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4722583293914795, "learning_rate": 1e-05, "loss": 0.4166, "mean_token_accuracy": 0.8842699527740479, "num_tokens": 885788394.0, "step": 23332 }, { "epoch": 2.9681974303523724, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.741353750228882, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8897435069084167, "num_tokens": 885820462.0, "step": 23333 }, { "epoch": 2.968324640630963, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4025373458862305, "learning_rate": 1e-05, "loss": 0.4101, "mean_token_accuracy": 0.8888131976127625, "num_tokens": 885853714.0, "step": 23334 }, { "epoch": 2.9684518509095534, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.811269521713257, "learning_rate": 1e-05, "loss": 0.4347, "mean_token_accuracy": 0.8802757859230042, "num_tokens": 885887508.0, "step": 23335 }, { "epoch": 2.968579061188144, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.3342106342315674, "learning_rate": 1e-05, "loss": 0.4391, "mean_token_accuracy": 0.8747998476028442, "num_tokens": 885923314.0, "step": 23336 }, { "epoch": 2.9687062714667345, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.9309983253479004, "learning_rate": 1e-05, "loss": 0.4079, "mean_token_accuracy": 0.8905990123748779, "num_tokens": 885964466.0, "step": 23337 }, { "epoch": 2.968833481745325, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.982948303222656e-05, "grad_norm": 3.298163890838623, "learning_rate": 1e-05, "loss": 0.4323, "mean_token_accuracy": 0.8777273893356323, "num_tokens": 886001483.0, "step": 23338 }, { "epoch": 2.9689606920239155, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.106889724731445, "learning_rate": 1e-05, "loss": 0.4428, "mean_token_accuracy": 0.8815270662307739, "num_tokens": 886040130.0, "step": 23339 }, { "epoch": 2.969087902302506, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.3696022033691406, "learning_rate": 1e-05, "loss": 0.4322, "mean_token_accuracy": 0.8800342082977295, "num_tokens": 886076291.0, "step": 23340 }, { "epoch": 2.9692151125810966, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.229934215545654, "learning_rate": 1e-05, "loss": 0.4959, "mean_token_accuracy": 0.866946816444397, "num_tokens": 886109430.0, "step": 23341 }, { "epoch": 2.969342322859687, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.526528835296631, "learning_rate": 1e-05, "loss": 0.4293, "mean_token_accuracy": 0.8830915689468384, "num_tokens": 886147158.0, "step": 23342 }, { "epoch": 2.9694695331382777, "ewc_loss": 0.10205078125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.580352783203125e-05, "grad_norm": 4.193185329437256, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8804939985275269, "num_tokens": 886184810.0, "step": 23343 }, { "epoch": 2.969596743416868, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.6641697883605957, "learning_rate": 1e-05, "loss": 0.4281, "mean_token_accuracy": 0.8823322653770447, "num_tokens": 886225511.0, "step": 23344 }, { "epoch": 2.9697239536954587, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 4.040143966674805, "learning_rate": 1e-05, "loss": 0.4255, "mean_token_accuracy": 0.8818315863609314, "num_tokens": 886265185.0, "step": 23345 }, { "epoch": 2.9698511639740492, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.71451473236084, "learning_rate": 1e-05, "loss": 0.4479, "mean_token_accuracy": 0.8784874677658081, "num_tokens": 886304897.0, "step": 23346 }, { "epoch": 2.9699783742526398, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.888938903808594e-05, "grad_norm": 3.8448326587677, "learning_rate": 1e-05, "loss": 0.4661, "mean_token_accuracy": 0.8689736127853394, "num_tokens": 886343097.0, "step": 23347 }, { "epoch": 2.97010558453123, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.699995279312134, "learning_rate": 1e-05, "loss": 0.4453, "mean_token_accuracy": 0.8765971660614014, "num_tokens": 886377428.0, "step": 23348 }, { "epoch": 2.970232794809821, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.668196201324463, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8733040690422058, "num_tokens": 886418344.0, "step": 23349 }, { "epoch": 2.970360005088411, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.677905559539795, "learning_rate": 1e-05, "loss": 0.5111, "mean_token_accuracy": 0.8609078526496887, "num_tokens": 886458983.0, "step": 23350 }, { "epoch": 2.970487215367002, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5353317260742188, "learning_rate": 1e-05, "loss": 0.3868, "mean_token_accuracy": 0.892772912979126, "num_tokens": 886492174.0, "step": 23351 }, { "epoch": 2.970614425645592, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.652937412261963, "learning_rate": 1e-05, "loss": 0.4729, "mean_token_accuracy": 0.8676109313964844, "num_tokens": 886537145.0, "step": 23352 }, { "epoch": 2.970741635924183, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.466611623764038, "learning_rate": 1e-05, "loss": 0.4795, "mean_token_accuracy": 0.8687010407447815, "num_tokens": 886573420.0, "step": 23353 }, { "epoch": 2.970868846202773, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.656968593597412, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.880605161190033, "num_tokens": 886610489.0, "step": 23354 }, { "epoch": 2.9709960564813636, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3943285942077637, "learning_rate": 1e-05, "loss": 0.3953, "mean_token_accuracy": 0.8884679079055786, "num_tokens": 886646954.0, "step": 23355 }, { "epoch": 2.971123266759954, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.684478759765625, "learning_rate": 1e-05, "loss": 0.3999, "mean_token_accuracy": 0.8901487588882446, "num_tokens": 886686979.0, "step": 23356 }, { "epoch": 2.9712504770385446, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.372861862182617, "learning_rate": 1e-05, "loss": 0.5456, "mean_token_accuracy": 0.8510230779647827, "num_tokens": 886719734.0, "step": 23357 }, { "epoch": 2.971377687317135, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.739694595336914, "learning_rate": 1e-05, "loss": 0.4206, "mean_token_accuracy": 0.8826945424079895, "num_tokens": 886754762.0, "step": 23358 }, { "epoch": 2.9715048975957257, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.292513608932495, "learning_rate": 1e-05, "loss": 0.402, "mean_token_accuracy": 0.8851888179779053, "num_tokens": 886794830.0, "step": 23359 }, { "epoch": 2.971632107874316, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.856173515319824, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.8734909296035767, "num_tokens": 886834952.0, "step": 23360 }, { "epoch": 2.9717593181529067, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.243861436843872, "learning_rate": 1e-05, "loss": 0.4657, "mean_token_accuracy": 0.8655910491943359, "num_tokens": 886874236.0, "step": 23361 }, { "epoch": 2.9718865284314973, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.050424575805664, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8687950372695923, "num_tokens": 886907723.0, "step": 23362 }, { "epoch": 2.972013738710088, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.262282371520996, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.875442624092102, "num_tokens": 886939914.0, "step": 23363 }, { "epoch": 2.9721409489886783, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.269022464752197, "learning_rate": 1e-05, "loss": 0.4805, "mean_token_accuracy": 0.868982195854187, "num_tokens": 886970533.0, "step": 23364 }, { "epoch": 2.972268159267269, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.4043197631835938, "learning_rate": 1e-05, "loss": 0.4782, "mean_token_accuracy": 0.8646079897880554, "num_tokens": 887012103.0, "step": 23365 }, { "epoch": 2.9723953695458594, "ewc_loss": 0.10546875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.914138793945312e-05, "grad_norm": 4.37791109085083, "learning_rate": 1e-05, "loss": 0.4873, "mean_token_accuracy": 0.866747260093689, "num_tokens": 887055789.0, "step": 23366 }, { "epoch": 2.97252257982445, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.6250972747802734, "learning_rate": 1e-05, "loss": 0.4105, "mean_token_accuracy": 0.8878533840179443, "num_tokens": 887094505.0, "step": 23367 }, { "epoch": 2.9726497901030404, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.314319610595703, "learning_rate": 1e-05, "loss": 0.4215, "mean_token_accuracy": 0.8892425298690796, "num_tokens": 887131009.0, "step": 23368 }, { "epoch": 2.972777000381631, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.776235342025757, "learning_rate": 1e-05, "loss": 0.3963, "mean_token_accuracy": 0.892356276512146, "num_tokens": 887170969.0, "step": 23369 }, { "epoch": 2.9729042106602215, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.134467601776123, "learning_rate": 1e-05, "loss": 0.4145, "mean_token_accuracy": 0.8871872425079346, "num_tokens": 887205123.0, "step": 23370 }, { "epoch": 2.973031420938812, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.800978183746338, "learning_rate": 1e-05, "loss": 0.4089, "mean_token_accuracy": 0.8898611068725586, "num_tokens": 887235908.0, "step": 23371 }, { "epoch": 2.9731586312174025, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.921004056930542, "learning_rate": 1e-05, "loss": 0.4148, "mean_token_accuracy": 0.8869174718856812, "num_tokens": 887273929.0, "step": 23372 }, { "epoch": 2.9732858414959926, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.7251195907592773, "learning_rate": 1e-05, "loss": 0.4254, "mean_token_accuracy": 0.8833085298538208, "num_tokens": 887309379.0, "step": 23373 }, { "epoch": 2.9734130517745836, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.7800114154815674, "learning_rate": 1e-05, "loss": 0.4817, "mean_token_accuracy": 0.8711004257202148, "num_tokens": 887346108.0, "step": 23374 }, { "epoch": 2.9735402620531737, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.646242380142212, "learning_rate": 1e-05, "loss": 0.4059, "mean_token_accuracy": 0.8920443654060364, "num_tokens": 887380066.0, "step": 23375 }, { "epoch": 2.9736674723317646, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6881656646728516, "learning_rate": 1e-05, "loss": 0.4446, "mean_token_accuracy": 0.876645028591156, "num_tokens": 887414043.0, "step": 23376 }, { "epoch": 2.9737946826103547, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.569032907485962, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8758202791213989, "num_tokens": 887451921.0, "step": 23377 }, { "epoch": 2.9739218928889453, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6613035202026367, "learning_rate": 1e-05, "loss": 0.4448, "mean_token_accuracy": 0.8806829452514648, "num_tokens": 887483977.0, "step": 23378 }, { "epoch": 2.974049103167536, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4911344051361084, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8823373913764954, "num_tokens": 887516795.0, "step": 23379 }, { "epoch": 2.9741763134461263, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.657257556915283, "learning_rate": 1e-05, "loss": 0.4424, "mean_token_accuracy": 0.8749731779098511, "num_tokens": 887554791.0, "step": 23380 }, { "epoch": 2.974303523724717, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.410224437713623, "learning_rate": 1e-05, "loss": 0.423, "mean_token_accuracy": 0.8846078515052795, "num_tokens": 887598451.0, "step": 23381 }, { "epoch": 2.9744307340033074, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.717010974884033, "learning_rate": 1e-05, "loss": 0.3939, "mean_token_accuracy": 0.8940610289573669, "num_tokens": 887633242.0, "step": 23382 }, { "epoch": 2.974557944281898, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.350353479385376, "learning_rate": 1e-05, "loss": 0.4035, "mean_token_accuracy": 0.8853932619094849, "num_tokens": 887665084.0, "step": 23383 }, { "epoch": 2.9746851545604884, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.820188522338867, "learning_rate": 1e-05, "loss": 0.5243, "mean_token_accuracy": 0.852701723575592, "num_tokens": 887706086.0, "step": 23384 }, { "epoch": 2.974812364839079, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.2810394763946533, "learning_rate": 1e-05, "loss": 0.4538, "mean_token_accuracy": 0.8704084157943726, "num_tokens": 887747395.0, "step": 23385 }, { "epoch": 2.9749395751176695, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.9844889640808105, "learning_rate": 1e-05, "loss": 0.4718, "mean_token_accuracy": 0.8714616298675537, "num_tokens": 887789834.0, "step": 23386 }, { "epoch": 2.97506678539626, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.2583489418029785, "learning_rate": 1e-05, "loss": 0.4843, "mean_token_accuracy": 0.8615882396697998, "num_tokens": 887831416.0, "step": 23387 }, { "epoch": 2.9751939956748505, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.191160678863525, "learning_rate": 1e-05, "loss": 0.4598, "mean_token_accuracy": 0.8690710067749023, "num_tokens": 887869781.0, "step": 23388 }, { "epoch": 2.975321205953441, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.373342514038086, "learning_rate": 1e-05, "loss": 0.4381, "mean_token_accuracy": 0.8810954093933105, "num_tokens": 887906844.0, "step": 23389 }, { "epoch": 2.9754484162320316, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.326473236083984, "learning_rate": 1e-05, "loss": 0.4762, "mean_token_accuracy": 0.8709378242492676, "num_tokens": 887939514.0, "step": 23390 }, { "epoch": 2.975575626510622, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.079673767089844e-05, "grad_norm": 3.613776922225952, "learning_rate": 1e-05, "loss": 0.4863, "mean_token_accuracy": 0.8654084205627441, "num_tokens": 887976041.0, "step": 23391 }, { "epoch": 2.9757028367892127, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.257113933563232, "learning_rate": 1e-05, "loss": 0.497, "mean_token_accuracy": 0.8661260604858398, "num_tokens": 888010007.0, "step": 23392 }, { "epoch": 2.975830047067803, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.7573790550231934, "learning_rate": 1e-05, "loss": 0.4289, "mean_token_accuracy": 0.8806845545768738, "num_tokens": 888045654.0, "step": 23393 }, { "epoch": 2.9759572573463937, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.042586803436279, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8831276297569275, "num_tokens": 888084075.0, "step": 23394 }, { "epoch": 2.9760844676249842, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7647788524627686, "learning_rate": 1e-05, "loss": 0.4551, "mean_token_accuracy": 0.8766912817955017, "num_tokens": 888124891.0, "step": 23395 }, { "epoch": 2.9762116779035748, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.827824831008911, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8851555585861206, "num_tokens": 888161078.0, "step": 23396 }, { "epoch": 2.9763388881821653, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.71476149559021, "learning_rate": 1e-05, "loss": 0.4184, "mean_token_accuracy": 0.8836376667022705, "num_tokens": 888197013.0, "step": 23397 }, { "epoch": 2.9764660984607554, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.6739258766174316, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8727074861526489, "num_tokens": 888231676.0, "step": 23398 }, { "epoch": 2.9765933087393464, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.555152893066406e-05, "grad_norm": 3.6601662635803223, "learning_rate": 1e-05, "loss": 0.4178, "mean_token_accuracy": 0.8845853805541992, "num_tokens": 888272676.0, "step": 23399 }, { "epoch": 2.9767205190179364, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.340576171875e-05, "grad_norm": 3.5604000091552734, "learning_rate": 1e-05, "loss": 0.4146, "mean_token_accuracy": 0.8831356167793274, "num_tokens": 888310064.0, "step": 23400 }, { "epoch": 2.9768477292965274, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.63126802444458, "learning_rate": 1e-05, "loss": 0.4325, "mean_token_accuracy": 0.8806375861167908, "num_tokens": 888344103.0, "step": 23401 }, { "epoch": 2.9769749395751175, "ewc_loss": 0.08837890625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.508331537246704, "learning_rate": 1e-05, "loss": 0.4568, "mean_token_accuracy": 0.8727984428405762, "num_tokens": 888379284.0, "step": 23402 }, { "epoch": 2.977102149853708, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6014480590820312, "learning_rate": 1e-05, "loss": 0.4106, "mean_token_accuracy": 0.8889978528022766, "num_tokens": 888414058.0, "step": 23403 }, { "epoch": 2.9772293601322986, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4710724353790283, "learning_rate": 1e-05, "loss": 0.4567, "mean_token_accuracy": 0.8681751489639282, "num_tokens": 888454756.0, "step": 23404 }, { "epoch": 2.977356570410889, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5885565280914307, "learning_rate": 1e-05, "loss": 0.4331, "mean_token_accuracy": 0.8800272345542908, "num_tokens": 888491995.0, "step": 23405 }, { "epoch": 2.9774837806894796, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4414217472076416, "learning_rate": 1e-05, "loss": 0.4957, "mean_token_accuracy": 0.860907793045044, "num_tokens": 888533073.0, "step": 23406 }, { "epoch": 2.97761099096807, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.583890199661255, "learning_rate": 1e-05, "loss": 0.3715, "mean_token_accuracy": 0.8989859819412231, "num_tokens": 888570392.0, "step": 23407 }, { "epoch": 2.9777382012466607, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.397946834564209, "learning_rate": 1e-05, "loss": 0.4313, "mean_token_accuracy": 0.8789815902709961, "num_tokens": 888612078.0, "step": 23408 }, { "epoch": 2.977865411525251, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6222524642944336, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8862958550453186, "num_tokens": 888653603.0, "step": 23409 }, { "epoch": 2.9779926218038417, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.360323190689087, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.8703078627586365, "num_tokens": 888687089.0, "step": 23410 }, { "epoch": 2.9781198320824323, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.6807916164398193, "learning_rate": 1e-05, "loss": 0.4628, "mean_token_accuracy": 0.8653304576873779, "num_tokens": 888732221.0, "step": 23411 }, { "epoch": 2.978247042361023, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.2866342067718506, "learning_rate": 1e-05, "loss": 0.3857, "mean_token_accuracy": 0.8929781913757324, "num_tokens": 888768992.0, "step": 23412 }, { "epoch": 2.9783742526396133, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.8095805644989014, "learning_rate": 1e-05, "loss": 0.4094, "mean_token_accuracy": 0.888923168182373, "num_tokens": 888809834.0, "step": 23413 }, { "epoch": 2.978501462918204, "ewc_loss": 0.0830078125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.696846008300781e-05, "grad_norm": 3.198125123977661, "learning_rate": 1e-05, "loss": 0.4108, "mean_token_accuracy": 0.8828775882720947, "num_tokens": 888851943.0, "step": 23414 }, { "epoch": 2.9786286731967944, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.03153657913208, "learning_rate": 1e-05, "loss": 0.4303, "mean_token_accuracy": 0.8804143071174622, "num_tokens": 888890465.0, "step": 23415 }, { "epoch": 2.978755883475385, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.174316167831421, "learning_rate": 1e-05, "loss": 0.4439, "mean_token_accuracy": 0.8765811920166016, "num_tokens": 888932954.0, "step": 23416 }, { "epoch": 2.9788830937539754, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.338079452514648, "learning_rate": 1e-05, "loss": 0.4472, "mean_token_accuracy": 0.88042151927948, "num_tokens": 888973916.0, "step": 23417 }, { "epoch": 2.979010304032566, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.3518505096435547, "learning_rate": 1e-05, "loss": 0.4937, "mean_token_accuracy": 0.8598970770835876, "num_tokens": 889017202.0, "step": 23418 }, { "epoch": 2.9791375143111565, "ewc_loss": 0.109375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 7.343292236328125e-05, "grad_norm": 4.570598602294922, "learning_rate": 1e-05, "loss": 0.5048, "mean_token_accuracy": 0.8650652766227722, "num_tokens": 889053604.0, "step": 23419 }, { "epoch": 2.979264724589747, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 3.7509469985961914, "learning_rate": 1e-05, "loss": 0.4493, "mean_token_accuracy": 0.8773746490478516, "num_tokens": 889087328.0, "step": 23420 }, { "epoch": 2.979391934868337, "ewc_loss": 0.10595703125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.446565628051758, "learning_rate": 1e-05, "loss": 0.4748, "mean_token_accuracy": 0.8728444576263428, "num_tokens": 889125209.0, "step": 23421 }, { "epoch": 2.979519145146928, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 3.9448494911193848, "learning_rate": 1e-05, "loss": 0.4705, "mean_token_accuracy": 0.8735692501068115, "num_tokens": 889160968.0, "step": 23422 }, { "epoch": 2.979646355425518, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.204562664031982, "learning_rate": 1e-05, "loss": 0.4804, "mean_token_accuracy": 0.8711742162704468, "num_tokens": 889201710.0, "step": 23423 }, { "epoch": 2.979773565704109, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.9040584564208984, "learning_rate": 1e-05, "loss": 0.5098, "mean_token_accuracy": 0.8620610237121582, "num_tokens": 889237003.0, "step": 23424 }, { "epoch": 2.979900775982699, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 4.001418590545654, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8686212301254272, "num_tokens": 889278775.0, "step": 23425 }, { "epoch": 2.98002798626129, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.7605409622192383, "learning_rate": 1e-05, "loss": 0.4765, "mean_token_accuracy": 0.871721088886261, "num_tokens": 889320888.0, "step": 23426 }, { "epoch": 2.9801551965398803, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8740499019622803, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8805191516876221, "num_tokens": 889356746.0, "step": 23427 }, { "epoch": 2.980282406818471, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.621047019958496, "learning_rate": 1e-05, "loss": 0.4151, "mean_token_accuracy": 0.8878048658370972, "num_tokens": 889400771.0, "step": 23428 }, { "epoch": 2.9804096170970613, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.822180986404419, "learning_rate": 1e-05, "loss": 0.4601, "mean_token_accuracy": 0.8719785213470459, "num_tokens": 889440459.0, "step": 23429 }, { "epoch": 2.980536827375652, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.269050598144531e-05, "grad_norm": 3.5131824016571045, "learning_rate": 1e-05, "loss": 0.4247, "mean_token_accuracy": 0.8839753866195679, "num_tokens": 889479782.0, "step": 23430 }, { "epoch": 2.9806640376542424, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.745887756347656e-05, "grad_norm": 3.83239483833313, "learning_rate": 1e-05, "loss": 0.3847, "mean_token_accuracy": 0.8950745463371277, "num_tokens": 889520167.0, "step": 23431 }, { "epoch": 2.980791247932833, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.078315734863281e-05, "grad_norm": 3.4147534370422363, "learning_rate": 1e-05, "loss": 0.4334, "mean_token_accuracy": 0.8812406063079834, "num_tokens": 889567589.0, "step": 23432 }, { "epoch": 2.9809184582114234, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.936622619628906e-05, "grad_norm": 3.9241068363189697, "learning_rate": 1e-05, "loss": 0.4503, "mean_token_accuracy": 0.8782687783241272, "num_tokens": 889607703.0, "step": 23433 }, { "epoch": 2.981045668490014, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.369401216506958, "learning_rate": 1e-05, "loss": 0.4416, "mean_token_accuracy": 0.8728708028793335, "num_tokens": 889645735.0, "step": 23434 }, { "epoch": 2.9811728787686045, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 4.065908432006836, "learning_rate": 1e-05, "loss": 0.4252, "mean_token_accuracy": 0.8834230899810791, "num_tokens": 889680035.0, "step": 23435 }, { "epoch": 2.981300089047195, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.3755691051483154, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8777980804443359, "num_tokens": 889720234.0, "step": 23436 }, { "epoch": 2.9814272993257855, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.200307369232178, "learning_rate": 1e-05, "loss": 0.4183, "mean_token_accuracy": 0.8872080445289612, "num_tokens": 889756070.0, "step": 23437 }, { "epoch": 2.981554509604376, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.513355255126953, "learning_rate": 1e-05, "loss": 0.4351, "mean_token_accuracy": 0.8803136348724365, "num_tokens": 889796499.0, "step": 23438 }, { "epoch": 2.9816817198829666, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.23433256149292, "learning_rate": 1e-05, "loss": 0.5215, "mean_token_accuracy": 0.859473466873169, "num_tokens": 889830804.0, "step": 23439 }, { "epoch": 2.981808930161557, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.008148193359375e-05, "grad_norm": 3.983626365661621, "learning_rate": 1e-05, "loss": 0.4783, "mean_token_accuracy": 0.8662701845169067, "num_tokens": 889865228.0, "step": 23440 }, { "epoch": 2.9819361404401477, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.837318181991577, "learning_rate": 1e-05, "loss": 0.4456, "mean_token_accuracy": 0.8755877614021301, "num_tokens": 889900780.0, "step": 23441 }, { "epoch": 2.982063350718738, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.7188100814819336, "learning_rate": 1e-05, "loss": 0.4489, "mean_token_accuracy": 0.8797003626823425, "num_tokens": 889944366.0, "step": 23442 }, { "epoch": 2.9821905609973287, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.609729051589966, "learning_rate": 1e-05, "loss": 0.3803, "mean_token_accuracy": 0.8965870141983032, "num_tokens": 889979739.0, "step": 23443 }, { "epoch": 2.9823177712759192, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.732038736343384, "learning_rate": 1e-05, "loss": 0.4259, "mean_token_accuracy": 0.8796522617340088, "num_tokens": 890017288.0, "step": 23444 }, { "epoch": 2.9824449815545098, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.460871696472168, "learning_rate": 1e-05, "loss": 0.4076, "mean_token_accuracy": 0.8832948207855225, "num_tokens": 890050848.0, "step": 23445 }, { "epoch": 2.9825721918331, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.7665865421295166, "learning_rate": 1e-05, "loss": 0.4298, "mean_token_accuracy": 0.8802983164787292, "num_tokens": 890087985.0, "step": 23446 }, { "epoch": 2.982699402111691, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.3506438732147217, "learning_rate": 1e-05, "loss": 0.4602, "mean_token_accuracy": 0.8696392774581909, "num_tokens": 890128292.0, "step": 23447 }, { "epoch": 2.982826612390281, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.847456693649292, "learning_rate": 1e-05, "loss": 0.4345, "mean_token_accuracy": 0.8809516429901123, "num_tokens": 890176130.0, "step": 23448 }, { "epoch": 2.982953822668872, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.275000810623169, "learning_rate": 1e-05, "loss": 0.4348, "mean_token_accuracy": 0.8777270913124084, "num_tokens": 890212547.0, "step": 23449 }, { "epoch": 2.983081032947462, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.010876655578613, "learning_rate": 1e-05, "loss": 0.4677, "mean_token_accuracy": 0.8735109567642212, "num_tokens": 890257423.0, "step": 23450 }, { "epoch": 2.983208243226053, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.2593231201171875, "learning_rate": 1e-05, "loss": 0.4376, "mean_token_accuracy": 0.8773495554924011, "num_tokens": 890299465.0, "step": 23451 }, { "epoch": 2.983335453504643, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.210739612579346, "learning_rate": 1e-05, "loss": 0.4253, "mean_token_accuracy": 0.8856158256530762, "num_tokens": 890340010.0, "step": 23452 }, { "epoch": 2.9834626637832335, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.37670636177063, "learning_rate": 1e-05, "loss": 0.448, "mean_token_accuracy": 0.8765925765037537, "num_tokens": 890380865.0, "step": 23453 }, { "epoch": 2.983589874061824, "ewc_loss": 0.10595703125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.351336479187012, "learning_rate": 1e-05, "loss": 0.4195, "mean_token_accuracy": 0.8907009363174438, "num_tokens": 890416710.0, "step": 23454 }, { "epoch": 2.9837170843404146, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.6038498878479004, "learning_rate": 1e-05, "loss": 0.414, "mean_token_accuracy": 0.8900404572486877, "num_tokens": 890452963.0, "step": 23455 }, { "epoch": 2.983844294619005, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.301085472106934, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8831999897956848, "num_tokens": 890493308.0, "step": 23456 }, { "epoch": 2.9839715048975957, "ewc_loss": 0.099609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.341934204101562e-05, "grad_norm": 3.7749149799346924, "learning_rate": 1e-05, "loss": 0.4672, "mean_token_accuracy": 0.8733468651771545, "num_tokens": 890529279.0, "step": 23457 }, { "epoch": 2.984098715176186, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.117822170257568, "learning_rate": 1e-05, "loss": 0.504, "mean_token_accuracy": 0.8588082790374756, "num_tokens": 890568866.0, "step": 23458 }, { "epoch": 2.9842259254547767, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 3.764849901199341, "learning_rate": 1e-05, "loss": 0.4337, "mean_token_accuracy": 0.8822681307792664, "num_tokens": 890613077.0, "step": 23459 }, { "epoch": 2.9843531357333672, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.927574396133423, "learning_rate": 1e-05, "loss": 0.4403, "mean_token_accuracy": 0.8795506358146667, "num_tokens": 890653599.0, "step": 23460 }, { "epoch": 2.9844803460119578, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.7078418731689453, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.879397988319397, "num_tokens": 890694581.0, "step": 23461 }, { "epoch": 2.9846075562905483, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.804776668548584, "learning_rate": 1e-05, "loss": 0.411, "mean_token_accuracy": 0.8844974040985107, "num_tokens": 890725912.0, "step": 23462 }, { "epoch": 2.984734766569139, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.4836273193359375e-05, "grad_norm": 3.6300623416900635, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8764932751655579, "num_tokens": 890763028.0, "step": 23463 }, { "epoch": 2.9848619768477294, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.7166852951049805, "learning_rate": 1e-05, "loss": 0.4406, "mean_token_accuracy": 0.8786427974700928, "num_tokens": 890803318.0, "step": 23464 }, { "epoch": 2.98498918712632, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.551746368408203, "learning_rate": 1e-05, "loss": 0.4511, "mean_token_accuracy": 0.8755893707275391, "num_tokens": 890841405.0, "step": 23465 }, { "epoch": 2.9851163974049104, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6690683364868164, "learning_rate": 1e-05, "loss": 0.4186, "mean_token_accuracy": 0.8830378651618958, "num_tokens": 890884271.0, "step": 23466 }, { "epoch": 2.985243607683501, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.478898525238037, "learning_rate": 1e-05, "loss": 0.4087, "mean_token_accuracy": 0.8865680694580078, "num_tokens": 890917747.0, "step": 23467 }, { "epoch": 2.9853708179620915, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6885159015655518, "learning_rate": 1e-05, "loss": 0.4256, "mean_token_accuracy": 0.8836290240287781, "num_tokens": 890958659.0, "step": 23468 }, { "epoch": 2.985498028240682, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4083786010742188, "learning_rate": 1e-05, "loss": 0.3903, "mean_token_accuracy": 0.893955409526825, "num_tokens": 890993792.0, "step": 23469 }, { "epoch": 2.9856252385192725, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.747292995452881, "learning_rate": 1e-05, "loss": 0.4211, "mean_token_accuracy": 0.8817279934883118, "num_tokens": 891030873.0, "step": 23470 }, { "epoch": 2.9857524487978626, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.3365135192871094, "learning_rate": 1e-05, "loss": 0.3891, "mean_token_accuracy": 0.893893837928772, "num_tokens": 891068062.0, "step": 23471 }, { "epoch": 2.9858796590764536, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.8551433086395264, "learning_rate": 1e-05, "loss": 0.4726, "mean_token_accuracy": 0.8686695098876953, "num_tokens": 891108749.0, "step": 23472 }, { "epoch": 2.9860068693550437, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.863739013671875e-05, "grad_norm": 3.2926859855651855, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8768528699874878, "num_tokens": 891144749.0, "step": 23473 }, { "epoch": 2.9861340796336346, "ewc_loss": 0.09765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.151199340820312e-05, "grad_norm": 4.032957553863525, "learning_rate": 1e-05, "loss": 0.4925, "mean_token_accuracy": 0.8661569356918335, "num_tokens": 891180795.0, "step": 23474 }, { "epoch": 2.9862612899122247, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.30147123336792, "learning_rate": 1e-05, "loss": 0.4478, "mean_token_accuracy": 0.8724104166030884, "num_tokens": 891221333.0, "step": 23475 }, { "epoch": 2.9863885001908153, "ewc_loss": 0.1025390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.628036499023438e-05, "grad_norm": 4.199859142303467, "learning_rate": 1e-05, "loss": 0.4483, "mean_token_accuracy": 0.8797097206115723, "num_tokens": 891258414.0, "step": 23476 }, { "epoch": 2.986515710469406, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.4409804344177246, "learning_rate": 1e-05, "loss": 0.4674, "mean_token_accuracy": 0.8711211681365967, "num_tokens": 891296864.0, "step": 23477 }, { "epoch": 2.9866429207479963, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.249102592468262, "learning_rate": 1e-05, "loss": 0.4016, "mean_token_accuracy": 0.894852876663208, "num_tokens": 891338191.0, "step": 23478 }, { "epoch": 2.986770131026587, "ewc_loss": 0.09619140625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.62522554397583, "learning_rate": 1e-05, "loss": 0.4558, "mean_token_accuracy": 0.8748595714569092, "num_tokens": 891377354.0, "step": 23479 }, { "epoch": 2.9868973413051774, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.137421607971191, "learning_rate": 1e-05, "loss": 0.4356, "mean_token_accuracy": 0.881436288356781, "num_tokens": 891415868.0, "step": 23480 }, { "epoch": 2.987024551583768, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.7192306518554688, "learning_rate": 1e-05, "loss": 0.4241, "mean_token_accuracy": 0.8817272186279297, "num_tokens": 891449548.0, "step": 23481 }, { "epoch": 2.9871517618623584, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.946108341217041, "learning_rate": 1e-05, "loss": 0.4996, "mean_token_accuracy": 0.85910964012146, "num_tokens": 891485937.0, "step": 23482 }, { "epoch": 2.987278972140949, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.716101884841919, "learning_rate": 1e-05, "loss": 0.4232, "mean_token_accuracy": 0.8827745318412781, "num_tokens": 891522458.0, "step": 23483 }, { "epoch": 2.9874061824195395, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.745525598526001, "learning_rate": 1e-05, "loss": 0.476, "mean_token_accuracy": 0.8673563599586487, "num_tokens": 891560804.0, "step": 23484 }, { "epoch": 2.98753339269813, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.69020414352417, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.889488697052002, "num_tokens": 891595692.0, "step": 23485 }, { "epoch": 2.9876606029767205, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.584846258163452, "learning_rate": 1e-05, "loss": 0.4137, "mean_token_accuracy": 0.8872373700141907, "num_tokens": 891632007.0, "step": 23486 }, { "epoch": 2.987787813255311, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.435943603515625e-05, "grad_norm": 3.667109251022339, "learning_rate": 1e-05, "loss": 0.4208, "mean_token_accuracy": 0.8848861455917358, "num_tokens": 891671605.0, "step": 23487 }, { "epoch": 2.9879150235339016, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.4952964782714844, "learning_rate": 1e-05, "loss": 0.4694, "mean_token_accuracy": 0.8674308061599731, "num_tokens": 891708369.0, "step": 23488 }, { "epoch": 2.988042233812492, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6604676246643066, "learning_rate": 1e-05, "loss": 0.4078, "mean_token_accuracy": 0.8861520886421204, "num_tokens": 891744673.0, "step": 23489 }, { "epoch": 2.9881694440910826, "ewc_loss": 0.0869140625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.054473876953125e-05, "grad_norm": 3.4220516681671143, "learning_rate": 1e-05, "loss": 0.4393, "mean_token_accuracy": 0.8764457106590271, "num_tokens": 891781971.0, "step": 23490 }, { "epoch": 2.988296654369673, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.692047357559204, "learning_rate": 1e-05, "loss": 0.4112, "mean_token_accuracy": 0.8889532089233398, "num_tokens": 891814079.0, "step": 23491 }, { "epoch": 2.9884238646482637, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.361531972885132, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8851767778396606, "num_tokens": 891852188.0, "step": 23492 }, { "epoch": 2.9885510749268542, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.53131103515625e-05, "grad_norm": 3.75000262260437, "learning_rate": 1e-05, "loss": 0.4443, "mean_token_accuracy": 0.8757283091545105, "num_tokens": 891894048.0, "step": 23493 }, { "epoch": 2.9886782852054448, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.288496971130371, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8780055642127991, "num_tokens": 891935199.0, "step": 23494 }, { "epoch": 2.9888054954840353, "ewc_loss": 0.09423828125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.817413330078125e-05, "grad_norm": 3.8714444637298584, "learning_rate": 1e-05, "loss": 0.4357, "mean_token_accuracy": 0.8804755210876465, "num_tokens": 891977686.0, "step": 23495 }, { "epoch": 2.9889327057626254, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.792213439941406e-05, "grad_norm": 3.2477524280548096, "learning_rate": 1e-05, "loss": 0.4578, "mean_token_accuracy": 0.8691550493240356, "num_tokens": 892017501.0, "step": 23496 }, { "epoch": 2.9890599160412163, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 4.054586887359619, "learning_rate": 1e-05, "loss": 0.4715, "mean_token_accuracy": 0.8680791258811951, "num_tokens": 892057369.0, "step": 23497 }, { "epoch": 2.9891871263198064, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.2586796283721924, "learning_rate": 1e-05, "loss": 0.4339, "mean_token_accuracy": 0.8779503107070923, "num_tokens": 892093998.0, "step": 23498 }, { "epoch": 2.9893143365983974, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.273205280303955, "learning_rate": 1e-05, "loss": 0.4171, "mean_token_accuracy": 0.8912848234176636, "num_tokens": 892125835.0, "step": 23499 }, { "epoch": 2.9894415468769875, "ewc_loss": 0.09326171875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.7220458984375e-05, "grad_norm": 3.4217498302459717, "learning_rate": 1e-05, "loss": 0.4469, "mean_token_accuracy": 0.8788472414016724, "num_tokens": 892160340.0, "step": 23500 }, { "epoch": 2.989568757155578, "ewc_loss": 0.10498046875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.866455078125e-05, "grad_norm": 4.366230487823486, "learning_rate": 1e-05, "loss": 0.4555, "mean_token_accuracy": 0.8757428526878357, "num_tokens": 892202096.0, "step": 23501 }, { "epoch": 2.9896959674341685, "ewc_loss": 0.0986328125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.246566772460938e-05, "grad_norm": 3.649385690689087, "learning_rate": 1e-05, "loss": 0.4218, "mean_token_accuracy": 0.8873474597930908, "num_tokens": 892241942.0, "step": 23502 }, { "epoch": 2.989823177712759, "ewc_loss": 0.10400390625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.771087646484375e-05, "grad_norm": 4.278569221496582, "learning_rate": 1e-05, "loss": 0.4818, "mean_token_accuracy": 0.8681782484054565, "num_tokens": 892284724.0, "step": 23503 }, { "epoch": 2.9899503879913496, "ewc_loss": 0.09912109375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.29425048828125e-05, "grad_norm": 3.771228075027466, "learning_rate": 1e-05, "loss": 0.4136, "mean_token_accuracy": 0.8905044198036194, "num_tokens": 892324761.0, "step": 23504 }, { "epoch": 2.99007759826994, "ewc_loss": 0.10107421875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.4849853515625e-05, "grad_norm": 4.08475923538208, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.8892982006072998, "num_tokens": 892358128.0, "step": 23505 }, { "epoch": 2.9902048085485307, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.7653181552886963, "learning_rate": 1e-05, "loss": 0.4653, "mean_token_accuracy": 0.8735266923904419, "num_tokens": 892400380.0, "step": 23506 }, { "epoch": 2.990332018827121, "ewc_loss": 0.095703125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.984306335449219e-05, "grad_norm": 3.8989973068237305, "learning_rate": 1e-05, "loss": 0.4721, "mean_token_accuracy": 0.8714711666107178, "num_tokens": 892439141.0, "step": 23507 }, { "epoch": 2.9904592291057117, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.841255187988281e-05, "grad_norm": 3.70699405670166, "learning_rate": 1e-05, "loss": 0.4401, "mean_token_accuracy": 0.8767512440681458, "num_tokens": 892480983.0, "step": 23508 }, { "epoch": 2.9905864393843022, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.650520324707031e-05, "grad_norm": 3.7474448680877686, "learning_rate": 1e-05, "loss": 0.4505, "mean_token_accuracy": 0.8785781860351562, "num_tokens": 892518220.0, "step": 23509 }, { "epoch": 2.9907136496628928, "ewc_loss": 0.09130859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.507469177246094e-05, "grad_norm": 3.6427037715911865, "learning_rate": 1e-05, "loss": 0.4311, "mean_token_accuracy": 0.8798505663871765, "num_tokens": 892549308.0, "step": 23510 }, { "epoch": 2.9908408599414833, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6511752605438232, "learning_rate": 1e-05, "loss": 0.4474, "mean_token_accuracy": 0.8775979280471802, "num_tokens": 892590154.0, "step": 23511 }, { "epoch": 2.990968070220074, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.5615015029907227, "learning_rate": 1e-05, "loss": 0.3814, "mean_token_accuracy": 0.8964002132415771, "num_tokens": 892630951.0, "step": 23512 }, { "epoch": 2.9910952804986644, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.6154837608337402, "learning_rate": 1e-05, "loss": 0.454, "mean_token_accuracy": 0.8733813762664795, "num_tokens": 892675345.0, "step": 23513 }, { "epoch": 2.991222490777255, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.503098726272583, "learning_rate": 1e-05, "loss": 0.4584, "mean_token_accuracy": 0.871417760848999, "num_tokens": 892718129.0, "step": 23514 }, { "epoch": 2.9913497010558454, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.6103098392486572, "learning_rate": 1e-05, "loss": 0.4409, "mean_token_accuracy": 0.8786488771438599, "num_tokens": 892757159.0, "step": 23515 }, { "epoch": 2.991476911334436, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.441296339035034, "learning_rate": 1e-05, "loss": 0.4277, "mean_token_accuracy": 0.8809058666229248, "num_tokens": 892797872.0, "step": 23516 }, { "epoch": 2.9916041216130265, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.643279790878296, "learning_rate": 1e-05, "loss": 0.4596, "mean_token_accuracy": 0.8711969256401062, "num_tokens": 892837213.0, "step": 23517 }, { "epoch": 2.991731331891617, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.364429473876953, "learning_rate": 1e-05, "loss": 0.4388, "mean_token_accuracy": 0.8739221096038818, "num_tokens": 892878651.0, "step": 23518 }, { "epoch": 2.991858542170207, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.725820302963257, "learning_rate": 1e-05, "loss": 0.3973, "mean_token_accuracy": 0.888085126876831, "num_tokens": 892913739.0, "step": 23519 }, { "epoch": 2.991985752448798, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.3021106719970703, "learning_rate": 1e-05, "loss": 0.4341, "mean_token_accuracy": 0.8767306804656982, "num_tokens": 892949649.0, "step": 23520 }, { "epoch": 2.992112962727388, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.7697296142578125e-05, "grad_norm": 3.938286542892456, "learning_rate": 1e-05, "loss": 0.4327, "mean_token_accuracy": 0.8824129104614258, "num_tokens": 892988831.0, "step": 23521 }, { "epoch": 2.992240173005979, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.234740972518921, "learning_rate": 1e-05, "loss": 0.4269, "mean_token_accuracy": 0.8797409534454346, "num_tokens": 893025311.0, "step": 23522 }, { "epoch": 2.992367383284569, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.12601375579834, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8722363710403442, "num_tokens": 893072461.0, "step": 23523 }, { "epoch": 2.99249459356316, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.2819809913635254, "learning_rate": 1e-05, "loss": 0.4579, "mean_token_accuracy": 0.8719290494918823, "num_tokens": 893109268.0, "step": 23524 }, { "epoch": 2.9926218038417502, "ewc_loss": 0.10595703125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.961822509765625e-05, "grad_norm": 4.368912220001221, "learning_rate": 1e-05, "loss": 0.4868, "mean_token_accuracy": 0.8714510798454285, "num_tokens": 893139033.0, "step": 23525 }, { "epoch": 2.992749014120341, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.5089991092681885, "learning_rate": 1e-05, "loss": 0.4181, "mean_token_accuracy": 0.8841618895530701, "num_tokens": 893180056.0, "step": 23526 }, { "epoch": 2.9928762243989313, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.417145729064941, "learning_rate": 1e-05, "loss": 0.4668, "mean_token_accuracy": 0.8747519254684448, "num_tokens": 893214164.0, "step": 23527 }, { "epoch": 2.993003434677522, "ewc_loss": 0.1005859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.437301635742188e-05, "grad_norm": 3.7723453044891357, "learning_rate": 1e-05, "loss": 0.4426, "mean_token_accuracy": 0.8794506788253784, "num_tokens": 893255030.0, "step": 23528 }, { "epoch": 2.9931306449561124, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.242918968200684, "learning_rate": 1e-05, "loss": 0.4609, "mean_token_accuracy": 0.872552752494812, "num_tokens": 893287661.0, "step": 23529 }, { "epoch": 2.993257855234703, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 3.8506455421447754, "learning_rate": 1e-05, "loss": 0.5144, "mean_token_accuracy": 0.8561239242553711, "num_tokens": 893328461.0, "step": 23530 }, { "epoch": 2.9933850655132934, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.9965977668762207, "learning_rate": 1e-05, "loss": 0.4288, "mean_token_accuracy": 0.8826373815536499, "num_tokens": 893367543.0, "step": 23531 }, { "epoch": 2.993512275791884, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 3.777602434158325, "learning_rate": 1e-05, "loss": 0.4128, "mean_token_accuracy": 0.8897524476051331, "num_tokens": 893401180.0, "step": 23532 }, { "epoch": 2.9936394860704745, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.8288021087646484, "learning_rate": 1e-05, "loss": 0.3804, "mean_token_accuracy": 0.8966964483261108, "num_tokens": 893429823.0, "step": 23533 }, { "epoch": 2.993766696349065, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.6743621826171875e-05, "grad_norm": 3.6836066246032715, "learning_rate": 1e-05, "loss": 0.433, "mean_token_accuracy": 0.876662015914917, "num_tokens": 893472730.0, "step": 23534 }, { "epoch": 2.9938939066276555, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.602836608886719e-05, "grad_norm": 3.709728717803955, "learning_rate": 1e-05, "loss": 0.4244, "mean_token_accuracy": 0.8820182085037231, "num_tokens": 893516506.0, "step": 23535 }, { "epoch": 2.994021116906246, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.6020822525024414, "learning_rate": 1e-05, "loss": 0.4402, "mean_token_accuracy": 0.8777173757553101, "num_tokens": 893560539.0, "step": 23536 }, { "epoch": 2.9941483271848366, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.6678645610809326, "learning_rate": 1e-05, "loss": 0.435, "mean_token_accuracy": 0.8793085217475891, "num_tokens": 893595986.0, "step": 23537 }, { "epoch": 2.994275537463427, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.1975250244140625e-05, "grad_norm": 3.5229198932647705, "learning_rate": 1e-05, "loss": 0.4086, "mean_token_accuracy": 0.886223316192627, "num_tokens": 893632997.0, "step": 23538 }, { "epoch": 2.9944027477420176, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.66951060295105, "learning_rate": 1e-05, "loss": 0.4594, "mean_token_accuracy": 0.8747403621673584, "num_tokens": 893671889.0, "step": 23539 }, { "epoch": 2.994529958020608, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.0067901611328125e-05, "grad_norm": 3.4449777603149414, "learning_rate": 1e-05, "loss": 0.4011, "mean_token_accuracy": 0.8885853290557861, "num_tokens": 893708781.0, "step": 23540 }, { "epoch": 2.9946571682991987, "ewc_loss": 0.0908203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.459785461425781e-05, "grad_norm": 3.7014389038085938, "learning_rate": 1e-05, "loss": 0.44, "mean_token_accuracy": 0.8741931915283203, "num_tokens": 893745590.0, "step": 23541 }, { "epoch": 2.9947843785777892, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.361121654510498, "learning_rate": 1e-05, "loss": 0.4132, "mean_token_accuracy": 0.8844048976898193, "num_tokens": 893783599.0, "step": 23542 }, { "epoch": 2.9949115888563798, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.626678466796875e-05, "grad_norm": 3.783112049102783, "learning_rate": 1e-05, "loss": 0.3812, "mean_token_accuracy": 0.8958707451820374, "num_tokens": 893822603.0, "step": 23543 }, { "epoch": 2.99503879913497, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.76837158203125e-05, "grad_norm": 3.298969268798828, "learning_rate": 1e-05, "loss": 0.4243, "mean_token_accuracy": 0.8830281496047974, "num_tokens": 893857318.0, "step": 23544 }, { "epoch": 2.995166009413561, "ewc_loss": 0.09521484375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.91278076171875e-05, "grad_norm": 3.917586326599121, "learning_rate": 1e-05, "loss": 0.4382, "mean_token_accuracy": 0.8744750618934631, "num_tokens": 893896269.0, "step": 23545 }, { "epoch": 2.995293219692151, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.887580871582031e-05, "grad_norm": 3.2680625915527344, "learning_rate": 1e-05, "loss": 0.4666, "mean_token_accuracy": 0.8657878637313843, "num_tokens": 893936877.0, "step": 23546 }, { "epoch": 2.995420429970742, "ewc_loss": 0.10009765625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.389617919921875e-05, "grad_norm": 4.105928897857666, "learning_rate": 1e-05, "loss": 0.4738, "mean_token_accuracy": 0.8774497509002686, "num_tokens": 893970014.0, "step": 23547 }, { "epoch": 2.995547640249332, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.322355270385742, "learning_rate": 1e-05, "loss": 0.4352, "mean_token_accuracy": 0.8791525959968567, "num_tokens": 894005568.0, "step": 23548 }, { "epoch": 2.995674850527923, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 4.263301372528076, "learning_rate": 1e-05, "loss": 0.452, "mean_token_accuracy": 0.8785090446472168, "num_tokens": 894053880.0, "step": 23549 }, { "epoch": 2.995802060806513, "ewc_loss": 0.09375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.793571472167969e-05, "grad_norm": 3.5043954849243164, "learning_rate": 1e-05, "loss": 0.3948, "mean_token_accuracy": 0.8914852142333984, "num_tokens": 894083981.0, "step": 23550 }, { "epoch": 2.9959292710851035, "ewc_loss": 0.10302734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.67572021484375e-05, "grad_norm": 4.273972988128662, "learning_rate": 1e-05, "loss": 0.4467, "mean_token_accuracy": 0.8820579051971436, "num_tokens": 894125315.0, "step": 23551 }, { "epoch": 2.996056481363694, "ewc_loss": 0.09814453125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.198883056640625e-05, "grad_norm": 3.6813242435455322, "learning_rate": 1e-05, "loss": 0.4621, "mean_token_accuracy": 0.8738288879394531, "num_tokens": 894165195.0, "step": 23552 }, { "epoch": 2.9961836916422846, "ewc_loss": 0.1015625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.532669067382812e-05, "grad_norm": 4.133170127868652, "learning_rate": 1e-05, "loss": 0.4683, "mean_token_accuracy": 0.8710222244262695, "num_tokens": 894206443.0, "step": 23553 }, { "epoch": 2.996310901920875, "ewc_loss": 0.09716796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.103515625e-05, "grad_norm": 3.75040340423584, "learning_rate": 1e-05, "loss": 0.4344, "mean_token_accuracy": 0.8821150064468384, "num_tokens": 894240897.0, "step": 23554 }, { "epoch": 2.9964381121994657, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.031990051269531e-05, "grad_norm": 3.9280102252960205, "learning_rate": 1e-05, "loss": 0.412, "mean_token_accuracy": 0.8886562585830688, "num_tokens": 894277475.0, "step": 23555 }, { "epoch": 2.996565322478056, "ewc_loss": 0.0947265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.8650970458984375e-05, "grad_norm": 3.717745780944824, "learning_rate": 1e-05, "loss": 0.472, "mean_token_accuracy": 0.8683575391769409, "num_tokens": 894320231.0, "step": 23556 }, { "epoch": 2.9966925327566467, "ewc_loss": 0.0927734375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.698204040527344e-05, "grad_norm": 3.753527879714966, "learning_rate": 1e-05, "loss": 0.4373, "mean_token_accuracy": 0.8793059587478638, "num_tokens": 894359196.0, "step": 23557 }, { "epoch": 2.9968197430352372, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.6670548915863037, "learning_rate": 1e-05, "loss": 0.4548, "mean_token_accuracy": 0.8723281621932983, "num_tokens": 894401362.0, "step": 23558 }, { "epoch": 2.9969469533138278, "ewc_loss": 0.09033203125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.412101745605469e-05, "grad_norm": 3.624661684036255, "learning_rate": 1e-05, "loss": 0.4066, "mean_token_accuracy": 0.8871278762817383, "num_tokens": 894435445.0, "step": 23559 }, { "epoch": 2.9970741635924183, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.3882598876953125e-05, "grad_norm": 3.6173129081726074, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8741837739944458, "num_tokens": 894475968.0, "step": 23560 }, { "epoch": 2.997201373871009, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.2928924560546875e-05, "grad_norm": 3.558788299560547, "learning_rate": 1e-05, "loss": 0.4141, "mean_token_accuracy": 0.883628249168396, "num_tokens": 894511692.0, "step": 23561 }, { "epoch": 2.9973285841495994, "ewc_loss": 0.08935546875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.316734313964844e-05, "grad_norm": 3.5870845317840576, "learning_rate": 1e-05, "loss": 0.4624, "mean_token_accuracy": 0.8738843202590942, "num_tokens": 894550886.0, "step": 23562 }, { "epoch": 2.99745579442819, "ewc_loss": 0.08740234375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.125999450683594e-05, "grad_norm": 3.5082597732543945, "learning_rate": 1e-05, "loss": 0.3992, "mean_token_accuracy": 0.8901336789131165, "num_tokens": 894584811.0, "step": 23563 }, { "epoch": 2.9975830047067804, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.14984130859375e-05, "grad_norm": 3.5602405071258545, "learning_rate": 1e-05, "loss": 0.4482, "mean_token_accuracy": 0.8755590319633484, "num_tokens": 894622866.0, "step": 23564 }, { "epoch": 2.997710214985371, "ewc_loss": 0.08642578125, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.030632019042969e-05, "grad_norm": 3.465116262435913, "learning_rate": 1e-05, "loss": 0.3974, "mean_token_accuracy": 0.8935356140136719, "num_tokens": 894663513.0, "step": 23565 }, { "epoch": 2.9978374252639615, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5614848136901855, "learning_rate": 1e-05, "loss": 0.4491, "mean_token_accuracy": 0.8735164999961853, "num_tokens": 894702873.0, "step": 23566 }, { "epoch": 2.997964635542552, "ewc_loss": 0.0859375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.9591064453125e-05, "grad_norm": 3.4430298805236816, "learning_rate": 1e-05, "loss": 0.4177, "mean_token_accuracy": 0.8820775747299194, "num_tokens": 894737990.0, "step": 23567 }, { "epoch": 2.9980918458211425, "ewc_loss": 0.0888671875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.245208740234375e-05, "grad_norm": 3.5747270584106445, "learning_rate": 1e-05, "loss": 0.4285, "mean_token_accuracy": 0.8829303979873657, "num_tokens": 894775408.0, "step": 23568 }, { "epoch": 2.9982190560997326, "ewc_loss": 0.08544921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.935264587402344e-05, "grad_norm": 3.4445064067840576, "learning_rate": 1e-05, "loss": 0.4431, "mean_token_accuracy": 0.8732773661613464, "num_tokens": 894815702.0, "step": 23569 }, { "epoch": 2.9983462663783236, "ewc_loss": 0.087890625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.173683166503906e-05, "grad_norm": 3.5906927585601807, "learning_rate": 1e-05, "loss": 0.4304, "mean_token_accuracy": 0.8799490928649902, "num_tokens": 894849116.0, "step": 23570 }, { "epoch": 2.9984734766569137, "ewc_loss": 0.0849609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.9114227294921875e-05, "grad_norm": 3.3839712142944336, "learning_rate": 1e-05, "loss": 0.3965, "mean_token_accuracy": 0.8874373435974121, "num_tokens": 894882020.0, "step": 23571 }, { "epoch": 2.9986006869355046, "ewc_loss": 0.08984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.364418029785156e-05, "grad_norm": 3.6327176094055176, "learning_rate": 1e-05, "loss": 0.4326, "mean_token_accuracy": 0.8809329271316528, "num_tokens": 894916994.0, "step": 23572 }, { "epoch": 2.9987278972140947, "ewc_loss": 0.083984375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.8160552978515625e-05, "grad_norm": 3.328448534011841, "learning_rate": 1e-05, "loss": 0.4537, "mean_token_accuracy": 0.872651219367981, "num_tokens": 894952763.0, "step": 23573 }, { "epoch": 2.9988551074926852, "ewc_loss": 0.091796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.8263466358184814, "learning_rate": 1e-05, "loss": 0.4301, "mean_token_accuracy": 0.8805375695228577, "num_tokens": 894983427.0, "step": 23574 }, { "epoch": 2.9989823177712758, "ewc_loss": 0.08349609375, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.744529724121094e-05, "grad_norm": 3.214864492416382, "learning_rate": 1e-05, "loss": 0.4799, "mean_token_accuracy": 0.8663198351860046, "num_tokens": 895019082.0, "step": 23575 }, { "epoch": 2.9991095280498663, "ewc_loss": 0.0966796875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.0558319091796875e-05, "grad_norm": 4.000448226928711, "learning_rate": 1e-05, "loss": 0.408, "mean_token_accuracy": 0.8908421397209167, "num_tokens": 895054051.0, "step": 23576 }, { "epoch": 2.999236738328457, "ewc_loss": 0.08447265625, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 4.839897155761719e-05, "grad_norm": 3.1468334197998047, "learning_rate": 1e-05, "loss": 0.4246, "mean_token_accuracy": 0.8795061111450195, "num_tokens": 895090185.0, "step": 23577 }, { "epoch": 2.9993639486070474, "ewc_loss": 0.1044921875, "ewc_loss_diag": 0.0003604888916015625, "ewc_loss_parallel": 6.818771362304688e-05, "grad_norm": 4.337739944458008, "learning_rate": 1e-05, "loss": 0.4618, "mean_token_accuracy": 0.8773923516273499, "num_tokens": 895130159.0, "step": 23578 }, { "epoch": 2.999491158885638, "ewc_loss": 0.09228515625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 5.5789947509765625e-05, "grad_norm": 3.295658826828003, "learning_rate": 1e-05, "loss": 0.4379, "mean_token_accuracy": 0.8781321048736572, "num_tokens": 895169639.0, "step": 23579 }, { "epoch": 2.9996183691642284, "ewc_loss": 0.1103515625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 7.43865966796875e-05, "grad_norm": 4.633676052093506, "learning_rate": 1e-05, "loss": 0.4083, "mean_token_accuracy": 0.8942294120788574, "num_tokens": 895209939.0, "step": 23580 }, { "epoch": 2.999745579442819, "ewc_loss": 0.103515625, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 6.723403930664062e-05, "grad_norm": 3.751300573348999, "learning_rate": 1e-05, "loss": 0.4685, "mean_token_accuracy": 0.8718714118003845, "num_tokens": 895253073.0, "step": 23581 }, { "epoch": 2.9998727897214095, "ewc_loss": 0.109375, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 7.295608520507812e-05, "grad_norm": 4.546489715576172, "learning_rate": 1e-05, "loss": 0.464, "mean_token_accuracy": 0.8759496808052063, "num_tokens": 895295058.0, "step": 23582 }, { "epoch": 3.0, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 7.009506225585938e-05, "grad_norm": 4.018104076385498, "learning_rate": 1e-05, "loss": 0.4359, "mean_token_accuracy": 0.8839433789253235, "num_tokens": 895335272.0, "step": 23583 }, { "epoch": 3.0, "ewc_loss": 0.1064453125, "ewc_loss_diag": 0.0003643035888671875, "ewc_loss_parallel": 7.009506225585938e-05, "step": 23583, "total_flos": 5.156325133756924e+19, "train_loss": 0.46243270478726584, "train_runtime": 51121.2309, "train_samples_per_second": 7.381, "train_steps_per_second": 0.461 } ], "logging_steps": 1, "max_steps": 23583, "num_input_tokens_seen": 0, "num_train_epochs": 3, "save_steps": 11792, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 5.156325133756924e+19, "train_batch_size": 4, "trial_name": null, "trial_params": null }